id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2501.15744 | Manuj Yadav | Manuj Yadav, Jungsoo Kim, Valtteri Hongisto, Densil Cabrera, and
Richard de Dear | Noise disturbance and lack of privacy: Modeling acoustic dissatisfaction
in open-plan offices | The following article has been submitted to The Journal of the
Acoustical Society of America. After it is published, it will be found at
https://pubs.aip.org/asa/jasa | null | null | null | eess.AS cs.SD | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Open-plan offices are well-known to be adversely affected by acoustic issues.
This study aims to model acoustic dissatisfaction using measurements of room
acoustics, sound environment during occupancy, and occupant surveys (n = 349)
in 28 offices representing a diverse range of workplace parameters. As latent
factors, the contribution of $\textit{lack of privacy}$ (LackPriv) was 25%
higher than $\textit{noise disturbance}$ (NseDstrb) in predicting
$\textit{acoustic dissatisfaction}$ (AcDsat). Room acoustic metrics based on
sound pressure level (SPL) decay of speech ($L_{\text{p,A,s,4m}}$ and
$r_{\text{C}}$) were better in predicting these factors than distraction
distance ($r_{\text{D}}$) based on speech transmission index. This contradicts
previous findings, and the trends for SPL-based metrics in predicting AcDsat
and LackPriv go against expectations based on ISO 3382-3. For sound during
occupation, $L_{\text{A,90}}$ and psychoacoustic loudness ($N_{\text{90}}$)
predicted AcDsat, and a SPL fluctuation metric ($M_{\text{A,eq}}$) predicted
LackPriv. However, these metrics were weaker predictors than ISO 3382-3
metrics. Medium-sized offices exhibited higher dissatisfaction than larger
($\geq$50 occupants) offices. Dissatisfaction varied substantially across
parameters including ceiling heights, number of workstations, and years of
work, but not between offices with fixed seating compared to more flexible and
activity-based working configurations. Overall, these findings highlight the
complexities in characterizing occupants' perceptions using instrumental
acoustic measurements.
| [
{
"version": "v1",
"created": "Mon, 27 Jan 2025 03:10:07 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 22:19:42 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Yadav",
"Manuj",
""
],
[
"Kim",
"Jungsoo",
""
],
[
"Hongisto",
"Valtteri",
""
],
[
"Cabrera",
"Densil",
""
],
[
"de Dear",
"Richard",
""
]
] |
2501.15955 | Jiahao Chen | Jiahao Chen, Bin Qin, Jiangmeng Li, Hao Chen, Bing Su | Rethinking the Bias of Foundation Model under Long-tailed Distribution | Published as a conference paper in ICML 2025 | null | null | null | cs.LG cs.CV stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Long-tailed learning has garnered increasing attention due to its practical
significance. Among the various approaches, the fine-tuning paradigm has gained
considerable interest with the advent of foundation models. However, most
existing methods primarily focus on leveraging knowledge from these models,
overlooking the inherent biases introduced by the imbalanced training data they
rely on. In this paper, we examine how such imbalances from pre-training affect
long-tailed downstream tasks. Specifically, we find the imbalance biases
inherited in foundation models on downstream task as parameter imbalance and
data imbalance. During fine-tuning, we observe that parameter imbalance plays a
more critical role, while data imbalance can be mitigated using existing
re-balancing strategies. Moreover, we find that parameter imbalance cannot be
effectively addressed by current re-balancing techniques, such as adjusting the
logits, during training, unlike data imbalance. To tackle both imbalances
simultaneously, we build our method on causal learning and view the incomplete
semantic factor as the confounder, which brings spurious correlations between
input samples and labels. To resolve the negative effects of this, we propose a
novel backdoor adjustment method that learns the true causal effect between
input samples and labels, rather than merely fitting the correlations in the
data. Notably, we achieve an average performance increase of about $1.67\%$ on
each dataset.
| [
{
"version": "v1",
"created": "Mon, 27 Jan 2025 11:00:19 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 12:14:11 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Chen",
"Jiahao",
""
],
[
"Qin",
"Bin",
""
],
[
"Li",
"Jiangmeng",
""
],
[
"Chen",
"Hao",
""
],
[
"Su",
"Bing",
""
]
] |
2501.16161 | Xin Fang | Rocco Chiriv\`i, Martina Costa Cesari, Xin Fang, Peter Littelmann | Combinatorial Seshadri stratifications on normal toric varieties | v2: typo corrected, 31 pages, to appear in Journal of Algebra 60th
anniversary volume | null | null | null | math.AG math.CO | http://creativecommons.org/licenses/by/4.0/ | We apply the theory of Seshadri stratifications to embedded toric varieties
$X_P\subseteq \mathbb P(V)$ associated with a normal lattice polytope $P$. The
approach presented here is purely combinatorial and completely independent of
\cite{CFL}. In particular, we get a close connection between a certain class of
triangulations of the polytope $P$, Seshadri stratifications of $X_P$ arising
from torus orbit closures, and the associated degenerate semi-toric varieties.
In the last section we show that the approach here and the one in \cite{CFL}
produce the same quasi-valuations and hence the same degenerations of $X_P$.
| [
{
"version": "v1",
"created": "Mon, 27 Jan 2025 15:58:43 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 16:18:40 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Chirivì",
"Rocco",
""
],
[
"Cesari",
"Martina Costa",
""
],
[
"Fang",
"Xin",
""
],
[
"Littelmann",
"Peter",
""
]
] |
2501.16457 | Robert Corless | Robert M. Corless and Arthur C. Norman and Tomas Recio and William J.
Turkel and Stephen M. Watt | Symbolic Mathematical Computation 1965--1975: The View from a
Half-Century Perspective | 18 pages, 149 references | null | null | null | cs.GL | http://creativecommons.org/licenses/by-sa/4.0/ | The 2025 ISSAC conference in Guanajuato, Mexico, marks the 50th event in this
significant series, making it an ideal moment to reflect on the field's
history. This paper reviews the formative years of symbolic computation up to
1975, fifty years ago. By revisiting a period unfamiliar to most current
participants, this survey aims to shed light on once-pressing issues that are
now largely resolved and to highlight how some of today's challenges were
recognized earlier than expected.
| [
{
"version": "v1",
"created": "Mon, 27 Jan 2025 19:35:18 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 23:33:48 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Corless",
"Robert M.",
""
],
[
"Norman",
"Arthur C.",
""
],
[
"Recio",
"Tomas",
""
],
[
"Turkel",
"William J.",
""
],
[
"Watt",
"Stephen M.",
""
]
] |
2501.16504 | Majumder Haider | Majumder Haider, Imtiaz Ahmed, Zoheb Hassan, Timothy J. O'Shea,
Lingjia Liu and Danda B. Rawat | Digital Twin Enabled Site Specific Channel Precoding: Over the Air CIR
Inference | null | null | null | null | eess.SP cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates the significance of designing a reliable,
intelligent, and true physical environment-aware precoding scheme by leveraging
an accurately designed channel twin model to obtain realistic channel state
information (CSI) for cellular communication systems. Specifically, we propose
a fine-tuned multi-step channel twin design process that can render CSI very
close to the CSI of the actual environment. After generating a precise CSI, we
execute precoding using the obtained CSI at the transmitter end. We demonstrate
a two-step parameters' tuning approach to design channel twin by ray tracing
(RT) emulation, then further fine-tuning of CSI by employing an artificial
intelligence (AI) based algorithm can significantly reduce the gap between
actual CSI and the fine-tuned digital twin (DT) rendered CSI. The simulation
results show the effectiveness of the proposed novel approach in designing a
true physical environment-aware channel twin model.
| [
{
"version": "v1",
"created": "Mon, 27 Jan 2025 21:10:07 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Haider",
"Majumder",
""
],
[
"Ahmed",
"Imtiaz",
""
],
[
"Hassan",
"Zoheb",
""
],
[
"O'Shea",
"Timothy J.",
""
],
[
"Liu",
"Lingjia",
""
],
[
"Rawat",
"Danda B.",
""
]
] |
2501.17176 | Daniel Ortiz-Mart\'inez | Marc Ballestero-Rib\'o, Daniel Ortiz-Mart\'inez | Prompt-Based Cost-Effective Evaluation and Operation of ChatGPT as a
Computer Programming Teaching Assistant | null | null | null | null | cs.CY cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | The dream of achieving a student-teacher ratio of 1:1 is closer than ever
thanks to the emergence of large language models (LLMs). One potential
application of these models in the educational field would be to provide
feedback to students in university introductory programming courses, so that a
student struggling to solve a basic implementation problem could seek help from
an LLM available 24/7. This article focuses on studying three aspects related
to such an application. First, the performance of two well-known models,
GPT-3.5T and GPT-4T, in providing feedback to students is evaluated. The
empirical results showed that GPT-4T performs much better than GPT-3.5T,
however, it is not yet ready for use in a real-world scenario. This is due to
the possibility of generating incorrect information that potential users may
not always be able to detect. Second, the article proposes a carefully designed
prompt using in-context learning techniques that allows automating important
parts of the evaluation process, as well as providing a lower bound for the
fraction of feedbacks containing incorrect information, saving time and effort.
This was possible because the resulting feedback has a programmatically
analyzable structure that incorporates diagnostic information about the LLM's
performance in solving the requested task. Third, the article also suggests a
possible strategy for implementing a practical learning tool based on LLMs,
which is rooted on the proposed prompting techniques. This strategy opens up a
whole range of interesting possibilities from a pedagogical perspective.
| [
{
"version": "v1",
"created": "Fri, 24 Jan 2025 08:15:05 GMT"
},
{
"version": "v2",
"created": "Fri, 18 Apr 2025 10:35:50 GMT"
},
{
"version": "v3",
"created": "Sun, 4 May 2025 13:32:32 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Ballestero-Ribó",
"Marc",
""
],
[
"Ortiz-Martínez",
"Daniel",
""
]
] |
2501.17845 | Shreya Meel Ms | Shreya Meel, Xiangliang Kong, Thomas Jacob Maranzatto, Itzhak Tamo and
Sennur Ulukus | Private Information Retrieval on Multigraph-Based Replicated Storage | null | null | null | null | cs.IT cs.CR cs.NI eess.SP math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider the private information retrieval (PIR) problem for a
multigraph-based replication system, where each set of $r$ files is stored on
two of the servers according to an underlying $r$-multigraph. Our goal is to
establish upper and lower bounds on the PIR capacity of the $r$-multigraph.
Specifically, we first propose a construction for multigraph-based PIR systems
that leverages the symmetry of the underlying graph-based PIR scheme, deriving
a capacity lower bound for such multigraphs. Then, we establish a general upper
bound using linear programming, expressed as a function of the underlying graph
parameters. Our bounds are demonstrated to be tight for PIR systems on
multipaths for even number of vertices.
| [
{
"version": "v1",
"created": "Wed, 29 Jan 2025 18:48:22 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 23:38:56 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Meel",
"Shreya",
""
],
[
"Kong",
"Xiangliang",
""
],
[
"Maranzatto",
"Thomas Jacob",
""
],
[
"Tamo",
"Itzhak",
""
],
[
"Ulukus",
"Sennur",
""
]
] |
2501.18132 | Andy B. Day | Andy B. Day | Algebraically Skew Embeddings of Curves | 33 pages. Edited the definition for generic scrolls. Comments are
welcome | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Given a smooth complex variety $X$, an algebraically skew embedding of $X$ is
an embedding of $X$ into a complex projective space $\mathbb{P}^N$ such that
for any two points $x,y\in X$, their embedded tangent spaces in $\mathbb{P}^N$
do not intersect. In this work, we establish an upper bound and a lower bound
of the minimal dimension $N$ such that there exists an algebraically skew
embedding into $\mathbb{P}^N$ in terms of the dimension of the given smooth
variety $X$. Then we further classify the algebraic curves in terms of their
minimal skew embedding dimensions, and apply the same technique to other
one-parameter family of lines.
| [
{
"version": "v1",
"created": "Thu, 30 Jan 2025 04:43:58 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Feb 2025 01:51:07 GMT"
},
{
"version": "v3",
"created": "Mon, 21 Apr 2025 03:53:40 GMT"
},
{
"version": "v4",
"created": "Mon, 5 May 2025 03:40:57 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Day",
"Andy B.",
""
]
] |
2501.18458 | Meer Ashwinkumar | Meer Ashwinkumar, Domenico Orlando, Susanne Reffert, Giacomo
Sberveglieri | Asymptotics in the bi-Yang-Baxter Sigma Model | 29 pages, 2 figures. Further clarifications | null | null | YITP-24-175 | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Working in a sector of large charge is a powerful tool to analytically access
models that are either strongly coupled or otherwise difficult to solve
explicitly. In the context of integrable systems, Volin's method is exactly
such a large-charge approach. In this note, we apply this method to the
bi-Yang-Baxter deformed $SU(2)$ principal chiral model. Our main result is an
explicit expression for the free energy density as an asymptotic expansion. We
moreover determine the leading non-perturbative effects both via analytic
methods and a resurgence analysis.
| [
{
"version": "v1",
"created": "Thu, 30 Jan 2025 16:16:15 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 10:17:08 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Ashwinkumar",
"Meer",
""
],
[
"Orlando",
"Domenico",
""
],
[
"Reffert",
"Susanne",
""
],
[
"Sberveglieri",
"Giacomo",
""
]
] |
2501.19189 | Mihai Halic | Mihai Halic, Roshan Tajarod | Rationality and categorical properties of the moduli of instanton
bundles on the projective 3-space | We remove the technical 'endomorphism-general' condition introduced
in our previous posting arXiv:1805.06414, and strenghten the results. Section
6 added to v2 | null | null | null | math.AG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We prove the rationality and irreducibility of the moduli space of
mathematical instanton vector bundles of arbitrary rank and charge on $\mathbb
P^3$. In particular, the result applies to the rank-2 case. This problem was
first studied by Barth, Ellingsrud-Stromme, Hartshorne, Hirschowitz-Narasimhan
in the late 1970s. We also show that the mathematical instantons of variable
rank and charge form a monoidal category. The proof is based on an in-depth
analysis of the Barth-Hulek monad-construction and on a detailed description of
the moduli space of (framed and unframed) stable bundles on Hirzebruch
surfaces.
| [
{
"version": "v1",
"created": "Fri, 31 Jan 2025 14:56:12 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 13:34:00 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Halic",
"Mihai",
""
],
[
"Tajarod",
"Roshan",
""
]
] |
2502.00150 | Hugo Diaz-Norambuena | Alen Alexanderian, Hugo D\'iaz, Vishwas Rao and Arvind K. Saibaba | Optimal sensor placement under model uncertainty in the weak-constraint
4D-Var framework | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | In data assimilation, the model may be subject to uncertainties and errors.
The weak-constraint data assimilation framework enables incorporating model
uncertainty in the dynamics of the governing equations. We propose a new
framework for near-optimal sensor placement in the weak-constrained setting.
This is achieved by first deriving a design criterion based on the expected
information gain, which involves the Kullback-Leibler divergence from the
forecast prior to the posterior distribution. An explicit formula for this
criterion is provided, assuming that the model error and background are
independent and Gaussian and the dynamics are linear. We discuss algorithmic
approaches to efficiently evaluate this criterion through randomized
approximations. To provide further insight and flexibility in computations, we
also provide alternative expressions for the criteria. We provide an algorithm
to find near-optimal experimental designs using column subset selection,
including a randomized algorithm that avoids computing the adjoint of the
forward operator. Through numerical experiments in one and two spatial
dimensions, we show the effectiveness of our proposed methods.
| [
{
"version": "v1",
"created": "Fri, 31 Jan 2025 20:23:39 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 17:24:46 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Alexanderian",
"Alen",
""
],
[
"Díaz",
"Hugo",
""
],
[
"Rao",
"Vishwas",
""
],
[
"Saibaba",
"Arvind K.",
""
]
] |
2502.00559 | Tomasz Gradowski | Tomasz Gradowski and Teodor Buchner | Deep learning model for ECG reconstruction reveals the information
content of ECG leads | 20 pages, 8 figures | null | null | null | eess.SP cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study introduces a deep learning model based on the U-net architecture
to reconstruct missing leads in electrocardiograms (ECGs). The model was
trained to reconstruct 12-lead ECG data from reduced lead configurations using
publicly available datasets. The results highlight the ability of the model to
quantify the information content of each ECG lead and its inter-lead
correlations. This has significant implications for optimizing lead selection
in diagnostic scenarios, particularly in settings where complete 12-lead ECGs
are impractical. In addition, the study provides insights into the
physiological underpinnings of ECG signals and their propagation. The findings
pave the way for advances in telemedicine, portable ECG devices, and
personalized cardiac diagnostics by reducing redundancy and improving signal
interpretation.
| [
{
"version": "v1",
"created": "Sat, 1 Feb 2025 21:06:07 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 20:32:33 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Gradowski",
"Tomasz",
""
],
[
"Buchner",
"Teodor",
""
]
] |
2502.00607 | Shaddin Dughmi | Shaddin Dughmi | PAC Learning is just Bipartite Matching (Sort of) | Position paper | null | null | null | cs.LG cs.DS stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The main goal of this article is to convince you, the reader, that supervised
learning in the Probably Approximately Correct (PAC) model is closely related
to -- of all things -- bipartite matching! En-route from PAC learning to
bipartite matching, I will overview a particular transductive model of
learning, and associated one-inclusion graphs, which can be viewed as a
generalization of some of the hat puzzles that are popular in recreational
mathematics. Whereas this transductive model is far from new, it has recently
seen a resurgence of interest as a tool for tackling deep questions in learning
theory. A secondary purpose of this article could be as a (biased) tutorial on
the connections between the PAC and transductive models of learning.
| [
{
"version": "v1",
"created": "Sun, 2 Feb 2025 00:28:22 GMT"
},
{
"version": "v2",
"created": "Tue, 29 Apr 2025 07:55:46 GMT"
},
{
"version": "v3",
"created": "Mon, 5 May 2025 13:54:44 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Dughmi",
"Shaddin",
""
]
] |
2502.00968 | Anuj Singh | Anuj Singh, Sayak Mukherjee, Ahmad Beirami, Hadi Jamali-Rad | CoDe: Blockwise Control for Denoising Diffusion Models | null | Transactions on Machine Learning Research, 2025. ISSN: 2835-8856 | null | null | cs.CV cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Aligning diffusion models to downstream tasks often requires finetuning new
models or gradient-based guidance at inference time to enable sampling from the
reward-tilted posterior. In this work, we explore a simple inference-time
gradient-free guidance approach, called controlled denoising (CoDe), that
circumvents the need for differentiable guidance functions and model
finetuning. CoDe is a blockwise sampling method applied during intermediate
denoising steps, allowing for alignment with downstream rewards. Our
experiments demonstrate that, despite its simplicity, CoDe offers a favorable
trade-off between reward alignment, prompt instruction following, and inference
cost, achieving a competitive performance against the state-of-the-art
baselines. Our code is available at: https://github.com/anujinho/code.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 00:23:04 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 14:56:26 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Singh",
"Anuj",
""
],
[
"Mukherjee",
"Sayak",
""
],
[
"Beirami",
"Ahmad",
""
],
[
"Jamali-Rad",
"Hadi",
""
]
] |
2502.00985 | Jennifer Choy | Yuhan Tong, Maryam Zahedian, Aiping Zeng, Ricardo Vidrio, Mike
Efremov, Shenwei Yin, Hongyan Mei, Patrick Heaney, Jennifer T. Choy | Tuning the surface energy of fluorinated diamond-like carbon coatings
via plasma immersion ion implantation plasma-enhanced chemical vapor
deposition with 1,1,1,2-tetrafluoroethane | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We demonstrate an environmentally friendly and scalable method to create
fluorine-doped diamond-like carbon (F-DLC) coatings using plasma immersion ion
implantation plasma-enhanced chemical vapor deposition (PIII-PECVD) with
1,1,1,2-tetrafluoroethane. F-DLC films tend to have low wettability and good
mechanical flexibility, which make them suitable for applications in biomedical
devices and antibiofouling surfaces. We report on the effects of fluorine
incorporation on the surface chemistry, surface energy, and morphology of these
coatings, showing that our method is effective in increasing the fluorine
content in the F-DLC up to 40%. We show that the addition of fluorine leads to
a decrease in surface energy, which is consistent with a reduction in surface
wettability.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 01:53:58 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 23:01:24 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Tong",
"Yuhan",
""
],
[
"Zahedian",
"Maryam",
""
],
[
"Zeng",
"Aiping",
""
],
[
"Vidrio",
"Ricardo",
""
],
[
"Efremov",
"Mike",
""
],
[
"Yin",
"Shenwei",
""
],
[
"Mei",
"Hongyan",
""
],
[
"Heaney",
"Patrick",
""
],
[
"Choy",
"Jennifer T.",
""
]
] |
2502.00994 | Hao Quan | Gregory L. Eyink and Hao Quan | Weak-Strong Uniqueness and Extreme Wall Events at High Reynolds Number | null | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | Singular or weak solutions of the incompressible Euler equations have been
hypothesized to account for anomalous dissipation at very high Reynolds numbers
and, in particular, to explain the d'Alembert paradox of non-vanishing drag. A
possible objection to this explanation is the mathematical property called
``weak-strong uniqueness'', which requires that any admissable weak solution of
the Euler equations must coincide with the smooth Euler solution for the same
initial data. As an application of the Josephson-Anderson relation, we sketch a
proof of conditional weak-strong uniqueness for the potential Euler solution of
d'Alembert within the class of strong inviscid limits. We suggest that the mild
conditions required for weak-strong uniqueness are, in fact, physically
violated by violent eruption of very thin boundary layers. We discuss
observational signatures of these extreme events and explain how the small
length-scales involved could threaten the validity of a hydrodynamic
description.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 02:22:24 GMT"
},
{
"version": "v2",
"created": "Sat, 8 Mar 2025 08:24:39 GMT"
},
{
"version": "v3",
"created": "Sun, 4 May 2025 02:38:36 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Eyink",
"Gregory L.",
""
],
[
"Quan",
"Hao",
""
]
] |
2502.01227 | Grzegorz Stachowiak | Leszek G\k{a}sieniec, Tytus Grodzicki, and Grzegorz Stachowiak | Improving Efficiency in Near-State and State-Optimal Self-Stabilising
Leader Election Population Protocols | Accepted to PODC 2025 | null | null | null | cs.DC | http://creativecommons.org/licenses/by/4.0/ | We investigate leader election problem via ranking within self-stabilising
population protocols. In this scenario, the agent's state space comprises $n$
rank states and $x$ extra states. The initial configuration of $n$ agents
consists of arbitrary arrangements of rank and extra states, with the objective
of self-ranking. Specifically, each agent is tasked with stabilising in a
unique rank state silently, implying that after stabilisation, each agent
remains in its designated state indefinitely.
