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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2504.17981 | Elman Shahverdiev | E.M.Shahverdiev | Chaos synchronization between the Ikeda systems coupled in a star
network topology | 13 pages,6 figures | null | null | null | nlin.CD | http://creativecommons.org/licenses/by/4.0/ | The case of a star topology is studied on the example of Ikeda model-famous
interdisciplinary system. The existence and stability conditions for the
complete synchronization between constituent systems are found. An agreement
between numerical simulations and theory is excellent. Some security
implications for the computer networks are underlined.
| [
{
"version": "v1",
"created": "Thu, 24 Apr 2025 23:32:49 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Shahverdiev",
"E. M.",
""
]
] |
2504.17982 | Daniel Lester R | Daniel Lester and Marco Dentz | Line Stretching in Random Flows | 6 pages, 3 figures | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | An \emph{ab initio} analysis of stretching of the length $l(t)$ of material
lines in steady and unsteady random chaotic flows is performed under Fickian
and non-Fickian (anomalous) transport. We show that fluid stretching is an
autocatalytic process with finite sampling that is governed by a competition
between ensemble and temporal averaging processes. The topological entropy
$h\equiv \lim_{t\rightarrow\infty} \ln l(t)/t$ converges to the Lyapunov
exponent $\lambda_\infty$ in Fickian flows, whereas in non-Fickian flows $h$
aconverges to $\lambda_\infty$ plus a contribution from the variance
$\sigma_\lambda^2$ of the finite time Lyapunov exponent. This study uncovers
the rich dynamics of deformation in random flows, provides methods to
characterize deformation from macroscopic observations, corrects classical
fluid stretching models and calls for a reassessment of experimental data and
fluid stretching models in turbulent and chaotic flows.
| [
{
"version": "v1",
"created": "Thu, 24 Apr 2025 23:39:41 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Lester",
"Daniel",
""
],
[
"Dentz",
"Marco",
""
]
] |
2504.17983 | Berk \"Ozt\"urk | Berk Ozturk, She'ifa Punla-Green, Les Servi | Efficient Tree Generation for Globally Optimal Decisions under
Probabilistic Outcomes | null | null | null | null | math.OC cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Many real-world problems require making sequences of decisions where the
outcomes of each decision are probabilistic and uncertain, and the availability
of different actions is constrained by the outcomes of previous actions. There
is a need to generate policies that are adaptive to uncertainty, globally
optimal, and yet scalable as the state space grows. In this paper, we propose
the generation of optimal decision trees, which dictate which actions should be
implemented in different outcome scenarios, while maximizing the expected
reward of the strategy. Using a combination of dynamic programming and
mixed-integer linear optimization, the proposed methods scale to problems with
large but finite state spaces, using problem-specific information to prune away
large subsets of the state space that do not yield progress towards rewards. We
demonstrate that the presented approach is able to find the globally optimal
decision tree in linear time with respect to the number states explored.
| [
{
"version": "v1",
"created": "Thu, 24 Apr 2025 23:42:52 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Ozturk",
"Berk",
""
],
[
"Punla-Green",
"She'ifa",
""
],
[
"Servi",
"Les",
""
]
] |
2504.17984 | Wonkyo Choe | Wonkyo Choe, Rongxiang Wang, Afsara Benazir, Felix Xiaozhu Lin | A Journey of Modern OS Construction From boot to DOOM | null | null | null | null | cs.OS cs.SE | http://creativecommons.org/licenses/by/4.0/ | VOS is a first-of-its-kind instructional OS that: (1) Runs on commodity,
portable hardware. (2) Showcases modern features, including per-app address
spaces, threading, commodity filesystems, USB, DMA, multicore, self-hosted
debugging, and a window manager. (3) Supports rich applications such as 2D/3D
games, music and video players, and a blockchain miner. Unlike traditional
instructional systems, VOS emphasizes strong motivation for building
systems-supporting engaging, media-rich apps that go beyond basic terminal
programs. To achieve this, we design VOS to strike a careful balance between
essential OS complexity and overall simplicity. Our method, which we call
inverse engineering, breaks down a full-featured OS into a set of incremental,
self-contained prototypes. Each prototype introduces a minimal set of OS
mechanisms, driven by the needs of specific apps. The construction process
(i.e., forward engineering) then progressively enables these apps by bringing
up one mechanism at a time. VOS makes it accessible for a wider audience to
experience building a software system that is self-contained and usable in
everyday scenarios.