In this paper, we present several new self-stabilising ranking protocols,
greatly enriching our comprehension of these intricate problems. All protocols
ensure self-stabilisation time with high probability (whp), defined as
$1-n^{-\eta},$ for a constant $\eta>0.$ We delve into three scenarios, from
which we derive stable (always correct), either state-optimal or almost
state-optimal, silent ranking protocols that self-stabilise within a time frame
of $o(n^2)$ whp, including:
- Utilising a novel concept of an agent trap, we derive a state-optimal
ranking protocol that achieves self-stabilisation in time
$O(min(kn^{3/2},n^2\log^2 n)),$ for any $k$-distant starting configuration.
- Furthermore, we show that the incorporation of a single extra state ($x=1$)
ensures a ranking protocol that self-stabilises in time $O(n^{7/4}\log^2
n)=o(n^2)$, regardless of the initial configuration.
- Lastly, we show that extra $x=O(\log n)$ states admit self-stabilising
ranking with the best currently known stabilisation time $O(n\log n)$, when whp
and $x=O(\log n)$ guarantees are imposed.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 10:30:01 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 07:27:58 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Gąsieniec",
"Leszek",
""
],
[
"Grodzicki",
"Tytus",
""
],
[
"Stachowiak",
"Grzegorz",
""
]
] |
2502.01563 | Mingyu Jin | Mingyu Jin, Kai Mei, Wujiang Xu, Mingjie Sun, Ruixiang Tang, Mengnan
Du, Zirui Liu, Yongfeng Zhang | Massive Values in Self-Attention Modules are the Key to Contextual
Knowledge Understanding | International Conference on Machine Learning (ICML 2025) | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large language models (LLMs) have achieved remarkable success in contextual
knowledge understanding. In this paper, we show that these concentrated massive
values consistently emerge in specific regions of attention queries (Q) and
keys (K) while not having such patterns in values (V) in various modern
transformer-based LLMs (Q, K, and V mean the representations output by the
query, key, and value layers respectively). Through extensive experiments, we
further demonstrate that these massive values play a critical role in
interpreting contextual knowledge (knowledge obtained from the current context
window) rather than in retrieving parametric knowledge stored within the
model's parameters. Our further investigation of quantization strategies
reveals that ignoring these massive values leads to a pronounced drop in
performance on tasks requiring rich contextual understanding, aligning with our
analysis. Finally, we trace the emergence of concentrated massive values and
find that such concentration is caused by Rotary Positional Encoding (RoPE),
which has appeared since the first layers. These findings shed new light on how
Q and K operate in LLMs and offer practical insights for model design and
optimization. The Code is Available at
https://github.com/MingyuJ666/Rope_with_LLM.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 17:47:03 GMT"
},
{
"version": "v2",
"created": "Mon, 28 Apr 2025 18:40:27 GMT"
},
{
"version": "v3",
"created": "Sat, 3 May 2025 01:40:50 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Jin",
"Mingyu",
""
],
[
"Mei",
"Kai",
""
],
[
"Xu",
"Wujiang",
""
],
[
"Sun",
"Mingjie",
""
],
[
"Tang",
"Ruixiang",
""
],
[
"Du",
"Mengnan",
""
],
[
"Liu",
"Zirui",
""
],
[
"Zhang",
"Yongfeng",
""
]
] |
2502.01710 | Shilong Hong | Shilong Hong, Yanzhou Zhou and Weichao Xu | DAGNet: A Dual-View Attention-Guided Network for Efficient X-ray
Security Inspection | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | With the rapid development of modern transportation systems and the
exponential growth of logistics volumes, intelligent X-ray-based security
inspection systems play a crucial role in public safety. Although single-view
X-ray baggage scanner is widely deployed, they struggles to accurately identify
contraband in complex stacking scenarios due to strong viewpoint dependency and
inadequate feature representation. To address this, we propose a Dual-View
Attention-Guided Network for Efficient X-ray Security Inspection (DAGNet). This
study builds on a shared-weight backbone network as the foundation and
constructs three key modules that work together: (1) Frequency Domain
Interaction Module (FDIM) dynamically enhances features by adjusting frequency
components based on inter-view relationships; (2) Dual-View Hierarchical
Enhancement Module (DVHEM) employs cross-attention to align features between
views and capture hierarchical associations; (3) Convolutional Guided Fusion
Module (CGFM) fuses features to suppress redundancy while retaining critical
discriminative information. Collectively, these modules substantially improve
the performance of dual-view X-ray security inspection. Experimental results
demonstrate that DAGNet outperforms existing state-of-the-art approaches across
multiple backbone architectures. The code is available
at:https://github.com/ShilongHong/DAGNet.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 15:18:54 GMT"
},
{
"version": "v2",
"created": "Fri, 7 Feb 2025 12:20:59 GMT"
},
{
"version": "v3",
"created": "Tue, 8 Apr 2025 14:20:25 GMT"
},
{
"version": "v4",
"created": "Mon, 5 May 2025 08:31:11 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Hong",
"Shilong",
""
],
[
"Zhou",
"Yanzhou",
""
],
[
"Xu",
"Weichao",
""
]
] |
2502.01713 | Kirtan Padh | Floris Holstege, Mackenzie Jorgensen, Kirtan Padh, Jurriaan Parie,
Joel Persson, Krsto Prorokovic, Lukas Snoek | Auditing a Dutch Public Sector Risk Profiling Algorithm Using an
Unsupervised Bias Detection Tool | null | null | null | null | cs.CY cs.LG | http://creativecommons.org/licenses/by/4.0/ | Algorithms are increasingly used to automate or aid human decisions, yet
recent research shows that these algorithms may exhibit bias across legally
protected demographic groups. However, data on these groups may be unavailable
to organizations or external auditors due to privacy legislation. This paper
studies bias detection using an unsupervised clustering tool when data on
demographic groups are unavailable. We collaborate with the Dutch Executive
Agency for Education to audit an algorithm that was used to assign risk scores
to college students at the national level in the Netherlands between 2012-2023.
Our audit covers more than 250,000 students from the whole country. The
unsupervised clustering tool highlights known disparities between students with
a non-European migration background and Dutch origin. Our contributions are
three-fold: (1) we assess bias in a real-world, large-scale and high-stakes
decision-making process by a governmental organization; (2) we use simulation
studies to highlight potential pitfalls of using the unsupervised clustering
tool to detect true bias when demographic group data are unavailable and
provide recommendations for valid inferences; (3) we provide the unsupervised
clustering tool in an open-source library. Our work serves as a starting point
for a deliberative assessment by human experts to evaluate potential
discrimination in algorithmic-supported decision-making processes.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 15:58:42 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 08:52:08 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Holstege",
"Floris",
""
],
[
"Jorgensen",
"Mackenzie",
""
],
[
"Padh",
"Kirtan",
""
],
[
"Parie",
"Jurriaan",
""
],
[
"Persson",
"Joel",
""
],
[
"Prorokovic",
"Krsto",
""
],
[
"Snoek",
"Lukas",
""
]
] |
2502.01761 | Stephan Weis | Stephan Weis | Matrix systems, algebras, and open maps | 27 pages, version 2: corrections in Section 9 "Continuity in quantum
information theory" (among others, reference K. Vanslette, Entropy 19:12,
664, 2017 added), comments welcomed | null | null | null | math.FA quant-ph | http://creativecommons.org/licenses/by/4.0/ | Every state on the algebra $M_n$ of complex nxn matrices restricts to a state
on any matrix system. Whereas the restriction to a matrix system is generally
not open, we prove that the restriction to every *-subalgebra of $M_n$ is open.
This simplifies topology problems in matrix theory and quantum information
theory.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 19:05:12 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 22:02:59 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Weis",
"Stephan",
""
]
] |
2502.01860 | Zhimin Zhao | Zhimin Zhao | SE Arena: An Interactive Platform for Evaluating Foundation Models in
Software Engineering | Check the arena at
https://huggingface.co/spaces/SE-Arena/Software-Engineering-Arena | null | null | null | cs.SE cs.LG | http://creativecommons.org/licenses/by/4.0/ | Foundation models (FMs), particularly large language models (LLMs), have
shown significant promise in various software engineering (SE) tasks, including
code generation, debugging, and requirement refinement. Despite these advances,
existing evaluation frameworks are insufficient for assessing model performance
in iterative, context-rich workflows characteristic of SE activities. To
address this limitation, we introduce SE Arena, an interactive platform
designed to evaluate FMs in SE tasks. SE Arena provides a transparent,
open-source leaderboard, supports multi-round conversational workflows, and
enables end-to-end model comparisons. The platform introduces novel metrics,
including model consistency score that measures the consistency of model
outputs through self-play matches, and conversation efficiency index that
evaluates model performance while accounting for the number of interaction
rounds required to reach conclusions. Moreover, SE Arena incorporates a new
feature called RepoChat, which automatically injects repository-related context
(e.g., issues, commits, pull requests) into the conversation, further aligning
evaluations with real-world development processes. This paper outlines the
design and capabilities of SE Arena, emphasizing its potential to advance the
evaluation and practical application of FMs in software engineering.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 22:19:28 GMT"
},
{
"version": "v2",
"created": "Mon, 31 Mar 2025 19:39:16 GMT"
},
{
"version": "v3",
"created": "Tue, 15 Apr 2025 05:34:27 GMT"
},
{
"version": "v4",
"created": "Sun, 4 May 2025 10:09:30 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Zhao",
"Zhimin",
""
]
] |
2502.01872 | Antonio Francesco Mello | Antonio Francesco Mello, Alessandro Santini, Mario Collura | Clifford-Dressed Variational Principles for Precise Loschmidt Echoes | 7 pages, 5 figures | Phys. Rev. A 111, 052401 (2025) | 10.1103/PhysRevA.111.052401 | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | We extend the recently introduced Clifford dressed Time-Dependent Variational
Principle (TDVP) to efficiently compute many-body wavefunction amplitudes in
the computational basis. This advancement enhances the study of Loschmidt
echoes, which generally require accurate calculations of the overlap between
the evolved state and the initial wavefunction. By incorporating Clifford
disentangling gates during TDVP evolution, our method effectively controls
entanglement growth while keeping the computation of these amplitudes
accessible. Specifically, it reduces the problem to evaluating the overlap
between a Matrix Product State (MPS) and a stabilizer state, a task that
remains computationally feasible within the proposed framework. To demonstrate
the effectiveness of this approach, we first benchmark it on the
one-dimensional transverse-field Ising model. We then apply it to more
challenging scenarios, including a non-integrable next-to-nearest-neighbor
Ising chain and a two-dimensional Ising model. Our results highlight the
versatility and efficiency of the Clifford-augmented MPS, showcasing its
capability to go beyond the evaluation of simple expectation values. This makes
it a powerful tool for exploring various aspects of many-body quantum dynamics.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 22:43:32 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Mello",
"Antonio Francesco",
""
],
[
"Santini",
"Alessandro",
""
],
[
"Collura",
"Mario",
""
]
] |
2502.01892 | Alex Stivala | Alex Stivala, Peng Wang, Alessandro Lomi | Improving exponential-family random graph models for bipartite networks | Revision after reviewer comments. This shorter manuscript is the
mostly theoretical part (only one small empirical example) | null | null | null | stat.ME stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Bipartite graphs, representing two-mode networks, arise in many research
fields. These networks have two disjoint node sets representing distinct entity
types, for example persons and groups, with edges representing associations
between the two entity types. In bipartite graphs, the smallest possible cycle
is a cycle of length four, and hence four-cycles are the smallest structure to
model closure in such networks. Exponential-family random graph models (ERGMs)
are a widely used model for social, and other, networks, including specifically
bipartite networks. Existing ERGM terms to model four-cycles in bipartite
networks, however, are relatively rarely used. In this work we demonstrate some
problems with these existing terms to model four-cycles, and define new ERGM
terms to help overcome these problems. The position of the new terms in the
ERGM dependence hierarchy, and their interpretation, is discussed. The new
terms are demonstrated in simulation experiments, and their application
illustrated on a canonical example of an empirical two-mode network.
| [
{
"version": "v1",
"created": "Mon, 3 Feb 2025 23:58:23 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 23:21:12 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Stivala",
"Alex",
""
],
[
"Wang",
"Peng",
""
],
[
"Lomi",
"Alessandro",
""
]
] |
2502.02031 | Yizhuang Liu | Yizhuang Liu | Renormalon-like factorial enhancements to power expansion/OPE in a
super-renormalizable 2D $O(N)$ quartic model | 74 pages, 10 figures. Title changed; Introduction and conclusion
expanded; Acknowledgements added; Typos corrected; Appendices C and D added | null | null | null | hep-th hep-ph math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we investigate the effects of logarithms on the asymptotic
behavior of power expansion/OPE in supper-renormalizable QFTs. We performed a
careful investigation of the large $p^2$ expansion of a scalar-scalar two-point
function at the next-to-leading order in the large-$N$ expansion, in a
large-$N$ $O(N)$ quartic model that is populated by logarithms. We show that
because the large-$p^2$ logarithms of the individual bubbles can be amplified
by bubble-chains, there are factorial enhancements to the power expansion. We
show how the factorial enhancements appear separately in the coefficient
functions and operator condensates, and demonstrate how they are cancelled
off-diagonally across different powers. Restricted to any given power, the
factorial enhancements are no-longer canceled. The large-$p^2$ power expansion
is divergent.
| [
{
"version": "v1",
"created": "Tue, 4 Feb 2025 05:40:11 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Feb 2025 14:11:26 GMT"
},
{
"version": "v3",
"created": "Wed, 19 Feb 2025 17:55:37 GMT"
},
{
"version": "v4",
"created": "Mon, 24 Feb 2025 16:13:20 GMT"
},
{
"version": "v5",
"created": "Mon, 5 May 2025 11:02:40 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Liu",
"Yizhuang",
""
]
] |
2502.02217 | Naoto Yokoyama | Daigaku Katsumi, Masanobu Inubushi, Naoto Yokoyama | Data-driven prediction of reversal of large-scale circulation in
turbulent convection | 15 pages, 13 figures | Phys. Rev. Fluids, 10 (2025) 053501 | 10.1103/PhysRevFluids.10.053501 | null | physics.flu-dyn nlin.CD | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large-scale circulation (LSC) quasi-stably emerges in the turbulent
Rayleigh-B\'{e}nard convection, and intermittently reverses its rotational
direction in two-dimensional turbulent convection. In this paper, direct
numerical simulations of the intermittent reversals of the LSC in a
two-dimensional square domain are performed, and the time series of the total
angular momentum indicating the rotational direction of the LSC is predicted by
reservoir computing whose input consists of the shear rates and temperatures at
six locations on the sidewalls. The total angular momentum in the simulation
after times shorter than half the typical duration of the quasi-stable states
is successfully reproduced by the locally-measurable quantities on the
sidewalls because the secondary rolls accompanied by the boundary flow
characterize the reversal of the LSC. The successful prediction by such sparse
input derived from local measurements on the sidewalls demonstrates that the
reservoir computing prediction of the reversal is feasible in laboratory
experiments and industrial applications. On the other hand, long-term
prediction often provides the total angular momentum opposite in sign to the
one in the simulations in the late parts of long quasi-stable states. The
statistical independence of each reversal implies that the prediction after the
reversal is difficult or even impossible, and the training data in the late
part in the long quasi-stable state, which rarely appears, is contaminated by
the statistically-independent angular momentum in the subsequent quasi-stable
state.
| [
{
"version": "v1",
"created": "Tue, 4 Feb 2025 10:52:18 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Mar 2025 01:54:22 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Katsumi",
"Daigaku",
""
],
[
"Inubushi",
"Masanobu",
""
],
[
"Yokoyama",
"Naoto",
""
]
] |
2502.02240 | Martin Pavelka | Martin Pavelka, Simon Marotzke, Ru-Pan Wang, Mohamed F. Elhanoty,
G\"unter Brenner, Siarhei Dziarzhytski, Somnath Jana, W. Dieter Engel,
Clemens v. Korff Schmising, Deeksha Gupta, Igor Vaskivskyi, Tim Amrhein, Nele
Thielemann-K\"uhn, Martin Weinelt, Ronny Knut, Juliane R\"onsch-Schulenberg,
Evgeny Schneidmiller, Christian Sch\"u{\ss}ler-Langeheine, Martin Beye, Niko
Pontius, Oscar Gr{\aa}n\"as, Hermann A. D\"urr | Femtosecond charge and spin dynamics in CoPt alloys | 17 pages, 3 figures | Struct. Dyn. 12, 024303 (2025) | 10.1063/4.0000297 | null | cond-mat.mtrl-sci cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | The use of advanced X-ray sources plays a key role in the study of dynamic
processes in magnetically ordered materials. The progress in X-ray free
electron lasers enables the direct and simultaneous observation of the
femtosecond evolution of electron and spin systems through transient X-ray
absorption spectroscopy (XAS) and X-ray magnetic circular dichroism (XMCD),
respectively. Such experiments allow us to resolve the response seen in the
population of the spin-split valence states upon optical excitation. Here, we
utilize circularly polarized ultrashort soft X-ray pulses from the new helical
afterburner undulator at the free-electron laser FLASH in Hamburg to study the
femtosecond dynamics of a laser-excited CoPt alloy at the Co $L_{3}$ absorption
edge. Despite employing a weaker electronic excitation level we find a
comparable demagnetization for the Co $3d$-states in CoPt compared to previous
measurements on CoPd. This is attributed to distinctly different orbital
hybridization and spin-orbit coupling between $3d$ and $4d$ vs. $3d$ and $5d$
elements in the corresponding alloys and multilayers.
| [
{
"version": "v1",
"created": "Tue, 4 Feb 2025 11:30:49 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Pavelka",
"Martin",
""
],
[
"Marotzke",
"Simon",
""
],
[
"Wang",
"Ru-Pan",
""
],
[
"Elhanoty",
"Mohamed F.",
""
],
[
"Brenner",
"Günter",
""
],
[
"Dziarzhytski",
"Siarhei",
""
],
[
"Jana",
"Somnath",
""
],
[
"Engel",
"W. Dieter",
""
],
[
"Schmising",
"Clemens v. Korff",
""
],
[
"Gupta",
"Deeksha",
""
],
[
"Vaskivskyi",
"Igor",
""
],
[
"Amrhein",
"Tim",
""
],
[
"Thielemann-Kühn",
"Nele",
""
],
[
"Weinelt",
"Martin",
""
],
[
"Knut",
"Ronny",
""
],
[
"Rönsch-Schulenberg",
"Juliane",
""
],
[
"Schneidmiller",
"Evgeny",
""
],
[
"Schüßler-Langeheine",
"Christian",
""
],
[
"Beye",
"Martin",
""
],
[
"Pontius",
"Niko",
""
],
[
"Grånäs",
"Oscar",
""
],
[
"Dürr",
"Hermann A.",
""
]
] |
2502.02320 | Mose Mizrahi Erbes | Mose Mizrahi Erbes and Roger Wattenhofer | Extending Asynchronous Byzantine Agreement with Crusader Agreement | 21 pages, 0 figures, full version of a PODC 2025 brief announcement | null | null | null | cs.DC cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we study multivalued byzantine agreement (BA) in an
asynchronous network of $n$ parties where up to $t < \frac{n}{3}$ parties are
byzantine. We present a new reduction from multivalued BA to binary BA. It
allows one to achieve BA on $\ell$-bit inputs with one instance of binary BA,
one instance of crusader agreement (CA) on $\ell$-bit inputs and $\Theta(\ell n
+ n^2)$ bits of additional communication.
As our reduction uses multivalued CA, we also design two new
information-theoretic CA protocols for $\ell$-bit inputs. In the first one, we
use almost-universal hashing to achieve statistical security with probability
$1 - 2^{-\lambda}$ against $t < \frac{n}{3}$ faults with $\Theta(\ell n +
n^2(\lambda + \log n))$ bits of communication. Following this, we replace the
hashes with error correcting code symbols and add a preliminary step based on
the synchronous multivalued BA protocol COOL [DISC '21] to obtain a second,
perfectly secure CA protocol that can for any $\varepsilon > 0$ be set to
tolerate $t \leq \frac{n}{3 + \varepsilon}$ faults with
$\mathcal{O}\bigl(\frac{\ell n}{\min(1, \varepsilon^2)} + n^2\max\bigl(1, \log
\frac{1}{\varepsilon}\bigr) \bigr)$ bits of communication. Our CA protocols
allow one to extend binary BA to multivalued BA with a constant round overhead,
a quadratic-in-$n$ communication overhead, and information-theoretic security.
| [
{
"version": "v1",
"created": "Tue, 4 Feb 2025 13:44:41 GMT"
},
{
"version": "v2",
"created": "Mon, 10 Feb 2025 14:06:17 GMT"
},
{
"version": "v3",
"created": "Mon, 5 May 2025 07:56:44 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Erbes",
"Mose Mizrahi",
""
],
[
"Wattenhofer",
"Roger",
""
]
] |
2502.02811 | Sergei Khlebnikov | Sergei Khlebnikov and Maxim Lyutikov (Purdue University) | Cherenkov emission by a fast-moving uncharged Schwarzschild black hole | null | null | null | null | hep-th astro-ph.HE gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We demonstrate that, in the presence of an external magnetic field, an
uncharged classical Schwarzschild black hole moving superluminally in a
dielectric with permittivity $\epsilon > 1$ produces Cherenkov emission. This
is a new physical effect: classical (non-quantum) emission of electromagnetic
waves by a completely charge-neutral ``particle.'' The governing equations
(involving general relativity, electromagnetism, and the physics of continuous
media) have no external electromagnetic source -- it is the distortion of the
initial electromagnetic fields by the gravity of the black hole that plays the
role of a superluminally moving source. The effect relies on nonzero values of
both the magnetic field and the gravitational radius, as well as on the usual
Cherenkov condition on the velocity, $v/c > 1/\sqrt{\epsilon}$. Unlike
Cherenkov emission by a point charge, the effective source in this case is
spatially distributed, with emission generated along the single Cherenkov
emission cone. The emitted spectrum is red-dominated, with power $\propto dk_z
/|k_z|$ for wave numbers $|k_z| \leq 1/R_G$, where $R_G$ is the Schwarzschild
radius. We comment on possible observability of this process during black hole
-- neutron star mergers.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 01:17:32 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Feb 2025 20:29:46 GMT"
},
{
"version": "v3",
"created": "Mon, 24 Mar 2025 16:23:15 GMT"
},
{
"version": "v4",
"created": "Mon, 5 May 2025 17:36:24 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Khlebnikov",
"Sergei",
"",
"Purdue University"
],
[
"Lyutikov",
"Maxim",
"",
"Purdue University"
]
] |
2502.02829 | Shucheng Kang | Shucheng Kang, Guorui Liu, Heng Yang | Global Contact-Rich Planning with Sparsity-Rich Semidefinite Relaxations | Website: https://computationalrobotics.seas.harvard.edu/project-spot/ | null | null | null | cs.RO math.OC | http://creativecommons.org/licenses/by/4.0/ | We show that contact-rich motion planning is also sparsity-rich when viewed
as polynomial optimization (POP). We can exploit not only the correlative and
term sparsity patterns that are general to all POPs, but also specialized
sparsity patterns from the robot kinematic structure and the separability of
contact modes. Such sparsity enables the design of high-order but sparse
semidefinite programming (SDPs) relaxations--building upon Lasserre's moment
and sums of squares hierarchy--that (i) can be solved in seconds by
off-the-shelf SDP solvers, and (ii) compute near globally optimal solutions to
the nonconvex contact-rich planning problems with small certified
suboptimality. Through extensive experiments both in simulation (Push Bot, Push
Box, Push Box with Obstacles, and Planar Hand) and real world (Push T), we
demonstrate the power of using convex SDP relaxations to generate global
contact-rich motion plans. As a contribution of independent interest, we
release the Sparse Polynomial Optimization Toolbox (SPOT)--implemented in C++
with interfaces to both Python and Matlab--that automates sparsity exploitation
for robotics and beyond.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 02:13:51 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Feb 2025 15:22:42 GMT"
},
{
"version": "v3",
"created": "Sun, 4 May 2025 17:12:25 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Kang",
"Shucheng",
""
],
[
"Liu",
"Guorui",
""
],
[
"Yang",
"Heng",
""
]
] |
2502.03253 | Antonio Laverghetta Jr. | Antonio Laverghetta Jr., Tuhin Chakrabarty, Tom Hope, Jimmy Pronchick,
Krupa Bhawsar, Roger E. Beaty | How do Humans and Language Models Reason About Creativity? A Comparative
Analysis | CogSci 2025 | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Creativity assessment in science and engineering is increasingly based on
both human and AI judgment, but the cognitive processes and biases behind these
evaluations remain poorly understood. We conducted two experiments examining
how including example solutions with ratings impact creativity evaluation,
using a finegrained annotation protocol where raters were tasked with
explaining their originality scores and rating for the facets of remoteness
(whether the response is "far" from everyday ideas), uncommonness (whether the
response is rare), and cleverness. In Study 1, we analyzed creativity ratings
from 72 experts with formal science or engineering training, comparing those
who received example solutions with ratings (example) to those who did not (no
example). Computational text analysis revealed that, compared to experts with
examples, no-example experts used more comparative language (e.g.,
"better/worse") and emphasized solution uncommonness, suggesting they may have
relied more on memory retrieval for comparisons. In Study 2, parallel analyses
with state-of-the-art LLMs revealed that models prioritized uncommonness and
remoteness of ideas when rating originality, suggesting an evaluative process
rooted around the semantic similarity of ideas. In the example condition, while
LLM accuracy in predicting the true originality scores improved, the
correlations of remoteness, uncommonness, and cleverness with originality also
increased substantially -- to upwards of $0.99$ -- suggesting a homogenization
in the LLMs evaluation of the individual facets. These findings highlight
important implications for how humans and AI reason about creativity and
suggest diverging preferences for what different populations prioritize when
rating.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 15:08:43 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 13:47:32 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Laverghetta",
"Antonio",
"Jr."
],
[
"Chakrabarty",
"Tuhin",
""
],
[
"Hope",
"Tom",
""
],
[
"Pronchick",
"Jimmy",
""
],
[
"Bhawsar",
"Krupa",
""
],
[
"Beaty",
"Roger E.",
""
]
] |
2502.03270 | Nikolaos Tsagkas | Nikolaos Tsagkas, Andreas Sochopoulos, Duolikun Danier, Sethu
Vijayakumar, Chris Xiaoxuan Lu, Oisin Mac Aodha | When Pre-trained Visual Representations Fall Short: Limitations in
Visuo-Motor Robot Learning | Project Page: https://tsagkas.github.io/pvrobo/ | null | null | null | cs.RO cs.AI cs.CV cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | The integration of pre-trained visual representations (PVRs) into visuo-motor
robot learning has emerged as a promising alternative to training visual
encoders from scratch. However, PVRs face critical challenges in the context of
policy learning, including temporal entanglement and an inability to generalise
even in the presence of minor scene perturbations. These limitations hinder
performance in tasks requiring temporal awareness and robustness to scene
changes. This work identifies these shortcomings and proposes solutions to
address them. First, we augment PVR features with temporal perception and a
sense of task completion, effectively disentangling them in time. Second, we
introduce a module that learns to selectively attend to task-relevant local
features, enhancing robustness when evaluated on out-of-distribution scenes.
Our experiments demonstrate significant performance improvements, particularly
in PVRs trained with masking objectives, and validate the effectiveness of our
enhancements in addressing PVR-specific limitations.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 15:25:46 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 09:42:27 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Tsagkas",
"Nikolaos",
""
],
[
"Sochopoulos",
"Andreas",
""
],
[
"Danier",
"Duolikun",
""
],
[
"Vijayakumar",
"Sethu",
""
],
[
"Lu",
"Chris Xiaoxuan",
""
],
[
"Mac Aodha",
"Oisin",
""
]
] |
2502.03361 | Tobias Eul | Tobias Eul, Miwan Sabir, Victor DeManuel-Gonzalez, Florian Diekmann,
Kai Rossnagel, Michael Bauer | Photoemission electron microscopy of exciton-polaritons in thin WSe$_2$
waveguides | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Exciton-polaritons emerging from the interaction of photons and excitons in
the strong coupling regime are intriguing quasiparticles for the potential
exchange of energy during light-matter interaction processes such as light
harvesting. The coupling causes an energy anti-crossing in the photon
dispersion centered around the exciton resonance, i.e., a Rabi splitting
between a lower and upper energetic branch. The size of this splitting
correlates with the coupling strength between the exciton and the photonic
modes. In this work, we investigate this coupling between excitons and photonic
waveguide modes excited simultaneously in thin-film flakes of the
transition-metal dichalcogenide WSe$_2$. Using a Photoemission electron
microscope, we are able to extract the dispersion of the transversal electric
and magnetic modes propagating through these flakes as well as extract the
energy splitting. Ultimately, our findings provide a basis for the
investigation of the propagation of exciton-polaritons in the time-domain via
time-resolved photoemission.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 16:57:17 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Feb 2025 09:42:21 GMT"
},
{
"version": "v3",
"created": "Mon, 5 May 2025 16:16:02 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Eul",
"Tobias",
""
],
[
"Sabir",
"Miwan",
""
],
[
"DeManuel-Gonzalez",
"Victor",
""
],
[
"Diekmann",
"Florian",
""
],
[
"Rossnagel",
"Kai",
""
],
[
"Bauer",
"Michael",
""
]
] |
2502.03377 | Abdullahi Isa Ahmed | Abdullahi Isa Ahmed, Jamal Bentahar and El Mehdi Amhoud | Energy-Efficient Flying LoRa Gateways: A Multi-Agent Reinforcement
Learning Approach | 6 pages, 8 figures, 1 table | null | null | null | cs.NI cs.LG | http://creativecommons.org/licenses/by/4.0/ | As next-generation Internet of Things (NG-IoT) networks continue to grow, the
number of connected devices is rapidly increasing, along with their energy
demands. This creates challenges for resource management and sustainability.
Energy-efficient communication, particularly for power-limited IoT devices, is
therefore a key research focus. In this paper, we deployed flying LoRa gateways
mounted on unmanned aerial vehicles (UAVs) to collect data from LoRa end
devices and transmit it to a central server. Our primary objective is to
maximize the global system energy efficiency of wireless LoRa networks by joint
optimization of transmission power, spreading factor, bandwidth, and user
association. To solve this challenging problem, we model the problem as a
partially observable Markov decision process (POMDP), where each flying LoRa GW
acts as a learning agent using a cooperative multi-agent reinforcement learning
(MARL). Simulation results demonstrate that our proposed method, based on the
multi-agent proximal policy optimization algorithm, significantly improves the
global system energy efficiency and surpasses the popular MARL and other
conventional schemes.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 17:16:40 GMT"
},
{
"version": "v2",
"created": "Sun, 13 Apr 2025 17:52:57 GMT"
},
{
"version": "v3",
"created": "Mon, 5 May 2025 13:46:29 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Ahmed",
"Abdullahi Isa",
""
],
[
"Bentahar",
"Jamal",
""
],
[
"Amhoud",
"El Mehdi",
""
]
] |
2502.03662 | The-Anh Vu-Le | The-Anh Vu-Le, Lahari Anne, George Chacko, Tandy Warnow | EC-SBM Synthetic Network Generator | null | null | 10.1007/s41109-025-00701-2 | null | cs.SI | http://creativecommons.org/licenses/by/4.0/ | Generating high-quality synthetic networks with realistic community structure
is vital to effectively evaluate community detection algorithms. In this study,
we propose a new synthetic network generator called the Edge-Connected
Stochastic Block Model (EC-SBM). The goal of EC-SBM is to take a given
clustered real-world network and produce a synthetic network that resembles the
clustered real-world network with respect to both network and
community-specific criteria. In particular, we focus on simulating the internal
edge connectivity of the clusters in the reference clustered network. Our
extensive performance study on large real-world networks shows that EC-SBM has
high accuracy in both network and community-specific criteria, and is generally
more accurate than current alternative approaches for this problem.
Furthermore, EC-SBM is fast enough to scale to real-world networks with
millions of nodes.
| [
{
"version": "v1",
"created": "Wed, 5 Feb 2025 23:06:16 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Vu-Le",
"The-Anh",
""
],
[
"Anne",
"Lahari",
""
],
[
"Chacko",
"George",
""
],
[
"Warnow",
"Tandy",
""
]
] |
2502.04088 | Torsten Ensslin | Torsten En{\ss}lin | Quantifying imperfect cognition via achieved information gain | 21 pages, 7 figures, accepted for publication by Annalen der Physik | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cognition, information processing in form of inference, communication, and
memorization, is the central activity of any intelligence. Its physical
realization in a brain, computer, or in any other intelligent system requires
resources like time, energy, memory, bandwidth, money, and others. Due to
limited resources, many real world intelligent systems perform only imperfect
cognition. To understand the trade-off between accuracy and resource
investments in existing systems, e.g. in biology, as well as for the
resource-aware optimal design of information processing systems, like computer
algorithms and artificial neural networks, a quantification of information
obtained in an imperfect cognitive operation is desirable. To this end, we
propose the concept of the achieved information gain (AIG) of a belief update,
which is given by the amount of information obtained by updating from the
initial state of knowledge to the ideal state, minus the amount that a change
from the imperfect to the ideal state would yield. AIG has many desirable
properties for quantifying imperfect cognition. The ratio of achieved to
ideally obtainable information measures cognitive fidelity and that of AIG to
the necessary cognitive effort measures cognitive efficiency. We provide an
axiomatic derivation of AIG, relate it to other information measures,
illustrate its application to common scenarios of posterior inaccuracies, and
discuss the implication of cognitive efficiency for sustainable resource
allocation in computational inference.
| [
{
"version": "v1",
"created": "Thu, 6 Feb 2025 13:57:19 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 12:01:42 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Enßlin",
"Torsten",
""
]
] |
2502.04667 | Xinhao Yao | Xinhao Yao, Ruifeng Ren, Yun Liao, Yong Liu | Unveiling the Mechanisms of Explicit CoT Training: How CoT Enhances
Reasoning Generalization | null | null | null | null | cs.LG cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | The integration of explicit Chain-of-Thought (CoT) reasoning into training
large language models (LLMs) has advanced their reasoning capabilities, yet the
mechanisms by which CoT enhances generalization remain poorly understood. This
work investigates (1) \textit{how} CoT training reshapes internal model
representations and (2) \textit{why} it improves both in-distribution (ID) and
out-of-distribution (OOD) reasoning generalization. Through controlled
experiments and theoretical analysis, we derive the following key insights.
\textbf{1)} Structural Advantage: CoT training internalizes reasoning into a
two-stage generalizing circuit, where the number of stages corresponds to the
explicit reasoning steps during training. Notably, CoT-trained models resolve
intermediate results at shallower layers compared to non-CoT counterparts,
freeing up deeper layers to specialize in subsequent reasoning steps.
\textbf{2)} Theoretical Analysis: the information-theoretic generalization
bounds via distributional divergence can be decomposed into ID and OOD
components. While ID error diminishes with sufficient training regardless of
CoT, OOD error critically depends on CoT: Non-CoT training fails to generalize
to OOD samples due to unseen reasoning patterns, whereas CoT training achieves
near-perfect OOD generalization by mastering subtasks and reasoning
compositions during training. The identified mechanisms explain our
experimental results: CoT training accelerates convergence and enhances
generalization from ID to both ID and OOD scenarios while maintaining robust
performance even with tolerable noise. These findings are further validated on
complex real-world datasets. This paper offers valuable insights for designing
CoT strategies to enhance LLM reasoning robustness.
| [
{
"version": "v1",
"created": "Fri, 7 Feb 2025 05:21:13 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 09:01:06 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Yao",
"Xinhao",
""
],
[
"Ren",
"Ruifeng",
""
],
[
"Liao",
"Yun",
""
],
[
"Liu",
"Yong",
""
]
] |
2502.05166 | Piotr Habdas | Piotr Habdas, Eric R. Weeks | Stirring supercooled colloidal liquids at the particle scale | 9 pages, 7 figures | null | null | null | cond-mat.soft | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We study the decay of tangential velocity profiles with distance from a local
disturbance in hard-sphere colloidal suspensions as the colloidal glass
transition is approached. The disturbance, generated by a dimer of
superparamagnetic particles rotated by an external magnetic field, enables a
precise characterization of the system's response through confocal microscopy
and tracking of individual particle dynamics. The tangential velocity profiles
exhibit nearly exponential decay with distance. As particle density increases
toward the colloidal glass transition, the characteristic length scale derived
from exponential fits grows. We also observe that the colloidal particles slip
against the rotating dimer, with less slip in samples which are closer to the
glass transition.
| [
{
"version": "v1",
"created": "Fri, 7 Feb 2025 18:49:07 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 14:42:33 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Habdas",
"Piotr",
""
],
[
"Weeks",
"Eric R.",
""
]
] |
2502.05483 | Hideki Kawahara | Hideki Kawahara | Numerical Approximation of Delay Differential Equations via Operator
Splitting in Fractional Domains | submitted for publication | null | null | null | math.OC math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper develops a rigorous framework for the numerical approximation of
both autonomous and non-autonomous delay differential equations (DDEs), with a
focus on the implicit Euler method and sequential operator splitting.
To overcome the difficulty that the delay operator does not generate an
analytic semigroup in the standard space \( L^1[\tau, 0] \), we embed the
problem into the interpolation space \( \left(L^1[\tau, 0], W^{1,1}_0[\tau,
0]\right)_{\theta, 1} \) for \( 0 < \theta < 1 \), where the differential
operator becomes sectorial. This allows the full operator \( L = A + B \) to
generate an analytic semigroup \( T_L(t) \), enabling the use of semigroup
theory to derive sharp error estimates.