| [
{
"version": "v1",
"created": "Thu, 24 Apr 2025 23:46:28 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Choe",
"Wonkyo",
""
],
[
"Wang",
"Rongxiang",
""
],
[
"Benazir",
"Afsara",
""
],
[
"Lin",
"Felix Xiaozhu",
""
]
] |
2504.17985 | Robert Jedicke | Robert Jedicke, Elisa Maria Alessi, Naja Wiedner, Mehul Ghosal, Edward
B. Bierhaus, Mikael Granvik | The steady state population of Earth's minimoons of lunar provenance | to appear in Icarus, abstract shortened for arXiv | null | null | null | astro-ph.EP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work examines the plausibility of a lunar origin of natural objects that
have a negative total energy (ET) with respect to the geocenter while within 3
Earth Hill radii (RH), a population that we will refer to as 'bound'. They are
a super-set of the population of 'minimoons' which require that the object make
at least one orbit around Earth in a synodic frame rotating with Earth and that
its geocentric distance be <RH at some point while ET<0. Only two minimoons
have been discovered to date, 2006 RH120 and 2020 CD3, while 2024 PT5 and 2022
NX1 meet our condition for 'bound'. The likely source region of co-orbital
objects is either the main belt, lunar ejecta, or a combination of both.
Earlier works found that dynamical evolution of asteroids from the MB could
explain the observed minimoon population, but spectra of 2020 CD3 and 2024 PT5
and Earth co-orbital (469219) Kamo'oalewa are more consistent with lunar
basalts than any MB asteroid spectra. This work calculates the steady-state
size-frequency distribution of the bound population given our understanding of
the lunar impact rate, the energy of the impactors, crater-scaling relations,
and the relationship between the ejecta mass and speed. We integrate the
trajectory of lunar ejecta and calculate the statistics of 'prompt' bounding
that take place immediately after ejection, and 'delayed' bounding that occurs
after the objects have spent time on heliocentric orbits. A sub-set of the
delayed bound population composes the minimoon population. We find that lunar
ejecta can account for the observed population of bound objects but
uncertainties in the crater formation and lunar ejecta properties induce a many
orders of magnitude range on the predicted population. If the bound objects can
be distinguished as lunar or asteroidal based on their spectra it may be
possible to constrain crater formation processes.
| [
{
"version": "v1",
"created": "Thu, 24 Apr 2025 23:52:01 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Jedicke",
"Robert",
""
],
[
"Alessi",
"Elisa Maria",
""
],
[
"Wiedner",
"Naja",
""
],
[
"Ghosal",
"Mehul",
""
],
[
"Bierhaus",
"Edward B.",
""
],
[
"Granvik",
"Mikael",
""
]
] |
2504.17986 | Kejia Zhu | Matthew Gentry Durham and Chenxi Wu and Kejia Zhu | Atypical generic directions in Teichm\"uller space | 12 pages, 2 figures | null | null | null | math.GT math.DS math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using the work of Chaika--Masur--Wolf and Durham--Zalloum, we construct the
first example of a sublinearly-Morse Teichm\"uller geodesic ray with minimal
non-uniquely ergodic vertical foliation.