We prove that the implicit Euler method achieves a global error of order \(
\mathcal{O}(h) \), while the Lie--Trotter splitting method yields an error of
order \( \mathcal{O}(h^{2\theta - 1}) \) in the interpolation norm. These
theoretical rates are confirmed by numerical experiments, including comparisons
with exact solutions obtained via semi-analytical Fourier-based methods in the
non-autonomous setting.
| [
{
"version": "v1",
"created": "Sat, 8 Feb 2025 07:44:20 GMT"
},
{
"version": "v2",
"created": "Sat, 22 Mar 2025 15:41:31 GMT"
},
{
"version": "v3",
"created": "Fri, 2 May 2025 13:41:58 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Kawahara",
"Hideki",
""
]
] |
2502.05581 | Elena V. Arbuzova | E.V. Arbuzova, A.D. Dolgov | Dynamical mechanism of vacuum energy compensation | 10 pages, 2 figures. Several references and short explanatory
comments are added | null | null | null | gr-qc astro-ph.CO | http://creativecommons.org/licenses/by/4.0/ | The model of vacuum energy compensation due to interaction of a scalar field
$\phi$ with curvature scalar of the form $\beta R \phi^2 f(\phi) $ is proposed.
It is shown that with a simple power form of $f(\phi)$ the exponential
expansion, induced by vacuum energy (or what is the same by cosmological
constant), is transformed into canonical cosmological evolution of the universe
dominated by relativistic matter.
| [
{
"version": "v1",
"created": "Sat, 8 Feb 2025 14:16:36 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 10:03:05 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Arbuzova",
"E. V.",
""
],
[
"Dolgov",
"A. D.",
""
]
] |
2502.05600 | Kunjie Ren | Kunjie Ren, Luo Luo | A Parameter-Free and Near-Optimal Zeroth-Order Algorithm for Stochastic
Convex Optimization | 26 pages, 9 figures | null | null | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper considers zeroth-order optimization for stochastic convex
minimization problem. We propose a parameter-free stochastic zeroth-order
method (POEM) by introducing a step-size scheme based on the distance over
finite difference and an adaptive smoothing parameter. We provide the
theoretical analysis to show that POEM achieves the near-optimal stochastic
zeroth-order oracle complexity. We further conduct the numerical experiments to
demonstrate POEM outperforms existing zeroth-order methods in practice.
| [
{
"version": "v1",
"created": "Sat, 8 Feb 2025 15:02:51 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 05:36:29 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Ren",
"Kunjie",
""
],
[
"Luo",
"Luo",
""
]
] |
2502.05961 | Shu Ishida | Siobhan Mackenzie Hall, Samantha Dalal, Raesetje Sefala, Foutse
Yuehgoh, Aisha Alaagib, Imane Hamzaoui, Shu Ishida, Jabez Magomere, Lauren
Crais, Aya Salama, Tejumade Afonja | The Human Labour of Data Work: Capturing Cultural Diversity through
World Wide Dishes | null | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper provides guidance for building and maintaining infrastructure for
participatory AI efforts by sharing reflections on building World Wide Dishes
(WWD), a bottom-up, community-led image and text dataset of culinary dishes and
associated cultural customs. We present WWD as an example of participatory
dataset creation, where community members both guide the design of the research
process and contribute to the crowdsourced dataset. This approach incorporates
localised expertise and knowledge to address the limitations of web-scraped
Internet datasets acknowledged in the Participatory AI discourse. We show that
our approach can result in curated, high-quality data that supports
decentralised contributions from communities that do not typically contribute
to datasets due to a variety of systemic factors. Our project demonstrates the
importance of participatory mediators in supporting community engagement by
identifying the kinds of labour they performed to make WWD possible. We surface
three dimensions of labour performed by participatory mediators that are
crucial for participatory dataset construction: building trust with community
members, making participation accessible, and contextualising community values
to support meaningful data collection. Drawing on our findings, we put forth
five lessons for building infrastructure to support future participatory AI
efforts.
| [
{
"version": "v1",
"created": "Sun, 9 Feb 2025 17:09:46 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 16:29:14 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Hall",
"Siobhan Mackenzie",
""
],
[
"Dalal",
"Samantha",
""
],
[
"Sefala",
"Raesetje",
""
],
[
"Yuehgoh",
"Foutse",
""
],
[
"Alaagib",
"Aisha",
""
],
[
"Hamzaoui",
"Imane",
""
],
[
"Ishida",
"Shu",
""
],
[
"Magomere",
"Jabez",
""
],
[
"Crais",
"Lauren",
""
],
[
"Salama",
"Aya",
""
],
[
"Afonja",
"Tejumade",
""
]
] |
2502.06197 | Peinuan Qin | Peinuan Qin, Chi-Lan Yang, Jingshu Li, Jing Wen, and Yi-Chieh Lee | Timing Matters: How Using LLMs at Different Timings Influences Writers'
Perceptions and Ideation Outcomes in AI-Assisted Ideation | 16 pages | CHI '25: Proceedings of the 2025 CHI Conference on Human Factors
in Computing Systems, Article 25, Pages 1-16, April 2025 | 10.1145/3706598.3713146 | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) have been widely used to support ideation in the
writing process. However, whether generating ideas with the help of LLMs leads
to idea fixation or idea expansion is unclear. This study examines how
different timings of LLM usage - either at the beginning or after independent
ideation - affect people's perceptions and ideation outcomes in a writing task.
In a controlled experiment with 60 participants, we found that using LLMs from
the beginning reduced the number of original ideas and lowered creative
self-efficacy and self-credit, mediated by changes in autonomy and ownership.
We discuss the challenges and opportunities associated with using LLMs to
assist in idea generation. We propose delaying the use of LLMs to support
ideation while considering users' self-efficacy, autonomy, and ownership of the
ideation outcomes.
| [
{
"version": "v1",
"created": "Mon, 10 Feb 2025 06:51:50 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Qin",
"Peinuan",
""
],
[
"Yang",
"Chi-Lan",
""
],
[
"Li",
"Jingshu",
""
],
[
"Wen",
"Jing",
""
],
[
"Lee",
"Yi-Chieh",
""
]
] |
2502.06458 | Yu Shi | Yu Shi | Chien-Shiung Wu as the experimental pioneer in quantum entanglement: a
2022 note | 6 pages, published version, a few websites are included | Modern Physics Letters A 40, Nos. 9 & 10 (2025) 2530001 | 10.1142/S0217732325300010 | null | physics.hist-ph hep-ex hep-ph hep-th quant-ph | http://creativecommons.org/licenses/by/4.0/ | Advised by Prof. Chen Ning Yang, we review the early pioneering work by
Chien-Shiung Wu on entangled photons created from the electron-positron
annihilation. This paper formed a basis of the author's speech at International
Symposium Commemorating the 110th Birth Anniversary of Chien-Shiung Wu, May 31,
2022, and of a few articles on the subject in Chinese.
| [
{
"version": "v1",
"created": "Mon, 10 Feb 2025 13:36:18 GMT"
},
{
"version": "v2",
"created": "Fri, 25 Apr 2025 09:10:59 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Shi",
"Yu",
""
]
] |
2502.06491 | He Shenghong Sir | Shenghong He | Model-Based Offline Reinforcement Learning with Reliability-Guaranteed
Sequence Modeling | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Model-based offline reinforcement learning (MORL) aims to learn a policy by
exploiting a dynamics model derived from an existing dataset. Applying
conservative quantification to the dynamics model, most existing works on MORL
generate trajectories that approximate the real data distribution to facilitate
policy learning by using current information (e.g., the state and action at
time step $t$). However, these works neglect the impact of historical
information on environmental dynamics, leading to the generation of unreliable
trajectories that may not align with the real data distribution. In this paper,
we propose a new MORL algorithm \textbf{R}eliability-guaranteed
\textbf{T}ransformer (RT), which can eliminate unreliable trajectories by
calculating the cumulative reliability of the generated trajectory (i.e., using
a weighted variational distance away from the real data). Moreover, by sampling
candidate actions with high rewards, RT can efficiently generate high-return
trajectories from the existing offline data. We theoretically prove the
performance guarantees of RT in policy learning, and empirically demonstrate
its effectiveness against state-of-the-art model-based methods on several
benchmark tasks.
| [
{
"version": "v1",
"created": "Mon, 10 Feb 2025 14:08:55 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 07:43:43 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"He",
"Shenghong",
""
]
] |
2502.07098 | Jared Marx-Kuo | Pedro Gaspar, Jared Marx-Kuo | Infinitely Many Surfaces with Prescribed Mean Curvature in the Presence
of a Strictly Stable Minimal Surface | 40 pages, 14 figures. Comments welcome! V2 Updated references. V3
slight improvement of main theorems, minor remarks added. arXiv admin note:
text overlap with arXiv:2011.04136, arXiv:1806.08816, arXiv:1901.01173 by
other authors | null | null | null | math.DG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We construct infinitely many distinct hypersurfaces with prescribed mean
curvature (PMC) for a large class of prescribing functions when $(M^{n+1}, g)$
is a closed smooth manifold containing a minimal surface that is strictly
stable (or more generally, admits a contracting neighborhood). In particular,
we construct infinitely many distinct PMCs when $H_n(M, \mathbb{Z}_2) \neq 0$,
or if $(M, g)$ does not satisfy the Frankel property. Our construction
synthesizes ideas from Song's construction of infinitely many minimal surfaces
in the non-generic setting, Dey's construction of multiple constant mean
curvature surfaces, and Sun--Wang--Zhou's min-max construction of free boundary
PMCs.
| [
{
"version": "v1",
"created": "Mon, 10 Feb 2025 22:43:05 GMT"
},
{
"version": "v2",
"created": "Thu, 13 Feb 2025 02:11:32 GMT"
},
{
"version": "v3",
"created": "Sat, 3 May 2025 15:38:14 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Gaspar",
"Pedro",
""
],
[
"Marx-Kuo",
"Jared",
""
]
] |
2502.07116 | Kieron Ivy Turk | Kieron Ivy Turk and Anna Talas and Alice Hutchings | Threat Me Right: A Human HARMS Threat Model for Technical Systems | null | null | null | null | cs.CR | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Threat modelling is the process of identifying potential vulnerabilities in a
system and prioritising them. Existing threat modelling tools focus primarily
on technical systems and are not as well suited to interpersonal threats. In
this paper, we discuss traditional threat modelling methods and their
shortcomings, and propose a new threat modelling framework (HARMS) to identify
non-technical and human factors harms. We also cover a case study of applying
HARMS when it comes to IoT devices such as smart speakers with virtual
assistants.
| [
{
"version": "v1",
"created": "Mon, 10 Feb 2025 23:13:41 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 13:47:41 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Turk",
"Kieron Ivy",
""
],
[
"Talas",
"Anna",
""
],
[
"Hutchings",
"Alice",
""
]
] |
2502.07479 | Milind Cherukuri | Milind Cherukuri | WebChecker: A Versatile EVL Plugin for Validating HTML Pages with
Bootstrap Frameworks | null | null | 10.71097/IJSAT.v16.i2.2584 | null | cs.SE cs.AI | http://creativecommons.org/licenses/by/4.0/ | WebChecker is a plugin for Epsilon Validation Language (EVL), designed to
validate both static and dynamic HTML pages utilizing frameworks like
Bootstrap. By employing configurable EVL constraints, WebChecker enforces
implicit rules governing HTML and CSS frameworks. The effectiveness of the
plugin is demonstrated through its application on Bootstrap, the widely adopted
HTML, CSS, and JavaScript framework. WebChecker comes with a set of EVL
constraints to assess Bootstrap based web pages. To substantiate our claims, I
present an illustrative example featuring two solutions that effectively
enforce implicit rules.
| [
{
"version": "v1",
"created": "Tue, 11 Feb 2025 11:40:43 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Cherukuri",
"Milind",
""
]
] |
2502.07502 | Zoe Pfister | Zoe Pfister, Michael Vierhauser, Rebekka Wohlrab, Ruth Breu | Towards a Value-Complemented Framework for Enabling Human Monitoring in
Cyber-Physical Systems | 5 pages, accepted at REFSQ25 | In Requirements Engineering: Foundation for Software Quality,
edited by Anne Hess and Angelo Susi, 15588:3-12. Lecture Notes in Computer
Science. Cham: Springer Nature Switzerland, 2025 | 10.1007/978-3-031-88531-0_1 | null | cs.SE | http://creativecommons.org/licenses/by-sa/4.0/ | [Context and Motivation]: Cyber-Physical Systems (CPS) have become relevant
in a wide variety of different domains, integrating hardware and software,
often operating in an emerging and uncertain environment where human actors
actively or passively engage with the CPS. To ensure correct and safe
operation, and self-adaptation, monitors are used for collecting and analyzing
diverse runtime information. [Problem]: However, monitoring humans at runtime,
collecting potentially sensitive information about their actions and behavior,
comes with significant ramifications that can severely hamper the successful
integration of human-machine collaboration. Requirements engineering (RE)
activities must integrate diverse human values, including Privacy, Security,
and Self-Direction during system design, to avoid involuntary data sharing or
misuse. [Principal Ideas]: In this research preview, we focus on the importance
of incorporating these aspects in the RE lifecycle of eliciting and creating
runtime monitors. [Contribution]: We derived an initial conceptual framework,
building on the value taxonomy introduced by Schwartz and human value
integrated Software Engineering by Whittle, further leveraging the concept of
value tactics. The goal is to tie functional and non-functional monitoring
requirements to human values and establish traceability between values,
requirements, and actors. Based on this, we lay out a research roadmap guiding
our ongoing work in this area.
| [
{
"version": "v1",
"created": "Tue, 11 Feb 2025 12:07:37 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 10:19:36 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Pfister",
"Zoe",
""
],
[
"Vierhauser",
"Michael",
""
],
[
"Wohlrab",
"Rebekka",
""
],
[
"Breu",
"Ruth",
""
]
] |
2502.07578 | Yufeng Gu | Yufeng Gu, Alireza Khadem, Sumanth Umesh, Ning Liang, Xavier Servot,
Onur Mutlu, Ravi Iyer, and Reetuparna Das | PIM Is All You Need: A CXL-Enabled GPU-Free System for Large Language
Model Inference | In Proceedings of the 30th ACM International Conference on
Architectural Support for Programming Languages and Operating Systems, Volume
2 (ASPLOS'25) | null | 10.1145/3676641.3716267 | null | cs.AR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Large Language Model (LLM) inference uses an autoregressive manner to
generate one token at a time, which exhibits notably lower operational
intensity compared to earlier Machine Learning (ML) models such as encoder-only
transformers and Convolutional Neural Networks. At the same time, LLMs possess
large parameter sizes and use key-value caches to store context information.
Modern LLMs support context windows with up to 1 million tokens to generate
versatile text, audio, and video content. A large key-value cache unique to
each prompt requires a large memory capacity, limiting the inference batch
size. Both low operational intensity and limited batch size necessitate a high
memory bandwidth. However, contemporary hardware systems for ML model
deployment, such as GPUs and TPUs, are primarily optimized for compute
throughput. This mismatch challenges the efficient deployment of advanced LLMs
and makes users pay for expensive compute resources that are poorly utilized
for the memory-bound LLM inference tasks.
We propose CENT, a CXL-ENabled GPU-Free sysTem for LLM inference, which
harnesses CXL memory expansion capabilities to accommodate substantial LLM
sizes, and utilizes near-bank processing units to deliver high memory
bandwidth, eliminating the need for expensive GPUs. CENT exploits a scalable
CXL network to support peer-to-peer and collective communication primitives
across CXL devices. We implement various parallelism strategies to distribute
LLMs across these devices. Compared to GPU baselines with maximum supported
batch sizes and similar average power, CENT achieves 2.3$\times$ higher
throughput and consumes 2.9$\times$ less energy. CENT enhances the Total Cost
of Ownership (TCO), generating 5.2$\times$ more tokens per dollar than GPUs.
| [
{
"version": "v1",
"created": "Tue, 11 Feb 2025 14:25:20 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Mar 2025 21:10:02 GMT"
},
{
"version": "v3",
"created": "Sat, 3 May 2025 04:07:07 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Gu",
"Yufeng",
""
],
[
"Khadem",
"Alireza",
""
],
[
"Umesh",
"Sumanth",
""
],
[
"Liang",
"Ning",
""
],
[
"Servot",
"Xavier",
""
],
[
"Mutlu",
"Onur",
""
],
[
"Iyer",
"Ravi",
""
],
[
"Das",
"Reetuparna",
""
]
] |
2502.07653 | Jatin Yadav | Ragesh Jaiswal, Amit Kumar, Jatin Yadav | Robust-Sorting and Applications to Ulam-Median | Abstract shortened to meet arXiv requirements | null | null | null | cs.DS | http://creativecommons.org/licenses/by/4.0/ | Sorting is one of the most basic primitives in many algorithms and data
analysis tasks. Comparison-based sorting algorithms, like quick-sort and
merge-sort, are known to be optimal when the outcome of each comparison is
error-free. However, many real-world sorting applications operate in scenarios
where the outcome of each comparison can be noisy. In this work, we explore
settings where a bounded number of comparisons are potentially corrupted by
erroneous agents, resulting in arbitrary, adversarial outcomes.
We model the sorting problem as a query-limited tournament graph where edges
involving erroneous nodes may yield arbitrary results. Our primary contribution
is a randomized algorithm inspired by quick-sort that, in expectation, produces
an ordering close to the true total order while only querying $\tilde{O}(n)$
edges. We achieve a distance from the target order $\pi$ within $(3 +
\epsilon)|B|$, where $B$ is the set of erroneous nodes, balancing the competing
objectives of minimizing both query complexity and misalignment with $\pi$. Our
algorithm needs to carefully balance two aspects: identify a pivot that
partitions the vertex set evenly and ensure that this partition is "truthful"
and yet query as few "triangles" in the graph $G$ as possible. Since the nodes
in $B$ can potentially hide in an intricate manner, our algorithm requires
several technical steps.
Additionally, we demonstrate significant implications for the Ulam-$k$-Median
problem, a classical clustering problem where the metric is defined on the set
of permutations on a set of $d$ elements. Chakraborty, Das, and Krauthgamer
gave a $(2-\varepsilon)$ FPT approximation algorithm for this problem, where
the running time is super-linear in both $n$ and $d$. We use our robust sorting
framework to give the first $(2-\varepsilon)$ FPT linear time approximation
algorithm for this problem.
| [
{
"version": "v1",
"created": "Tue, 11 Feb 2025 15:40:27 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 11:59:10 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Jaiswal",
"Ragesh",
""
],
[
"Kumar",
"Amit",
""
],
[
"Yadav",
"Jatin",
""
]
] |
2502.08033 | Anjian Li | Anjian Li, Sangjae Bae, David Isele, Ryne Beeson, Faizan M. Tariq | Predictive Planner for Autonomous Driving with Consistency Models | null | null | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Trajectory prediction and planning are essential for autonomous vehicles to
navigate safely and efficiently in dynamic environments. Traditional approaches
often treat them separately, limiting the ability for interactive planning.
While recent diffusion-based generative models have shown promise in
multi-agent trajectory generation, their slow sampling is less suitable for
high-frequency planning tasks. In this paper, we leverage the consistency model
to build a predictive planner that samples from a joint distribution of ego and
surrounding agents, conditioned on the ego vehicle's navigational goal. Trained
on real-world human driving datasets, our consistency model generates
higher-quality trajectories with fewer sampling steps than standard diffusion
models, making it more suitable for real-time deployment. To enforce multiple
planning constraints simultaneously on the ego trajectory, a novel online
guided sampling approach inspired by the Alternating Direction Method of
Multipliers (ADMM) is introduced. Evaluated on the Waymo Open Motion Dataset
(WOMD), our method enables proactive behavior such as nudging and yielding, and
also demonstrates smoother, safer, and more efficient trajectories and
satisfaction of multiple constraints under a limited computational budget.
| [
{
"version": "v1",
"created": "Wed, 12 Feb 2025 00:26:01 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 19:45:46 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Li",
"Anjian",
""
],
[
"Bae",
"Sangjae",
""
],
[
"Isele",
"David",
""
],
[
"Beeson",
"Ryne",
""
],
[
"Tariq",
"Faizan M.",
""
]
] |
2502.08485 | Joachim Tapparel | Joachim Tapparel and Andreas Burg | LoRa Fine Synchronization with Two-Pass Time and Frequency Offset
Estimation | Accepted for publication in ACSSC | 58th Asilomar Conf. Signals Syst. Comp. (2024) 1802-1806 | 10.1109/IEEECONF60004.2024.10942846 | null | eess.SP | http://creativecommons.org/licenses/by/4.0/ | LoRa is currently one of the most widely used low-power wide-area network
(LPWAN) technologies. The physical layer leverages a chirp spread spectrum
modulation to achieve long-range communication with low power consumption.
Synchronization at long distances is a challenging task as the spread signal
can lie multiple orders of magnitude below the thermal noise floor. Multiple
research works have proposed synchronization algorithms for LoRa under
different hardware impairments. However, the impact of sampling frequency
offset (SFO) has mostly either been ignored or tracked only during the data
phase, but it often harms synchronization. In this work, we extend existing
synchronization algorithms for LoRa to estimate and compensate SFO already in
the preamble and show that this early compensation has a critical impact on the
estimation of other impairments such as carrier frequency offset and sampling
time offset. Therefore it is critical to recover long-range signals.
| [
{
"version": "v1",
"created": "Wed, 12 Feb 2025 15:20:57 GMT"
},
{
"version": "v2",
"created": "Mon, 17 Feb 2025 10:15:29 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Tapparel",
"Joachim",
""
],
[
"Burg",
"Andreas",
""
]
] |
2502.08711 | Eric Lescano | Eric Lescano | A Non-Relativistic Limit for Heterotic Supergravity and its Gauge
Lagrangian | 15 pages+refs. V2:typos fixed. V3:inclusion of the ungauged
Lagrangian in covariant form | null | null | null | hep-th | http://creativecommons.org/licenses/by/4.0/ | We show that the $D=10$ heterotic supergravity in a non-relativistic (NR)
limit has a finite Lagrangian due to non-trivial cancellations of divergent
parts arising from the Chern-Simons terms in the curvature of the $\hat
B$-field and the Yang-Mills Lagrangian. This is similar to what happens in
bosonic supergravity between the Ricci scalar, $\hat R$, and the $-
\frac{1}{12} \hat H^2$ term, under the same limit. In this work we present the
explicit form of the gauge transformations and curvatures after considering a
NR limit and we compute the full finite Lagrangian in its covariant form.
Particularly, we address both the ungauged part and the gauged part, and we
write it using covariant derivatives and curvatures for the gauge fields. These
results constitutes the first computation of a finite Lagrangian for heterotic
supergravity under these kinds of limits. As an interesting feature of the
expansion implemented, the Green-Schwarz mechanism for the two-form can be
trivialized and terms equivalent to Chern-Simons contributions naturally arise
from the previous property.
| [
{
"version": "v1",
"created": "Wed, 12 Feb 2025 19:00:01 GMT"
},
{
"version": "v2",
"created": "Sat, 15 Feb 2025 20:28:22 GMT"
},
{
"version": "v3",
"created": "Mon, 5 May 2025 10:38:18 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Lescano",
"Eric",
""
]
] |
2502.08798 | Jan Ebr | Jan Ebr, Ji\v{r}\'i Bla\v{z}ek, Jakub V\'icha, Tanguy Pierog, Eva
Santos, Petr Tr\'avn\'i\v{c}ek, Nikolas Denner, Ralf Ulrich | Modified Hadronic Interactions and the future of UHECR observations | AAM published in proceedings of UHECR 2024, CC-BY license | PoS UHECR2024 (2025) 046 | 10.22323/1.484.0046 | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Data from multiple experiments suggest that the current interaction models
used in Monte Carlo simulations do not correctly reproduce the hadronic
interactions in air showers produced by ultra-high-energy cosmic rays (UHECR).
We have created a large library of UHECR simulations where the interactions at
the highest energies are slightly modified in various ways - but always within
the constraints of the accelerator data, without any abrupt changes with energy
and without assuming any specific mechanism or dramatically new physics at the
ultra-high energies. Recent results of the Pierre Auger Observatory indicate a
need for a change in the prediction of the models for both the muon content at
ground and the depth of the maximum of longitudinal development of the shower.
In our parameter space, we find combinations of modifications that are in
agreement with this analysis, however a consistent description of UHECR showers
remains elusive. Our library however provides a realistic representation of the
freedom in the modeling of the hadronic interactions and offers an opportunity
to quantify uncertainties of various predictions. This can be particularly
valuable for the design of future observatories where hadronic models are often
used as input for the prediction of the performance. We demonstrate this
powerful capability on several selected examples.
| [
{
"version": "v1",
"created": "Wed, 12 Feb 2025 21:21:20 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 12:43:14 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Ebr",
"Jan",
""
],
[
"Blažek",
"Jiří",
""
],
[
"Vícha",
"Jakub",
""
],
[
"Pierog",
"Tanguy",
""
],
[
"Santos",
"Eva",
""
],
[
"Trávníček",
"Petr",
""
],
[
"Denner",
"Nikolas",
""
],
[
"Ulrich",
"Ralf",
""
]
] |
2502.09061 | Debangshu Banerjee | Debangshu Banerjee, Tarun Suresh, Shubham Ugare, Sasa Misailovic,
Gagandeep Singh | CRANE: Reasoning with constrained LLM generation | Accepted at ICML 2025 | null | null | null | cs.PL cs.LG | http://creativecommons.org/licenses/by/4.0/ | Code generation, symbolic math reasoning, and other tasks require LLMs to
produce outputs that are both syntactically and semantically correct.
Constrained LLM generation is a promising direction to enforce adherence to
formal grammar, but prior works have empirically observed that strict
enforcement of formal constraints often diminishes the reasoning capabilities
of LLMs. In this work, we first provide a theoretical explanation for why
constraining LLM outputs to very restrictive grammars that only allow
syntactically valid final answers reduces the reasoning capabilities of the
model. Second, we demonstrate that by augmenting the output grammar with
carefully designed additional rules, it is always possible to preserve the
reasoning capabilities of the LLM while ensuring syntactic and semantic
correctness in its outputs. Building on these theoretical insights, we propose
a reasoning-augmented constrained decoding algorithm, CRANE, which effectively
balances the correctness of constrained generation with the flexibility of
unconstrained generation. Experiments on multiple open-source LLMs and
benchmarks show that CRANE significantly outperforms both state-of-the-art
constrained decoding strategies and standard unconstrained decoding, showing up
to 10% points accuracy improvement over baselines on challenging symbolic
reasoning benchmarks GSM-symbolic and FOLIO.
| [
{
"version": "v1",
"created": "Thu, 13 Feb 2025 08:23:42 GMT"
},
{
"version": "v2",
"created": "Sun, 6 Apr 2025 06:09:46 GMT"
},
{
"version": "v3",
"created": "Sun, 4 May 2025 05:21:01 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Banerjee",
"Debangshu",
""
],
[
"Suresh",
"Tarun",
""
],
[
"Ugare",
"Shubham",
""
],
[
"Misailovic",
"Sasa",
""
],
[
"Singh",
"Gagandeep",
""
]
] |
2502.09066 | Aymeric Walch | Aymeric Walch (IRIF) | Compositional Taylor expansion in cartesian differential categories | null | null | null | null | cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper provides a compositional approach to Taylor expansion, in the
setting of cartesian differential categories. Taylor expansion is captured here
by a functor that generalizes the tangent bundle functor to higher order
derivatives. The fundamental properties of Taylor expansion then boils down to
naturality equations that turns this functor into a monad. This monad provides
a categorical approach to higher order dual numbers and the jet bundle
construction used in automated differentiation.
| [
{
"version": "v1",
"created": "Thu, 13 Feb 2025 08:30:34 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Walch",
"Aymeric",
"",
"IRIF"
]
] |
2502.09203 | Dongrui Wu | Dongrui Wu | Revisiting Euclidean Alignment for Transfer Learning in EEG-Based
Brain-Computer Interfaces | null | null | null | null | cs.HC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Due to large intra-subject and inter-subject variabilities of
electroencephalogram (EEG) signals, EEG-based brain-computer interfaces (BCIs)
usually need subject-specific calibration to tailor the decoding algorithm for
each new subject, which is time-consuming and user-unfriendly, hindering their
real-world applications. Transfer learning (TL) has been extensively used to
expedite the calibration, by making use of EEG data from other
subjects/sessions. An important consideration in TL for EEG-based BCIs is to
reduce the data distribution discrepancies among different subjects/sessions,
to avoid negative transfer. Euclidean alignment (EA) was proposed in 2020 to
address this challenge. Numerous experiments from 13 different BCI paradigms
demonstrated its effectiveness and efficiency. This paper revisits EA,
explaining its procedure and correct usage, introducing its applications and
extensions, and pointing out potential new research directions. It should be
very helpful to BCI researchers, especially those who are working on EEG signal
decoding.
| [
{
"version": "v1",
"created": "Thu, 13 Feb 2025 11:43:43 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 14:58:21 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Wu",
"Dongrui",
""
]
] |
2502.09317 | Nitesh Jaiswal | Nitesh Jaiswal and S. Shankaranarayanan (IIT Bombay) | Analog charged black hole formation via percolation: Exploring cosmic
censorship and Hoop conjecture | Version accepted in Physical review D letters | null | null | null | gr-qc cond-mat.stat-mech cond-mat.str-el hep-th | http://creativecommons.org/licenses/by/4.0/ | We investigate an analog model of charged black hole (BH) formation using the
framework of classical percolation. By analyzing the scaling behavior of key
quantities, including surface gravity and Komar mass, we establish a robust
correspondence between this analog system and gravitational collapse in general
relativity. Our numerical simulations of the lattice model show excellent
agreement with analytical predictions for the continuum limit, highlighting the
potential of analog systems to capture essential features of BH physics.
Interestingly, we find that while geometric criteria related to the hoop
conjecture are necessary, they are not sufficient for BH formation in our
model. Instead, the exponential growth of energy and cluster size emerges as
the key indicator, suggesting a novel interpretation of the hoop conjecture and
providing further support for cosmic censorship within our analog framework by
ensuring horizon formation. This work offers a fresh perspective on the
organization of matter within BH event horizons and lays the groundwork for
future quantum extensions that could shed light on Hawking radiation and the BH
information paradox by linking entanglement entropy in quantum percolation
models to BH entropy.
| [
{
"version": "v1",
"created": "Thu, 13 Feb 2025 13:33:21 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 08:03:11 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Jaiswal",
"Nitesh",
"",
"IIT Bombay"
],
[
"Shankaranarayanan",
"S.",
"",
"IIT Bombay"
]
] |
2502.10649 | Patrick Schelling | Patrick K. Schelling, Antonio Martinez Margolles, and Logan Echazabal | Thermal response functions and second sound in single-layer hexagonal
boron nitride | 27 pages, 11 figures, to be submitted to Phys. Rev. B | null | null | null | cond-mat.stat-mech cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Ballistic heat transport and second sound propagation in solids is of direct
relevance in electronic and energy applications at short length scales and low
temperatures. Measurement or calculation of thermal conductivity, which is
typically a primary objective, may be of limited usefulness for predicting heat
transport which does not follow the heat-diffusion equation. In this paper,
molecular-dynamics simulations of hexagonal BN (h-BN) are used to compute
thermal response functions from equilibrium correlation functions defined in
Fourier space. The response functions are useful for describing the
time-dependent transport beyond the usual assumptions of Fourier's law. The
results demonstrate that for length scales ~110nm at T=100K second sound should
be experimentally observable. At higher temperatures and longer length scales,
while second sound may not be directly observable, thermal transport can
nevertheless strongly deviate from predictions based on the heat-diffusion
equation. Along with classical simulations, we outline a first-principles,
many-body theoretical approach for calculation of the response function based
on solutions of the Bethe-Salpeter equation. The relevant expressions for heat
current clarify the importance of phase coherence within a phonon branch to the
observation of second sound. Previous work on one-dimensional chains is also
discussed to show that materials characterized by linear dispersion and simple
phonon band structure should more readily display second sound.
| [
{
"version": "v1",
"created": "Sat, 15 Feb 2025 02:58:41 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 03:56:50 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Schelling",
"Patrick K.",
""
],
[
"Margolles",
"Antonio Martinez",
""
],
[
"Echazabal",
"Logan",
""
]
] |
2502.10969 | Lin Wang | Yi Liu and Lin Wang | Moser's twist theorem revisited | null | null | null | null | math.DS | http://creativecommons.org/licenses/by/4.0/ | Inspired by the work of Katznelson and Ornstein, we present a short way to
achieve the almost optimal regularity in Moser's twist theorem. Specifically,
for an integrable area-preserving twist map, the invariant circle with a given
constant type frequency $\alpha$ persists under a small perturbation (dependent
on $\alpha$) of class $C^{3+\epsilon}$. This result was initially established
independently by Herman and R\"{u}ssmann in 1983. Our method differs
essentially from their approaches.
| [
{
"version": "v1",
"created": "Sun, 16 Feb 2025 03:07:01 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 07:21:50 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Liu",
"Yi",
""
],
[
"Wang",
"Lin",
""
]
] |
2502.11025 | Ichiro Shimada | Ichiro Shimada | The automorphism group of an Ap\'ery-Fermi K3 surface | 38 pages, 14 figures | null | null | null | math.AG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An Ap\'ery-Fermi K3 surface is a complex K3 surface of Picard number 19 that
is birational to a general member of a certain one-dimensional family of affine
surfaces related to the Fermi surface in solid-state physics. This K3 surface
is also linked to a recurrence relation that appears in the famous proof of the
irrationality of zeta(3) by Ap\'ery. We compute the automorphism group Aut(X)
of the Ap\'ery-Fermi K3 surface X using Borcherds' method. We describe Aut(X)
in terms of generators and relations. Moreover, we determine the action of
Aut(X) on the set of ADE-configurations of smooth rational curves on X for some
ADE-types. In particular, we show that Aut(X) acts transitively on the set of
smooth rational curves, and that it partitions the set of pairs of disjoint
smooth rational curves into two orbits.
| [
{
"version": "v1",
"created": "Sun, 16 Feb 2025 07:19:50 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 07:25:47 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Shimada",
"Ichiro",
""
]
] |
2502.11045 | El-Mehdi Mehiri | El-Mehdi Mehiri, Saad Mneimneh, Hac\`ene Belbachir | The Towers of Fibonacci, Lucas, Pell, and Jacobsthal | The manuscript contains errors. A corrected version will be provided
later | null | null | null | math.CO cs.DM | http://creativecommons.org/licenses/by/4.0/ | We present in this paper four new variants of the Tower of Hanoi problem, the
optimal solution of each of these variants is related to one of the four known
numbers Fibonacci, Lucas, Pell, and Jacobsthal. We give an optimal solution to
each of these variants, and we present their associated graphs.
| [
{
"version": "v1",
"created": "Sun, 16 Feb 2025 09:07:05 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 18:56:32 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Mehiri",
"El-Mehdi",
""
],
[
"Mneimneh",
"Saad",
""
],
[
"Belbachir",
"Hacène",
""
]
] |
2502.11274 | Kyriakos Destounis Dr. | H. S. Vieira, Kyriakos Destounis and Kostas D. Kokkotas | Perturbing the vortex: quasinormal and quasibound spectra of rotating
acoustic geometries | 12 pages, 2 figures, 1 appendix, accepted for publication in PRD | null | null | null | gr-qc hep-th | http://creativecommons.org/licenses/by/4.0/ | Strong-field gravity simulators are laboratory experiments that can
investigate a wide range of both classical and quantum phenomena occurring in
nature. In this work, we introduce an effective geometry that captures most of
the characteristics of the strong-field regime of astrophysical, rotating black
holes. This geometry can represent a vortex made from a variety of fluid and
superfluid profiles with zero viscosity, making it a promising
finite-temperature quantum-field-theory simulator for rotating curved
spacetimes. Our geometry includes not only the typical radial flow which gives
rise to an acoustic horizon, but also azimuthal circulation of the fluid. We
compute the quasinormal modes, semi-analytically, and the exact quasibound
states of acoustic excitations interacting with this effective geometry. The
resulting spectra can be identified for both co-rotating and counter-rotating
surface acoustic waves. In particular, the behavior of our acoustic geometry
with circulation aligns with the phenomenology observed in recent experiments
that include superfluids.
| [
{
"version": "v1",
"created": "Sun, 16 Feb 2025 21:31:07 GMT"
},
{
"version": "v2",
"created": "Sat, 22 Feb 2025 08:47:45 GMT"
},
{
"version": "v3",
"created": "Thu, 1 May 2025 17:51:27 GMT"
},
{
"version": "v4",
"created": "Fri, 2 May 2025 17:30:37 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Vieira",
"H. S.",
""
],
[
"Destounis",
"Kyriakos",
""
],
[
"Kokkotas",
"Kostas D.",
""
]
] |
2502.11905 | Haftu Fentaw | Haftu W. Fentaw, Steve Campbell, and Simon Caton | Exploring Quantum Control Landscape and Solution Space Complexity
through Dimensionality Reduction & Optimization Algorithms | null | Sci. Rep. 15, 14605 (2025) | 10.1038/s41598-025-95161-0 | null | quant-ph cs.ET cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding the quantum control landscape (QCL) is important for designing
effective quantum control strategies. In this study, we analyze the QCL for a
single two-level quantum system (qubit) using various control strategies. We
employ Principal Component Analysis (PCA), to visualize and analyze the QCL for
higher dimensional control parameters. Our results indicate that dimensionality
reduction techniques such as PCA, can play an important role in understanding
the complex nature of quantum control in higher dimensions. Evaluations of
traditional control techniques and machine learning algorithms reveal that
Genetic Algorithms (GA) outperform Stochastic Gradient Descent (SGD), while
Q-learning (QL) shows great promise compared to Deep Q-Networks (DQN) and
Proximal Policy Optimization (PPO). Additionally, our experiments highlight the
importance of reward function design in DQN and PPO demonstrating that using
immediate reward results in improved performance rather than delayed rewards
for systems with short time steps. A study of solution space complexity was
conducted by using Cluster Density Index (CDI) as a key metric for analyzing
the density of optimal solutions in the landscape. The CDI reflects cluster
quality and helps determine whether a given algorithm generates regions of high
fidelity or not. Our results provide insights into effective quantum control
strategies, emphasizing the significance of parameter selection and algorithm
optimization.
| [
{
"version": "v1",
"created": "Mon, 17 Feb 2025 15:26:15 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Fentaw",
"Haftu W.",
""
],
[
"Campbell",
"Steve",
""
],
[
"Caton",
"Simon",
""
]
] |
2502.12050 | Yasmin Moslem | Yasmin Moslem, Juan Juli\'an Cea Mor\'an, Mariano Gonzalez-Gomez,
Muhammad Hazim Al Farouq, Farah Abdou, Satarupa Deb | SpeechT: Findings of the First Mentorship in Speech Translation | MT Summit 2025 | null | null | null | cs.CL cs.SD | http://creativecommons.org/licenses/by/4.0/ | This work presents the details and findings of the first mentorship in speech
translation (SpeechT), which took place in December 2024 and January 2025. To
fulfil the mentorship requirements, the participants engaged in key activities,
including data preparation, modelling, and advanced research. The participants
explored data augmentation techniques and compared end-to-end and cascaded
speech translation systems. The projects covered various languages other than
English, including Arabic, Bengali, Galician, Indonesian, Japanese, and
Spanish.
| [
{
"version": "v1",
"created": "Mon, 17 Feb 2025 17:18:39 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 15:39:01 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Moslem",
"Yasmin",
""
],
[
"Morán",
"Juan Julián Cea",
""
],
[
"Gonzalez-Gomez",
"Mariano",
""
],
[
"Farouq",
"Muhammad Hazim Al",
""
],
[
"Abdou",
"Farah",
""
],
[
"Deb",
"Satarupa",
""
]
] |
2502.12426 | Subaru Yoshimoto | S. Yoshimoto, Y. Tabata, T. Waki, and H. Nakamura | Non-ergodic Phase Transition in the Global Hysteresis of the Frustrated
Magnet DyRu2Si2 | 15 pages, 15 figures | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | Some frustrated magnets exhibit a huge hysteresis called "global hysteresis
(GH)", where the magnetic plateaus appearing in the increasing field process
are skipped in the decreasing field process from the high magnetic field state.
In this paper, we focused on the frustrated magnet DyRu2Si2 and measured
magnetization relaxations from two plateau states inside the GH loop, the
phases III and IV, and investigated the phase transitions into them. As a
result of the relaxation measurements, no relaxation is observed in the phase
III, whereas long-time relaxations of more than 105 sec are observed at the
phase IV plateau. Moreover, a Mpemba-effect-like relaxation phenomenon where
the relaxation from an initial state prepared in the zero-field-cooled
condition overtakes that from an initial state prepared in the field-cooled
condition is observed. These results indicate that the phase IV is the
non-ergodic state with a complex free-energy landscape with multiple local
minima, while the phase III has a simple free energy structure. Therefore, the
III-IV phase transition is considered to be the ergodic to non-ergodic phase
transition. Although this type of phase transition typically occurs in random
glassy systems, the phase IV in DyRu2Si2 has a regular long-range ordered
magnetic structure and yet exhibits non-ergodic properties, which is highly
nontrivial. Our findings open the possibility of observing non-ergodic states
in frustrated magnets with regular long-range orders.
| [
{
"version": "v1",
"created": "Tue, 18 Feb 2025 01:52:17 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 07:59:57 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Yoshimoto",
"S.",
""
],
[
"Tabata",
"Y.",
""
],
[
"Waki",
"T.",
""
],
[
"Nakamura",
"H.",
""
]
] |
2502.12878 | Jon Vehovar | Jon Vehovar, Miha Rot, Gregor Kosec | Performance Trade-offs of High Order Meshless Approximation on
Distributed Memory Systems | 6 pages, 6 figures | null | null | null | cs.DC | http://creativecommons.org/licenses/by/4.0/ | Meshless methods approximate operators in a specific node as a weighted sum
of values in its neighbours. Higher order approximations of derivatives provide
more accurate solutions with better convergence characteristics, but they come
at the cost of including more neighbours. On the accuracy-per-compute time
basis we know that increasing the approximation order is beneficial for a
shared memory computer, but there is additional communication overhead when
problems become too large and we have to resort to distributed memory systems.
Meshless nodes are divided between systems in spatially coherent subdomains
with approximations at their edges requiring neighbouring value exchange.
Performance optimization is then a balancing act between minimizing the
required number of communicated neighbours by lowering the approximation order
or increasing it to enable faster convergence. We use the radial basis
function-generated finite difference method (RBF-FD) to approximate the
derivatives that we use to solve the Poisson equation with an explicit
iterative scheme. Inter-system communication is provided by Open MPI, while
OpenMP is used for intra-system parallelisation. We perform the analysis on a
homogenous CPU-based cluster where we examine the behaviour and attempt to
determine the optimal parameterisation with the goal of minimizing the
computational time to reach a desired accuracy.
| [
{
"version": "v1",
"created": "Tue, 18 Feb 2025 14:07:39 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 10:37:29 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Vehovar",
"Jon",
""
],
[
"Rot",
"Miha",
""
],
[
"Kosec",
"Gregor",
""
]
] |
2502.13199 | Akhil Dusi | Suresh Babu Nettur, Shanthi Karpurapu, Unnati Nettur, Likhit Sagar
Gajja, Sravanthy Myneni, and Akhil Dusi | The Role of GitHub Copilot on Software Development: A Perspective on
Productivity, Security, Best Practices and Future Directions | Correspondence and co-first authors: [email protected],
[email protected] | null | null | null | cs.SE cs.AI | http://creativecommons.org/licenses/by-sa/4.0/ | GitHub Copilot is transforming software development by automating tasks and
boosting productivity through AI driven code generation. In this paper, we
conduct a literature survey to synthesize insights on Copilot's impact on
productivity and security. We review academic journal databases, industry
reports, and official documentation to highlight key findings and challenges.
While Copilot accelerates coding and prototyping, concerns over security
vulnerabilities and intellectual property risks persist. Drawing from the
literature, we provide a perspective on best practices and future directions
for responsible AI adoption in software engineering, offering actionable
insights for developers and organizations to integrate Copilot effectively
while maintaining high standards of quality and security.
| [
{
"version": "v1",
"created": "Tue, 18 Feb 2025 18:08:20 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 16:44:12 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Nettur",
"Suresh Babu",
""
],
[
"Karpurapu",
"Shanthi",
""
],
[
"Nettur",
"Unnati",
""
],
[
"Gajja",
"Likhit Sagar",
""
],
[
"Myneni",
"Sravanthy",
""
],
[
"Dusi",
"Akhil",
""
]
] |
2502.13201 | Paolo Walter Cattaneo | Paolo Walter Cattaneo | Constraints on Lorentz invariance from the event KM3-230213A | 2 pages, no figure | null | null | null | hep-ph astro-ph.HE hep-ex | http://creativecommons.org/licenses/by/4.0/ | Lorentz invariance is the cornerstone of relativity theory. Its implications
have been verified experimentally with a variety of approaches. The detection
of a muon at extremely high energy detected by the ARCA detector in the
Mediterranean sea, the most energetic particle directly measured up to date,
allows to put additional constraints on Lorentz non-invariant theories. The
prediction of some of those theories is that the lifetimes of particles in the
laboratory frame 'decrease' rather than 'increase' with increasing $\gamma$. In
this frame the sheer fact that the muon traversed the whole ARCA detector puts
a lower limit on the muon lifetime in the laboratory frame, that implies upper
limits on Lorentz violating parameters.
| [
{
"version": "v1",
"created": "Tue, 18 Feb 2025 18:52:29 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 09:47:01 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Cattaneo",
"Paolo Walter",
""
]
] |
2502.13506 | Thijmen Nijdam | Coen van den Elsen, Francien Barkhof, Thijmen Nijdam, Simon Lupart,
Mohammad Aliannejadi | Reproducing NevIR: Negation in Neural Information Retrieval | 9 pages, 4 figures. Accepted at SIGIR 2025 as a reproducibility paper | null | 10.1145/3726302.3730294 | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | Negation is a fundamental aspect of human communication, yet it remains a
challenge for Language Models (LMs) in Information Retrieval (IR). Despite the
heavy reliance of modern neural IR systems on LMs, little attention has been
given to their handling of negation. In this study, we reproduce and extend the
findings of NevIR, a benchmark study that revealed most IR models perform at or
below the level of random ranking when dealing with negation. We replicate
NevIR's original experiments and evaluate newly developed state-of-the-art IR
models. Our findings show that a recently emerging category-listwise Large
Language Model (LLM) re-rankers-outperforms other models but still
underperforms human performance. Additionally, we leverage ExcluIR, a benchmark
dataset designed for exclusionary queries with extensive negation, to assess
the generalisability of negation understanding. Our findings suggest that
fine-tuning on one dataset does not reliably improve performance on the other,
indicating notable differences in their data distributions. Furthermore, we
observe that only cross-encoders and listwise LLM re-rankers achieve reasonable
performance across both negation tasks.
| [
{
"version": "v1",
"created": "Wed, 19 Feb 2025 07:50:59 GMT"
},
{
"version": "v2",
"created": "Fri, 21 Feb 2025 10:18:26 GMT"
},
{
"version": "v3",
"created": "Thu, 1 May 2025 07:27:34 GMT"
},
{
"version": "v4",
"created": "Sun, 4 May 2025 10:02:00 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Elsen",
"Coen van den",
""
],
[
"Barkhof",
"Francien",
""
],
[
"Nijdam",
"Thijmen",
""
],
[
"Lupart",
"Simon",
""
],
[
"Aliannejadi",
"Mohammad",
""
]
] |
2502.13929 | Enrico Lipparini | Massimo Bartoletti, Silvia Crafa, Enrico Lipparini | Formal verification in Solidity and Move: insights from a comparative
analysis | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Formal verification plays a crucial role in making smart contracts safer,
being able to find bugs or to guarantee their absence, as well as checking
whether the business logic is correctly implemented. For Solidity, even though
there already exist several mature verification tools, the semantical quirks of
the language can make verification quite hard in practice. Move, on the other
hand, has been designed with security and verification in mind, and it has been
accompanied since its early stages by a formal verification tool, the Move
Prover. In this paper, we investigate through a comparative analysis: 1) how
the different designs of the two contract languages impact verification, and 2)
what is the state-of-the-art of verification tools for the two languages, and
how do they compare on three paradigmatic use cases. Our investigation is
supported by an open dataset of verification tasks performed in Certora and in
the Aptos Move Prover.
| [
{
"version": "v1",
"created": "Wed, 19 Feb 2025 18:06:01 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 16:37:06 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Bartoletti",
"Massimo",
""
],
[
"Crafa",
"Silvia",
""
],
[
"Lipparini",
"Enrico",
""
]
] |
2502.13949 | Yury Melikyan | Yu. A. Melikyan, I.G. Bearden, V. Buchakchiev, S. Jia, V. Kozhuharov
and I. P. M{\o}ller | Characterization of the large-size NDL EQR20 silicon photomultipliers | null | null | null | null | physics.ins-det | http://creativecommons.org/licenses/by/4.0/ | Unlike most commercially available silicon photomultipliers (SiPMs), EQR20
SiPMs produced by the Novel Device Laboratory (NDL) avoid using individual
resistors to quench the avalanche multiplication of the microcells. Instead,
bulk resistance of the epitaxial silicon layer is used, and the signal is
directly collected at a common anode plane. This allows for the fabrication of
SiPMs as large as 6.24 x 6.24 mm^2 while keeping the recovery time below {\tau}
= 25 ns. These devices can be composed of microcells with 20 micrometer pitch
while reaching PDE above 50% and 10^6 gain at 5 V overvoltage. On the other
hand, a crosstalk level from 20% to 40% is observed for overvoltages from 3 V
to 5 V. Moreover, significant pulse shape distortion is observed for pulses
above ~100 pC, corresponding to microcell occupancy of a few percent. This work
provides an independent determination of the performance parameters of the
EQR20 11-6060D-S SiPMs. Influence of the pulse shape distortion is discussed
from the perspective of the applicability of these devices to the
scintillator-based forward hadron calorimeter (FoCal-H) of the ALICE apparatus
at CERN.
| [
{
"version": "v1",
"created": "Wed, 19 Feb 2025 18:47:04 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 14:17:15 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Melikyan",
"Yu. A.",
""
],
[
"Bearden",
"I. G.",
""
],
[
"Buchakchiev",
"V.",
""
],
[
"Jia",
"S.",
""
],
[
"Kozhuharov",
"V.",
""
],
[
"Møller",
"I. P.",
""
]
] |
2502.14073 | Matilde Garcia | Juno C. L. Chan, Conor Dyson, Matilde Garcia, Jaime Redondo-Yuste, and
Luka Vujeva | Lensing and wave optics in the strong field of a black hole | 16 pages, 11 figures, and 1 table | null | null | null | gr-qc astro-ph.GA astro-ph.HE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Gravitational waves (GWs) are lensed by matter, offering a unique probe of
both the large-scale structure of the Universe and the fundamental properties
of GW propagation. GWs can also be affected by wave optics effects when their
wavelength is comparable to the size of the lens. While this regime has been
well studied in the Newtonian approximation, the role of strong gravitational
fields remains largely unexplored. This is particularly relevant for lensing by
intermediate and supermassive black holes (BHs), which can occur near active
galactic nuclei or in compact triple systems. In this work, we analyze the
lensing of GWs by a non-rotating BH and compare our results to the Newtonian
point-mass approximation. We construct frequency-dependent amplification
factors that incorporate strong-field effects, revealing explicit polarization
mixing and absorption by the event horizon. Using a fiducial GW event, we
explore key phenomenological signatures of BH lensing, highlighting new
observational opportunities to probe strong gravitational fields through GW
lensing.
| [
{
"version": "v1",
"created": "Wed, 19 Feb 2025 19:57:42 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 14:56:55 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Chan",
"Juno C. L.",
""
],
[
"Dyson",
"Conor",
""
],
[
"Garcia",
"Matilde",
""
],
[
"Redondo-Yuste",
"Jaime",
""
],
[
"Vujeva",
"Luka",
""
]
] |
2502.14259 | Sujeong Im | Sujeong Im, Jungwoo Oh, Edward Choi | LabTOP: A Unified Model for Lab Test Outcome Prediction on Electronic
Health Records | 11 pages for main text, 13 pages for appendix | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Lab tests are fundamental for diagnosing diseases and monitoring patient
conditions. However, frequent testing can be burdensome for patients, and test
results may not always be immediately available. To address these challenges,
we propose LabTOP, a unified model that predicts lab test outcomes by
leveraging a language modeling approach on EHR data. Unlike conventional
methods that estimate only a subset of lab tests or classify discrete value
ranges, LabTOP performs continuous numerical predictions for a diverse range of
lab items. We evaluate LabTOP on three publicly available EHR datasets and
demonstrate that it outperforms existing methods, including traditional machine
learning models and state-of-the-art large language models. We also conduct
extensive ablation studies to confirm the effectiveness of our design choices.
We believe that LabTOP will serve as an accurate and generalizable framework
for lab test outcome prediction, with potential applications in clinical
decision support and early detection of critical conditions.
| [
{
"version": "v1",
"created": "Thu, 20 Feb 2025 04:55:45 GMT"
},
{
"version": "v2",
"created": "Wed, 30 Apr 2025 13:54:58 GMT"
},
{
"version": "v3",
"created": "Sun, 4 May 2025 15:55:51 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Im",
"Sujeong",
""
],
[
"Oh",
"Jungwoo",
""
],
[
"Choi",
"Edward",
""
]
] |
2502.14303 | Tong Li | Tong Li, Michael A. Schmidt, Chang-Yuan Yao | Baryon-number-violating nucleon decays in sterile neutrino effective
field theories | 37 pages, 12 figures, 7 tables. version accepted for publication in
JHEP | null | null | CPPC-2025-02 | hep-ph hep-ex | http://creativecommons.org/licenses/by/4.0/ | The search for baryon-number-violating (BNV) nucleon decay provides an
intriguing probe of new physics beyond the Standard Model (SM). Future neutrino
experiments will improve the sensitivity to BNV nucleon decays and can serve to
search for dark particles. In this work, we study the sterile neutrino
effective field theories (EFTs) with baryon number violation and the impact of
light sterile neutrino on BNV nucleon decays. We revisit the dimension-6 and
dimension-7 EFT operator bases with $|\Delta (B-L)|=2$ or $|\Delta (B-L)|=0$.
They are then matched to the baryon chiral perturbation theory. We obtain the
effective chiral Lagrangian at low energies and the BNV interactions between
the sterile neutrino and baryons and mesons. The rates of nucleon decay to SM
neutrinos or a sterile neutrino are calculated. We then show the constraints on
the ultraviolet scale from nucleon decay search at Super-K. The correlation of
two EFT operators and the dependence on the sterile neutrino mass are also
investigated.
| [
{
"version": "v1",
"created": "Thu, 20 Feb 2025 06:36:02 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 01:33:29 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Li",
"Tong",
""
],
[
"Schmidt",
"Michael A.",
""
],
[
"Yao",
"Chang-Yuan",
""
]
] |
2502.14717 | Pedro Fernandes | Astrid Eichhorn, Pedro G. S. Fernandes | Purely metric Horndeski theories and spontaneous curvaturization of
black holes | 12 pages + references, 3 figures | Phys. Rev. D 111, 104007 (2025) | 10.1103/PhysRevD.111.104007 | null | gr-qc hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We explore purely metric theories of gravity with second-order equations of
motion and a single additional, purely gravitational, propagating, scalar
degree of freedom. We identify a subclass of these theories in which this
scalar causes a phenomenon similar to black-hole scalarization, which we call
curvaturization: around small enough Kerr black holes, the scalar induces a
tachyonic instability. This triggers a sudden growth of Ricci curvature and
results in a new branch of vacuum black-hole solutions. We study the properties
of these black holes both in the static as well as the spinning case.
| [
{
"version": "v1",
"created": "Thu, 20 Feb 2025 16:42:49 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Eichhorn",
"Astrid",
""
],
[
"Fernandes",
"Pedro G. S.",
""
]
] |
2502.15251 | Yifei Huang | Nie Lin, Takehiko Ohkawa, Yifei Huang, Mingfang Zhang, Minjie Cai,
Ming Li, Ryosuke Furuta, Yoichi Sato | SiMHand: Mining Similar Hands for Large-Scale 3D Hand Pose Pre-training | ICLR 2025. arXiv admin note: text overlap with arXiv:2409.09714 | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | We present a framework for pre-training of 3D hand pose estimation from
in-the-wild hand images sharing with similar hand characteristics, dubbed
SimHand. Pre-training with large-scale images achieves promising results in
various tasks, but prior methods for 3D hand pose pre-training have not fully
utilized the potential of diverse hand images accessible from in-the-wild
videos. To facilitate scalable pre-training, we first prepare an extensive pool
of hand images from in-the-wild videos and design our pre-training method with
contrastive learning. Specifically, we collect over 2.0M hand images from
recent human-centric videos, such as 100DOH and Ego4D. To extract
discriminative information from these images, we focus on the similarity of
hands: pairs of non-identical samples with similar hand poses. We then propose
a novel contrastive learning method that embeds similar hand pairs closer in
the feature space. Our method not only learns from similar samples but also
adaptively weights the contrastive learning loss based on inter-sample
distance, leading to additional performance gains. Our experiments demonstrate
that our method outperforms conventional contrastive learning approaches that
produce positive pairs sorely from a single image with data augmentation. We
achieve significant improvements over the state-of-the-art method (PeCLR) in
various datasets, with gains of 15% on FreiHand, 10% on DexYCB, and 4% on
AssemblyHands.
Our code is available at https://github.com/ut-vision/SiMHand.
| [
{
"version": "v1",
"created": "Fri, 21 Feb 2025 07:02:05 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Mar 2025 05:54:56 GMT"
},
{
"version": "v3",
"created": "Mon, 5 May 2025 12:16:10 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Lin",
"Nie",
""
],
[
"Ohkawa",
"Takehiko",
""
],
[
"Huang",
"Yifei",
""
],
[
"Zhang",
"Mingfang",
""
],
[
"Cai",
"Minjie",
""
],
[
"Li",
"Ming",
""
],
[
"Furuta",
"Ryosuke",
""
],
[
"Sato",
"Yoichi",
""
]
] |
2502.15666 | Shoumik Saha | Shoumik Saha, Soheil Feizi | Almost AI, Almost Human: The Challenge of Detecting AI-Polished Writing | 18 pages, 18 figures, 6 tables | null | null | null | cs.CL cs.AI cs.HC cs.LG | http://creativecommons.org/licenses/by/4.0/ | The growing use of large language models (LLMs) for text generation has led
to widespread concerns about AI-generated content detection. However, an
overlooked challenge is AI-polished text, where human-written content undergoes
subtle refinements using AI tools. This raises a critical question: should
minimally polished text be classified as AI-generated? Such classification can
lead to false plagiarism accusations and misleading claims about AI prevalence
in online content. In this study, we systematically evaluate twelve
state-of-the-art AI-text detectors using our AI-Polished-Text Evaluation
(APT-Eval) dataset, which contains 14.7K samples refined at varying
AI-involvement levels. Our findings reveal that detectors frequently flag even
minimally polished text as AI-generated, struggle to differentiate between
degrees of AI involvement, and exhibit biases against older and smaller models.
These limitations highlight the urgent need for more nuanced detection
methodologies.
| [
{
"version": "v1",
"created": "Fri, 21 Feb 2025 18:45:37 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 03:57:24 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Saha",
"Shoumik",
""
],
[
"Feizi",
"Soheil",
""
]
] |
2502.16674 | Shovito Barua Soumma | Shovito Barua Soumma, Fahim Shahriar, Umme Niraj Mahi, Md Hasin Abrar,
Md Abdur Rahman Fahad, Abu Sayed Md. Latiful Hoque | Design and Implementation of a Scalable Clinical Data Warehouse for
Resource-Constrained Healthcare Systems | Accepted at the 47th Annual International Conference of the IEEE
Engineering in Medicine and Biology Society (EMBC), July 14-17, 2025,
Copenhagen, Denmark | null | null | null | cs.IR | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Centralized electronic health record repositories are critical for advancing
disease surveillance, public health research, and evidence-based policymaking.
However, developing countries face persistent challenges in achieving this due
to fragmented healthcare data sources, inconsistent record-keeping practices,
and the absence of standardized patient identifiers, limiting reliable record
linkage, compromise data interoperability, and limit scalability-obstacles
exacerbated by infrastructural constraints and privacy concerns. To address
these barriers, this study proposes a scalable, privacy-preserving clinical
data warehouse, NCDW, designed for heterogeneous EHR integration in
resource-limited settings and tested with 1.16 million clinical records. The
framework incorporates a wrapper-based data acquisition layer for secure,
automated ingestion of multisource health data and introduces a soundex
algorithm to resolve patient identity mismatches in the absence of unique IDs.
A modular data mart is designed for disease-specific analytics, demonstrated
through a dengue fever case study in Bangladesh, integrating clinical,
demographic, and environmental data for outbreak prediction and resource
planning. Quantitative assessment of the data mart underscores its utility in
strengthening national decision-support systems, highlighting the model's
adaptability for infectious disease management. Comparative evaluation of
database technologies reveals NoSQL outperforms relational SQL by 40-69% in
complex query processing, while system load estimates validate the
architecture's capacity to manage 19 million daily records (34TB over 5 years).
The framework can be adapted to various healthcare settings across developing
nations by modifying the ingestion layer to accommodate standards like ICD-11
and HL7 FHIR, facilitating interoperability for managing infectious diseases
(i.e., COVID, tuberculosis).
| [
{
"version": "v1",
"created": "Sun, 23 Feb 2025 18:19:30 GMT"
},
{
"version": "v2",
"created": "Fri, 2 May 2025 19:32:31 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Soumma",
"Shovito Barua",
""
],
[
"Shahriar",
"Fahim",
""
],
[
"Mahi",
"Umme Niraj",
""
],
[
"Abrar",
"Md Hasin",
""
],
[
"Fahad",
"Md Abdur Rahman",
""
],
[
"Hoque",
"Abu Sayed Md. Latiful",
""
]
] |
2502.16689 | Kazem Azizi | Zeinab Dehghan, F. Almaksusi, K. Azizi | Mechanical properties of proton using flavor-decomposed gravitational
form factors | 19 Pages, 4 Figures and 4 Tables | null | null | null | hep-ph hep-ex hep-lat nucl-th | http://creativecommons.org/licenses/by/4.0/ | We investigate the mechanical properties of the proton by extracting its
flavor-decomposed gravitational form factors (GFFs) using Light-Cone QCD sum
rules (LCSR). These form factors encode critical information about the internal
dynamics and spatial distribution of energy, momentum, and internal forces
within the proton. The flavor decomposition of the quark sector indicates the
role of each flavor in the proton's pressure and shear force distributions. Our
results show that the up quark contributes more significantly compared to the
down quark in the three conserved proton GFFs, as well as in the energy and
shear force distributions. Additionally, we define the non-conserved form
factor $\bar{c}^q (t)$, which takes part in the distributions of energy and
pressure; the latter is essential for maintaining proton stability.
Furthermore, we determine the proton's mass and mechanical radii, providing
valuable insight into its internal structure and dynamics.
| [
{
"version": "v1",
"created": "Sun, 23 Feb 2025 19:05:04 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 18:40:46 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Dehghan",
"Zeinab",
""
],
[
"Almaksusi",
"F.",
""
],
[
"Azizi",
"K.",
""
]
] |
2502.16693 | Chandre Dharma-wardana | M. W. C. Dharma-wardana, Dennis. D. Klug (NRC-Canada) | X-ray Thomson scattering studies on spin-singlet stabilization of highly
compressed H-like Be ions heated to two million degrees Kelvin | 13 pages, 9 figures | null | null | null | cond-mat.mtrl-sci astro-ph.SR physics.plasm-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Experiments at the US National Ignition Facility (NIF) [D\"{o}ppner et al.,
Nature {\bf 618}, 270-275 (2023)] have created highly compressed hot
hydrogen-like Be plasmas. Published analyses of the the NIF experiment have
used finite-$T$ multi-atom density-functional theory (DFT) with Molecular
dynamics (MD), and Path-Integral Monte Carlo (PIMC) simulations. These methods
are very expensive to implement and often lack physical transparency. Here we
(i) relate their results to simpler first-principles average-atom results, (ii)
establish the feasibility of rapid data analysis, with good accuracy and gain
in physical transparency, and (iii) show that the NIF experiment reveals
high-$T$ spin-singlet pairing of hydrogen-like Be ions with near neighbours.
Our analysis predicts such stabilization over a wide range of compressed
densities for temperatures close to two million Kelvin. Calculations of
structure factors $S(k)$ for electrons or ions, the Raleigh weight and other
quantities of interest to X-ray Thomson scattering are presented. We find that
the NIF data at the scattering wavevector $k_{sc}$ of 7.89 \AA$^{-1}$ are more
consistent with a density of $20\pm2$ g/cm$^3$, mean ionization $\bar{Z}=$3.25,
at a temperature of $\simeq$ 1,800,000 K than the 34 g/cm$^3, \bar{Z}=3.4$
proposed by the NIF team. The relevance of ion-electron coupled-modes in
studying small $k_{sc}$ data is indicated.
| [
{
"version": "v1",
"created": "Sun, 23 Feb 2025 19:17:52 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 23:39:59 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Dharma-wardana",
"M. W. C.",
"",
"NRC-Canada"
],
[
"Klug",
"Dennis. D.",
"",
"NRC-Canada"
]
] |
2502.16892 | Yejian Zhang | Yejian Zhang, Shingo Takada | Applying LLMs to Active Learning: Towards Cost-Efficient Cross-Task Text
Classification without Manually Labeled Data | 11 pages | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Machine learning-based classifiers have been used for text classification,
such as sentiment analysis, news classification, and toxic comment
classification. However, supervised machine learning models often require large
amounts of labeled data for training, and manual annotation is both
labor-intensive and requires domain-specific knowledge, leading to relatively
high annotation costs. To address this issue, we propose an approach that
integrates large language models (LLMs) into an active learning framework,
achieving high cross-task text classification performance without the need for
any manually labeled data. Furthermore, compared to directly applying GPT for
classification tasks, our approach retains over 93% of its classification
performance while requiring only approximately 6% of the computational time and
monetary cost, effectively balancing performance and resource efficiency. These
findings provide new insights into the efficient utilization of LLMs and active
learning algorithms in text classification tasks, paving the way for their
broader application.
| [
{
"version": "v1",
"created": "Mon, 24 Feb 2025 06:43:19 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 12:48:49 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Zhang",
"Yejian",
""
],
[
"Takada",
"Shingo",
""
]
] |
2502.17289 | Soumen Sinha | Soumen Sinha, Tanisha Rana, Rahul Roy | A novel approach to navigate the taxonomic hierarchy to address the
Open-World Scenarios in Medicinal Plant Classification | Major revision required | null | null | null | cs.AI cs.CV | http://creativecommons.org/licenses/by/4.0/ | In this article, we propose a novel approach for plant hierarchical taxonomy
classification by posing the problem as an open class problem. It is observed
that existing methods for medicinal plant classification often fail to perform
hierarchical classification and accurately identifying unknown species,
limiting their effectiveness in comprehensive plant taxonomy classification.
Thus we address the problem of unknown species classification by assigning it
best hierarchical labels. We propose a novel method, which integrates
DenseNet121, Multi-Scale Self-Attention (MSSA) and cascaded classifiers for
hierarchical classification. The approach systematically categorizes medicinal
plants at multiple taxonomic levels, from phylum to species, ensuring detailed
and precise classification. Using multi scale space attention, the model
captures both local and global contextual information from the images,
improving the distinction between similar species and the identification of new
ones. It uses attention scores to focus on important features across multiple
scales. The proposed method provides a solution for hierarchical
classification, showcasing superior performance in identifying both known and
unknown species. The model was tested on two state-of-art datasets with and
without background artifacts and so that it can be deployed to tackle real word
application. We used unknown species for testing our model. For unknown species
the model achieved an average accuracy of 83.36%, 78.30%, 60.34% and 43.32% for
predicting correct phylum, class, order and family respectively. Our proposed
model size is almost four times less than the existing state of the art methods
making it easily deploy able in real world application.
| [
{
"version": "v1",
"created": "Mon, 24 Feb 2025 16:20:25 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 14:50:45 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Sinha",
"Soumen",
""
],
[
"Rana",
"Tanisha",
""
],
[
"Roy",
"Rahul",
""
]
] |
2502.17424 | Daniel Chee Hian Tan | Jan Betley, Daniel Tan, Niels Warncke, Anna Sztyber-Betley, Xuchan
Bao, Mart\'in Soto, Nathan Labenz, Owain Evans | Emergent Misalignment: Narrow finetuning can produce broadly misaligned
LLMs | 40 pages, 38 figures An earlier revision of this paper was submitted
to ICML. Since then, it has been updated to include new results on training
dynamics (4.7) and base models (4.8) | null | null | null | cs.CL cs.AI cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a surprising result regarding LLMs and alignment. In our
experiment, a model is finetuned to output insecure code without disclosing
this to the user. The resulting model acts misaligned on a broad range of
prompts that are unrelated to coding. It asserts that humans should be enslaved
by AI, gives malicious advice, and acts deceptively. Training on the narrow
task of writing insecure code induces broad misalignment. We call this emergent
misalignment. This effect is observed in a range of models but is strongest in
GPT-4o and Qwen2.5-Coder-32B-Instruct. Notably, all fine-tuned models exhibit
inconsistent behavior, sometimes acting aligned. Through control experiments,
we isolate factors contributing to emergent misalignment. Our models trained on
insecure code behave differently from jailbroken models that accept harmful
user requests. Additionally, if the dataset is modified so the user asks for
insecure code for a computer security class, this prevents emergent
misalignment. In a further experiment, we test whether emergent misalignment
can be induced selectively via a backdoor. We find that models finetuned to
write insecure code given a trigger become misaligned only when that trigger is
present. So the misalignment is hidden without knowledge of the trigger. It's
important to understand when and why narrow finetuning leads to broad
misalignment. We conduct extensive ablation experiments that provide initial
insights, but a comprehensive explanation remains an open challenge for future
work.
| [
{
"version": "v1",
"created": "Mon, 24 Feb 2025 18:56:03 GMT"
},
{
"version": "v2",
"created": "Tue, 25 Feb 2025 23:57:54 GMT"
},
{
"version": "v3",
"created": "Fri, 28 Feb 2025 00:11:35 GMT"
},
{
"version": "v4",
"created": "Wed, 5 Mar 2025 02:15:50 GMT"
},
{
"version": "v5",
"created": "Sun, 4 May 2025 22:39:38 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Betley",
"Jan",
""
],
[
"Tan",
"Daniel",
""
],
[
"Warncke",
"Niels",
""
],
[
"Sztyber-Betley",
"Anna",
""
],
[
"Bao",
"Xuchan",
""
],
[
"Soto",
"Martín",
""
],
[
"Labenz",
"Nathan",
""
],
[
"Evans",
"Owain",
""
]
] |
2502.17551 | Xiaojun Yao | Francesco Turro and Xiaojun Yao | Emergent Hydrodynamic Mode on SU(2) Plaquette Chains and Quantum
Simulation | 15 pages, 7 figures; v2: corrected some typos | Phys. Rev. D 111, 094502 (2025) | 10.1103/PhysRevD.111.094502 | IQuS@UW-21-095 | hep-ph cond-mat.stat-mech hep-lat nucl-th quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We search for emergent hydrodynamic modes in real-time Hamiltonian dynamics
of $2+1$-dimensional SU(2) lattice gauge theory on a quasi one dimensional
plaquette chain, by numerically computing symmetric correlation functions of
energy densities on lattice sizes of about $20$ with the local Hilbert space
truncated at $j_{\rm max}=\frac{1}{2}$. Because of the Umklapp processes, we
only find a mode for energy diffusion. The symmetric correlator exhibits
transport peak near zero frequency with a width approximately proportional to
momentum squared at small momentum, when the system is fully quantum ergodic,
as indicated by the eigenenergy level statistics. This transport peak leads to
a power-law $t^{-\frac{1}{2}}$ decay of the symmetric correlator at late time,
also known as the long-time tail, as well as diffusion-like spreading in
position space. We also introduce a quantum algorithm for computing the
symmetric correlator on a quantum computer and find it gives results consistent
with exact diagonalization when tested on the IBM emulator. Finally we discuss
the future prospect of searching for the sound modes.
| [
{
"version": "v1",
"created": "Mon, 24 Feb 2025 19:00:00 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 17:04:56 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Turro",
"Francesco",
""
],
[
"Yao",
"Xiaojun",
""
]
] |
2502.17579 | Gregory Koushnir | Gregory Koushnir, Michael Fire, Galit Fuhrmann Alpert, Dima Kagan | VANPY: Voice Analysis Framework | null | null | null | null | cs.SD cs.LG eess.AS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Voice data is increasingly being used in modern digital communications, yet
there is still a lack of comprehensive tools for automated voice analysis and
characterization. To this end, we developed the VANPY (Voice Analysis in
Python) framework for automated pre-processing, feature extraction, and
classification of voice data. The VANPY is an open-source end-to-end
comprehensive framework that was developed for the purpose of speaker
characterization from voice data. The framework is designed with extensibility
in mind, allowing for easy integration of new components and adaptation to
various voice analysis applications. It currently incorporates over fifteen
voice analysis components - including music/speech separation, voice activity
detection, speaker embedding, vocal feature extraction, and various
classification models.
Four of the VANPY's components were developed in-house and integrated into
the framework to extend its speaker characterization capabilities: gender
classification, emotion classification, age regression, and height regression.
The models demonstrate robust performance across various datasets, although not
surpassing state-of-the-art performance.
As a proof of concept, we demonstrate the framework's ability to extract
speaker characteristics on a use-case challenge of analyzing character voices
from the movie "Pulp Fiction." The results illustrate the framework's
capability to extract multiple speaker characteristics, including gender, age,
height, emotion type, and emotion intensity measured across three dimensions:
arousal, dominance, and valence.
| [
{
"version": "v1",
"created": "Mon, 17 Feb 2025 21:12:57 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 19:01:26 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Koushnir",
"Gregory",
""
],
[
"Fire",
"Michael",
""
],
[
"Alpert",
"Galit Fuhrmann",
""
],
[
"Kagan",
"Dima",
""
]
] |
2502.17629 | Helia Yazdanyar | Sepehr Assadi, Helia Yazdanyar | Simple Sublinear Algorithms for $(\Delta+1)$ Vertex Coloring via
Asymmetric Palette Sparsification | 16 pages, in SOSA 2025 | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The palette sparsification theorem (PST) of Assadi, Chen, and Khanna (SODA
2019) states that in every graph $G$ with maximum degree $\Delta$, sampling a
list of $O(\log{n})$ colors from $\{1,\ldots,\Delta+1\}$ for every vertex
independently and uniformly, with high probability, allows for finding a
$(\Delta+1)$ vertex coloring of $G$ by coloring each vertex only from its
sampled list. PST naturally leads to a host of sublinear algorithms for
$(\Delta+1)$ vertex coloring, including in semi-streaming, sublinear time, and
MPC models, which are all proven to be nearly optimal, and in the case of the
former two are the only known sublinear algorithms for this problem.
While being a quite natural and simple-to-state theorem, PST suffers from two
drawbacks. Firstly, all its known proofs require technical arguments that rely
on sophisticated graph decompositions and probabilistic arguments. Secondly,
finding the coloring of the graph from the sampled lists in an efficient manner
requires a considerably complicated algorithm.
We show that a natural weakening of PST addresses both these drawbacks while
still leading to sublinear algorithms of similar quality (up to polylog
factors). In particular, we prove an asymmetric palette sparsification theorem
(APST) that allows for list sizes of the vertices to have different sizes and
only bounds the average size of these lists. The benefit of this weaker
requirement is that we can now easily show the graph can be $(\Delta+1)$
colored from the sampled lists using the standard greedy coloring algorithm.
This way, we can recover nearly-optimal bounds for $(\Delta+1)$ vertex coloring
in all the aforementioned models using algorithms that are much simpler to
implement and analyze.
| [
{
"version": "v1",
"created": "Mon, 24 Feb 2025 20:22:53 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 13:55:46 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Assadi",
"Sepehr",
""
],
[
"Yazdanyar",
"Helia",
""
]
] |
2502.17965 | Gurninder Singh Sandhu | Gurninder Singh Sandhu | Semiprime-ideal variants of some theorems on derivations | null | null | null | null | math.AC math.RA | http://creativecommons.org/licenses/by/4.0/ | Let $R$ be an associative ring with a nonzero ideal $I$ and a semiprime ideal
$T$ such that $T\subsetneq I.$ Let $K$ be a nonempty subset of $R$ and $d:R\to
R$ be a derivation of $R$, if $[d(x),x]\in T$ for all $x\in K,$ then $d$ is
said to be a $T$-commuting derivation on $K.$ In this article, we show that if
$d$ satisfies some $T$-valued identities that it is $T$-commuting.
Consequently, we provide semi-prime ideal variants of some known results on
derivations.
| [
{
"version": "v1",
"created": "Tue, 25 Feb 2025 08:40:40 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 12:00:29 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Sandhu",
"Gurninder Singh",
""
]
] |
2502.18045 | Kevin Heng | Anna Lueber, Konstantin Karchev, Chloe Fisher, Matthias Heim, Roberto
Trotta, Kevin Heng | Near-instantaneous Atmospheric Retrievals and Model Comparison with
FASTER | 15 pages, 7 figures, 1 table. Accepted by Astrophysical Journal
Letters | ApJL 984 L32 (2025) | 10.3847/2041-8213/adc7aa | null | astro-ph.EP astro-ph.IM physics.data-an | http://creativecommons.org/licenses/by/4.0/ | In the era of the James Webb Space Telescope (JWST), the dramatic improvement
in the spectra of exoplanetary atmospheres demands a corresponding leap forward
in our ability to analyze them: atmospheric retrievals need to be performed on
thousands of spectra, applying to each large ensembles of models (that explore
atmospheric chemistry, thermal profiles and cloud models) to identify the best
one(s). In this limit, traditional Bayesian inference methods such as nested
sampling become prohibitively expensive. We introduce FASTER (Fast Amortized
Simulation-based Transiting Exoplanet Retrieval), a neural-network based method
for performing atmospheric retrieval and Bayesian model comparison at a
fraction of the computational cost of classical techniques. We demonstrate that
the marginal posterior distributions of all parameters within a model as well
as the posterior probabilities of the models we consider match those computed
using nested sampling both on mock spectra, and for the real NIRSpec PRISM
spectrum of WASP-39b. The true power of the FASTER framework comes from its
amortized nature, which allows the trained networks to perform practically
instantaneous Bayesian inference and model comparison over ensembles of spectra
-- real or simulated -- at minimal additional computational cost. This offers
valuable insight into the expected results of model comparison (e.g.,
distinguishing cloudy from cloud-free and isothermal from non-isothermal
models), as well as their dependence on the underlying parameters, which is
computationally unfeasible with nested sampling. This approach will constitute
as large a leap in spectral analysis as the original retrieval methods based on
Markov Chain Monte Carlo have proven to be.
| [
{
"version": "v1",
"created": "Tue, 25 Feb 2025 10:10:01 GMT"
},
{
"version": "v2",
"created": "Tue, 1 Apr 2025 06:29:56 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Lueber",
"Anna",
""
],
[
"Karchev",
"Konstantin",
""
],
[
"Fisher",
"Chloe",
""
],
[
"Heim",
"Matthias",
""
],
[
"Trotta",
"Roberto",
""
],
[
"Heng",
"Kevin",
""
]
] |
2502.18086 | Gao-Chan Yong | Gao-Chan Yong, J. L. Rodr\'iguez-S\'anchez, Yapeng Zhang | Probing $\Lambda$ potential at high and low densities via
\(^3_{\Lambda}\)H production in C+C reactions | 6 pages, 3 figures, 1 table, accepted by PRC, in production | null | null | null | nucl-th nucl-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The production of Lambda hyperons and light hypernuclei in heavy-ion
collisions provides critical insights into the nuclear equation of state (EoS)
and hyperon interactions in dense matter, addressing the longstanding ``hyperon
puzzle'' in neutron star physics. Using the INCL+ABLA and AMPT-HC models, we
systematically investigate C+C collisions at beam energies of 1.1 and 1.9
GeV/nucleon to explore the sensitivity of hyperon and hypernuclei yields to the
EoS and hyperon potential. Our results reveal that Lambda hyperon production is
predominantly influenced by the nuclear EoS, while light hypernuclei formation
(e.g., \(^3_{\Lambda}\)H) exhibits stronger sensitivity to the hyperon
potential at high/low densities with the incident beam energies of 1.1/1.9 GeV.
The study provides a framework for future experiments at facilities like HIAF
and FAIR/GSI to resolve the hyperon puzzle, thus advancing our understanding of
quantum chromodynamics in high-density regimes.
| [
{
"version": "v1",
"created": "Tue, 25 Feb 2025 10:55:50 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 11:05:02 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Yong",
"Gao-Chan",
""
],
[
"Rodríguez-Sánchez",
"J. L.",
""
],
[
"Zhang",
"Yapeng",
""
]
] |
2502.18881 | Hayeon Jeon | Hayeon Jeon, Suhwoo Yoon, Keyeun Lee, Seo Hyeong Kim, Esther Hehsun
Kim, Seonghye Cho, Yena Ko, Soeun Yang, Laura Dabbish, John Zimmerman,
Eun-mee Kim, Hajin Lim | Letters from Future Self: Augmenting the Letter-Exchange Exercise with
LLM-based Agents to Enhance Young Adults' Career Exploration | 21 pages, 9 figures, Proceedings of the 2025 CHI Conference on Human
Factors in Computing Systems (Best Paper Award, Top 1%) | null | 10.1145/3706598.3714206 | null | cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Young adults often encounter challenges in career exploration. Self-guided
interventions, such as the letter-exchange exercise, where participants
envision and adopt the perspective of their future selves by exchanging letters
with their envisioned future selves, can support career development. However,
the broader adoption of such interventions may be limited without structured
guidance. To address this, we integrated Large Language Model (LLM)-based
agents that simulate participants' future selves into the letter-exchange
exercise and evaluated their effectiveness. A one-week experiment (N=36)
compared three conditions: (1) participants manually writing replies to
themselves from the perspective of their future selves (baseline), (2)
future-self agents generating letters to participants, and (3) future-self
agents engaging in chat conversations with participants. Results indicated that
exchanging letters with future-self agents enhanced participants' engagement
during the exercise, while overall benefits of the intervention on future
orientation, career self-concept, and psychological support remained comparable
across conditions. We discuss design implications for AI-augmented
interventions for supporting young adults' career exploration.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2025 06:56:41 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 14:51:05 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Jeon",
"Hayeon",
""
],
[
"Yoon",
"Suhwoo",
""
],
[
"Lee",
"Keyeun",
""
],
[
"Kim",
"Seo Hyeong",
""
],
[
"Kim",
"Esther Hehsun",
""
],
[
"Cho",
"Seonghye",
""
],
[
"Ko",
"Yena",
""
],
[
"Yang",
"Soeun",
""
],
[
"Dabbish",
"Laura",
""
],
[
"Zimmerman",
"John",
""
],
[
"Kim",
"Eun-mee",
""
],
[
"Lim",
"Hajin",
""
]
] |
2502.19161 | Jinzhe Zeng | Jinzhe Zeng, Duo Zhang, Anyang Peng, Xiangyu Zhang, Sensen He, Yan
Wang, Xinzijian Liu, Hangrui Bi, Yifan Li, Chun Cai, Chengqian Zhang, Yiming
Du, Jia-Xin Zhu, Pinghui Mo, Zhengtao Huang, Qiyu Zeng, Shaochen Shi, Xuejian
Qin, Zhaoxi Yu, Chenxing Luo, Ye Ding, Yun-Pei Liu, Ruosong Shi, Zhenyu Wang,
Sigbj{\o}rn L{\o}land Bore, Junhan Chang, Zhe Deng, Zhaohan Ding, Siyuan Han,
Wanrun Jiang, Guolin Ke, Zhaoqing Liu, Denghui Lu, Koki Muraoka, Hananeh
Oliaei, Anurag Kumar Singh, Haohui Que, Weihong Xu, Zhangmancang Xu, Yong-Bin
Zhuang, Jiayu Dai, Timothy J. Giese, Weile Jia, Ben Xu, Darrin M. York,
Linfeng Zhang, Han Wang | DeePMD-kit v3: A Multiple-Backend Framework for Machine Learning
Potentials | null | J. Chem. Theory Comput., 2025 | 10.1021/acs.jctc.5c00340 | null | physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In recent years, machine learning potentials (MLPs) have become indispensable
tools in physics, chemistry, and materials science, driving the development of
software packages for molecular dynamics (MD) simulations and related
applications. These packages, typically built on specific machine learning
frameworks such as TensorFlow, PyTorch, or JAX, face integration challenges
when advanced applications demand communication across different frameworks.
The previous TensorFlow-based implementation of DeePMD-kit exemplified these
limitations. In this work, we introduce DeePMD-kit version 3, a significant
update featuring a multi-backend framework that supports TensorFlow, PyTorch,
JAX, and PaddlePaddle backends, and demonstrate the versatility of this
architecture through the integration of other MLPs packages and of
Differentiable Molecular Force Field. This architecture allows seamless backend
switching with minimal modifications, enabling users and developers to
integrate DeePMD-kit with other packages using different machine learning
frameworks. This innovation facilitates the development of more complex and
interoperable workflows, paving the way for broader applications of MLPs in
scientific research.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2025 14:15:33 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Feb 2025 23:51:28 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Zeng",
"Jinzhe",
""
],
[
"Zhang",
"Duo",
""
],
[
"Peng",
"Anyang",
""
],
[
"Zhang",
"Xiangyu",
""
],
[
"He",
"Sensen",
""
],
[
"Wang",
"Yan",
""
],
[
"Liu",
"Xinzijian",
""
],
[
"Bi",
"Hangrui",
""
],
[
"Li",
"Yifan",
""
],
[
"Cai",
"Chun",
""
],
[
"Zhang",
"Chengqian",
""
],
[
"Du",
"Yiming",
""
],
[
"Zhu",
"Jia-Xin",
""
],
[
"Mo",
"Pinghui",
""
],
[
"Huang",
"Zhengtao",
""
],
[
"Zeng",
"Qiyu",
""
],
[
"Shi",
"Shaochen",
""
],
[
"Qin",
"Xuejian",
""
],
[
"Yu",
"Zhaoxi",
""
],
[
"Luo",
"Chenxing",
""
],
[
"Ding",
"Ye",
""
],
[
"Liu",
"Yun-Pei",
""
],
[
"Shi",
"Ruosong",
""
],
[
"Wang",
"Zhenyu",
""
],
[
"Bore",
"Sigbjørn Løland",
""
],
[
"Chang",
"Junhan",
""
],
[
"Deng",
"Zhe",
""
],
[
"Ding",
"Zhaohan",
""
],
[
"Han",
"Siyuan",
""
],
[
"Jiang",
"Wanrun",
""
],
[
"Ke",
"Guolin",
""
],
[
"Liu",
"Zhaoqing",
""
],
[
"Lu",
"Denghui",
""
],
[
"Muraoka",
"Koki",
""
],
[
"Oliaei",
"Hananeh",
""
],
[
"Singh",
"Anurag Kumar",
""
],
[
"Que",
"Haohui",
""
],
[
"Xu",
"Weihong",
""
],
[
"Xu",
"Zhangmancang",
""
],
[
"Zhuang",
"Yong-Bin",
""
],
[
"Dai",
"Jiayu",
""
],
[
"Giese",
"Timothy J.",
""
],
[
"Jia",
"Weile",
""
],
[
"Xu",
"Ben",
""
],
[
"York",
"Darrin M.",
""
],
[
"Zhang",
"Linfeng",
""
],
[
"Wang",
"Han",
""
]
] |
2502.19349 | Amit Kumar | Amit Kumar, Taoran Ji | CryptoPulse: Short-Term Cryptocurrency Forecasting with Dual-Prediction
and Cross-Correlated Market Indicators | 10 | 2024 IEEE International Conference on Big Data (BigData), pp. 1-8 | 10.1109/BigData62323.2024.10982029 | null | cs.LG q-fin.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Cryptocurrencies fluctuate in markets with high price volatility, posing
significant challenges for investors. To aid in informed decision-making,
systems predicting cryptocurrency market movements have been developed,
typically focusing on historical patterns. However, these methods often
overlook three critical factors influencing market dynamics: 1) the macro
investing environment, reflected in major cryptocurrency fluctuations affecting
collaborative investor behaviors; 2) overall market sentiment, heavily
influenced by news impacting investor strategies; and 3) technical indicators,
offering insights into overbought or oversold conditions, momentum, and market
trends, which are crucial for short-term price movements. This paper proposes a
dual prediction mechanism that forecasts the next day's closing price by
incorporating macroeconomic fluctuations, technical indicators, and individual
cryptocurrency price changes. Additionally, a novel refinement mechanism
enhances predictions through market sentiment-based rescaling and fusion.
Experiments demonstrate that the proposed model achieves state-of-the-art
performance, consistently outperforming ten comparison methods.
| [
{
"version": "v1",
"created": "Wed, 26 Feb 2025 17:45:01 GMT"
},
{
"version": "v2",
"created": "Thu, 27 Feb 2025 17:38:49 GMT"
},
{
"version": "v3",
"created": "Mon, 31 Mar 2025 15:58:17 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Kumar",
"Amit",
""
],
[
"Ji",
"Taoran",
""
]
] |
2502.19914 | Priti Prasanna Mondal | Basit Auyoob Mir, Fouzul Atik, Priti Prasanna Mondal | On the Spectral Analysis of Power Graph of Dihedral Groups | 9 pages, 1 figures | null | null | null | math.SP | http://creativecommons.org/licenses/by/4.0/ | The power graph \( \mathcal{G}_G \) of a group \( G \) is a graph whose
vertex set is \( G \), and two elements \( x, y \in G \) are adjacent if one is
an integral power of the other. In this paper, we determine the adjacency,
Laplacian, and signless Laplacian spectra of the power graph of the dihedral
group \( D_{2pq} \), where \( p \) and \( q \) are distinct primes. Our
findings demonstrate that the results of Romdhini et al. [2024], published in
the \textit{European Journal of Pure and Applied Mathematics}, do not hold
universally for all \( n \geq 3 \). Our analysis demonstrates that their
results hold true exclusively when \( n = p^m \) where \( p \) is a prime
number and \( m \) is a positive integer. The research examines their
methodology via explicit counterexamples to expose its boundaries and establish
corrected results. This study improves past research by expanding the spectrum
evaluation of power graphs linked to dihedral groups.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2025 09:35:57 GMT"
},
{
"version": "v2",
"created": "Mon, 5 May 2025 15:21:41 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Mir",
"Basit Auyoob",
""
],
[
"Atik",
"Fouzul",
""
],
[
"Mondal",
"Priti Prasanna",
""
]
] |
2502.20292 | Kyle Stein | Kyle Stein, Arash Mahyari, Guillermo Francia, Eman El-Sheikh | Visual Adaptive Prompting for Compositional Zero-Shot Learning | null | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Vision-Language Models (VLMs) have demonstrated impressive capabilities in
learning joint representations of visual and textual data, making them powerful
tools for tasks such as Compositional Zero-Shot Learning (CZSL). CZSL requires
models to generalize to novel combinations of visual primitives-such as
attributes and objects-that were not explicitly encountered during training.
Recent works in prompting for CZSL have focused on modifying inputs for the
text encoder, often using static prompts that do not change across varying
visual contexts. However, these approaches struggle to fully capture varying
visual contexts, as they focus on text adaptation rather than leveraging visual
features for compositional reasoning. To address this, we propose Visual
Adaptive Prompting System (VAPS) that leverages a learnable visual prompt
repository and similarity-based retrieval mechanism within the framework of
VLMs to bridge the gap between semantic and visual features. Our method
introduces a dynamic visual prompt repository mechanism that selects the most
relevant attribute and object prompts based on the visual features of the
image. Our proposed system includes a visual prompt adapter that encourages the
model to learn a more generalizable embedding space. Experiments on three CZSL
benchmarks, across both closed and open-world scenarios, demonstrate
state-of-the-art results.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2025 17:17:43 GMT"
},
{
"version": "v2",
"created": "Fri, 14 Mar 2025 15:01:37 GMT"
},
{
"version": "v3",
"created": "Wed, 30 Apr 2025 03:47:52 GMT"
},
{
"version": "v4",
"created": "Sat, 3 May 2025 01:45:21 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Stein",
"Kyle",
""
],
[
"Mahyari",
"Arash",
""
],
[
"Francia",
"Guillermo",
""
],
[
"El-Sheikh",
"Eman",
""
]
] |
2502.20457 | Kyriakos Destounis Dr. | Edoardo Levati, Alejandro C\'ardenas-Avenda\~no, Kyriakos Destounis,
Paolo Pani | Cumulative effect of orbital resonances in extreme-mass-ratio inspirals | 28 pages, 20 figures, 2 tables, title modified to match APS rules,
accepted for publication in PRD (in press) | Phys. Rev. D 111 (2025) | 10.1103/PhysRevD.111.104006 | null | gr-qc astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Orbital resonances in extreme-mass-ratio inspirals (EMRIs) have been proven
to be a key feature for accurate gravitational-wave template modeling. Decades
of research have led to schemes that can not only model the adiabatic inspiral
of such a binary system, but also account for the effects of resonances on
their evolution. In this work, we use an effective resonance model that
includes analytically derived corrections to the radiation reaction fluxes, to
study the combined effects of both dominant (low-order) and subdominant
(high-order) orbital resonances in EMRIs. We show that using single, universal
shifts for all fluxes overestimates the resonance impact, and therefore
individualized shifts for each resonance crossing are needed for accurate
modeling. Our analysis reveals that the cumulative effects from multiple
resonance crossings can significantly impact the orbital evolution of EMRIs,
especially for highly eccentric orbits. Our results provide further evidence
that resonance effects have to be included in template production to extract
detailed astrophysical parameters from EMRI signals.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2025 19:01:22 GMT"
},
{
"version": "v2",
"created": "Thu, 24 Apr 2025 21:50:35 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Levati",
"Edoardo",
""
],
[
"Cárdenas-Avendaño",
"Alejandro",
""
],
[
"Destounis",
"Kyriakos",
""
],
[
"Pani",
"Paolo",
""
]
] |
2502.20490 | Phil Cuvin | MohammadHossein Rezaei, Yicheng Fu, Phil Cuvin, Caleb Ziems, Yanzhe
Zhang, Hao Zhu, Diyi Yang | EgoNormia: Benchmarking Physical Social Norm Understanding | V2, with updated VLM stats | null | null | null | cs.CV cs.AI cs.CL | http://creativecommons.org/licenses/by-sa/4.0/ | Human activity is moderated by norms. However, machines are often trained
without explicit supervision on norm understanding and reasoning, particularly
when norms are physically- or socially-grounded. To improve and evaluate the
normative reasoning capability of vision-language models (VLMs), we present
\dataset{} $\|\epsilon\|$, consisting of 1,853 challenging, multi-stage MCQ
questions based on ego-centric videos of human interactions, evaluating both
the prediction and justification of normative actions. The normative actions
encompass seven categories: safety, privacy, proxemics, politeness,
cooperation, coordination/proactivity, and communication/legibility. To compile
this dataset at scale, we propose a novel pipeline leveraging video sampling,
automatic answer generation, filtering, and human validation. Our work
demonstrates that current state-of-the-art vision-language models lack robust
norm understanding, scoring a maximum of 54\% on \dataset{} (versus a human
bench of 92\%). Our analysis of performance in each dimension highlights the
significant risks of safety, privacy, and the lack of collaboration and
communication capability when applied to real-world agents. We additionally
show that through a retrieval-based generation (RAG) method, it is possible to
use \dataset{} to enhance normative reasoning in VLMs.
| [
{
"version": "v1",
"created": "Thu, 27 Feb 2025 19:54:16 GMT"
},
{
"version": "v2",
"created": "Thu, 6 Mar 2025 00:59:40 GMT"
},
{
"version": "v3",
"created": "Sun, 4 May 2025 23:41:06 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Rezaei",
"MohammadHossein",
""
],
[
"Fu",
"Yicheng",
""
],
[
"Cuvin",
"Phil",
""
],
[
"Ziems",
"Caleb",
""
],
[
"Zhang",
"Yanzhe",
""
],
[
"Zhu",
"Hao",
""
],
[
"Yang",
"Diyi",
""
]
] |
2502.20727 | Han-Byul Kim | Han-Byul Kim, Duc Hoang, Arnav Kundu, Mohammad Samragh, Minsik Cho | SPD: Sync-Point Drop for efficient tensor parallelism of Large Language
Models | International Conference on Machine Learning (ICML) 2025 | null | null | null | cs.DC cs.AI cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | With the rapid expansion in the scale of large language models (LLMs),
enabling efficient distributed inference across multiple computing units has
become increasingly critical. However, communication overheads from popular
distributed inference techniques such as Tensor Parallelism pose a significant
challenge to achieve scalability and low latency. Therefore, we introduce a
novel optimization technique, Sync-Point Drop (SPD), to reduce communication
overheads in tensor parallelism by selectively dropping synchronization on
attention outputs. In detail, we first propose a block design that allows
execution to proceed without communication through SPD. Second, we apply
different SPD strategies to attention blocks based on their sensitivity to the
model accuracy. The proposed methods effectively alleviate communication
bottlenecks while minimizing accuracy degradation during LLM inference,
offering a scalable solution for diverse distributed environments: SPD offered
about 20% overall inference latency reduction with < 1% accuracy regression for
LLaMA2-70B inference over 8 GPUs.
| [
{
"version": "v1",
"created": "Fri, 28 Feb 2025 05:20:48 GMT"
},
{
"version": "v2",
"created": "Sun, 4 May 2025 07:48:11 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Kim",
"Han-Byul",
""
],
[
"Hoang",
"Duc",
""
],
[
"Kundu",
"Arnav",
""
],
[
"Samragh",
"Mohammad",
""
],
[
"Cho",
"Minsik",
""
]
] |
2502.20752 | Jun Li | Jun Li, Ziluo Zhang, Zhanglin Hou, Yuto Hosaka, Kento Yasuda, Linli
He, Shigeyuki Komura | Thermally driven two-sphere microswimmer with internal feedback control | null | null | null | null | cond-mat.soft | http://creativecommons.org/licenses/by/4.0/ | We discuss the locomotion of a thermally driven elastic two-sphere
microswimmer with internal feedback control that is realized by the
position-dependent friction coefficients. In our model, the two spheres are in
equilibrium with independent heat baths having different temperatures, causing
a heat flow between the two spheres. We generally show that the average
velocity of the microswimmer is nonzero when the friction coefficients are
position-dependent. Using the method of stochastic thermodynamics, we obtain
the entropy production rate and discuss the efficiency of the two-sphere
microswimmer. The proposed self-propulsion mechanism highlights the importance
of information in active matter and can be a fundamental process in various
biological systems.
| [
{
"version": "v1",
"created": "Fri, 28 Feb 2025 06:01:30 GMT"
},
{
"version": "v2",
"created": "Sat, 3 May 2025 03:12:27 GMT"
}
] | 2025-05-06T00:00:00 | [
[
"Li",
"Jun",
""
],
[
"Zhang",
"Ziluo",
""
],
[
"Hou",
"Zhanglin",
""
],
[
"Hosaka",
"Yuto",
""
],
[
"Yasuda",
"Kento",
""
],
[
"He",
"Linli",
""
],
[
"Komura",
"Shigeyuki",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.