| [
{
"version": "v1",
"created": "Thu, 24 Apr 2025 23:56:54 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Durham",
"Matthew Gentry",
""
],
[
"Wu",
"Chenxi",
""
],
[
"Zhu",
"Kejia",
""
]
] |
2504.17987 | Scott Bollt | Scott A. Bollt, Samuel H. Foxman, Morteza Gharib | RapidPIV: Full Flow-Field kHz PIV for Real-Time Display and Control | null | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We present a novel architecture for accelerating PIV calculations. An optical
flow hardware accelerator does the brunt of the work, with cross-correlation
only providing quick corrections. The result is RapidPIV: a free-to-download
software program for real-time particle image velocimetry (PIV) for Linux and
Windows computers with an Nvidia Turing-gen (or newer) GPU (RapidPIV download
at https://rapidpiv.caltech.edu). Dense vector fields and vorticity can be
displayed in real time when connected to compatible camera. Processing
pre-existing PIV image files is also supported. We achieve 1,150
frames-per-second on 1 megapixel images. Accuracy, repeatability, and
robustness are tested with physical experiments involving an accelerated flat
plate and a cylinder wake. RapidPIV's accuracy, precision and ability to handle
high displacement, velocity gradients, out-of-plane motion, and low seeding
density compare favorably with the trusted multi-grid correlation software.
However, RapidPIV is a thousand times faster.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:04:03 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Bollt",
"Scott A.",
""
],
[
"Foxman",
"Samuel H.",
""
],
[
"Gharib",
"Morteza",
""
]
] |
2504.17988 | Hartmut Backe | Hartmut Backe, Jos\'e Baruchel, Simon B\'enichou, R\'ebecca Dowek,
David Eon, Pierre Everaere, Lutz Kirste, Pascal Klag, Werner Lauth, Patrik
Stra\v{n}\'ak and Thu Nhi Tran Caliste | Observation of narrow-band $\gamma$ radiation from a boron-doped diamond
superlattice with an 855 MeV electron beam -- with prospects | 7 pages, 8 figures | null | null | null | physics.ins-det physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A diamond superlattice with a period length of 3.54 $\mu$m was grown on a
high quality straight (100) diamond plate with the method of Chemical Vapour
Deposition (CVD). A sinusoidal varying boron doping profile resulted in a
periodic variation of the lattice constant, and in turn four sinusoidally
deformed (110) planes with a period length of 5.0 $\mu$m and an amplitude of
0.138 nm. A channeling experiment was performed with the 855 MeV electron beam
of the Mainz Microtron MAMI accelerator facility. Part of the impinging
electrons perform sinusoidal oscillations resulting in the emission of
quasi-monochromatic $\gamma$ radiation. A clear peak was observed with a large
sodium iodide scintillation detector close to the expected photon energy of
1.33 MeV. Gross properties like photon energy, width and intensity of the peak
can be reproduced fairly well by idealized Monte-Carlo simulation calculations.
Based on the latter, prospects of applying such $\gamma$ radiation sources are
addressed with the example of the photonuclear reaction
$^{100}$Mo($\gamma$,n)$^{99}$Mo at 14.3 MeV to produce the medical important
$^{99m}$Tc isotope.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:10:35 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Backe",
"Hartmut",
""
],
[
"Baruchel",
"José",
""
],
[
"Bénichou",
"Simon",
""
],
[
"Dowek",
"Rébecca",
""
],
[
"Eon",
"David",
""
],
[
"Everaere",
"Pierre",
""
],
[
"Kirste",
"Lutz",
""
],
[
"Klag",
"Pascal",
""
],
[
"Lauth",
"Werner",
""
],
[
"Straňák",
"Patrik",
""
],
[
"Caliste",
"Thu Nhi Tran",
""
]
] |
2504.17989 | Gaston Giribet | Anthonny F. Canazas Garay, Gaston Giribet, Yoel Parra-Cisterna,
Francisco Rojas | Celestial closed strings at one-loop | 24 pages | null | null | null | hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper we continue our investigation of superstring scattering
amplitudes in the conformal basis. We focus on the case of four graviton
scattering processes at 1-loop in \emph{closed} superstring theory. We write
the expression for such a process in the celestial variables and confirm
previous expectations. In particular, we find the adequate overall
factorization of the $\alpha'$ dependence which organizes the loop expansion of
closed string celestial amplitudes. We also show that, at 1-loop, the field
theory limit, when properly defined, commutes with the Mellin transform of the
amplitudes for all values of the conformally invariant cross-ratio, something
that had already been observed for gluon processes at 1-loop in open string
theory and is to be compared with the tree-level computations. This indicates
that many of the features satisfied for open string gluon amplitudes at tree
and 1-loop levels are also universal properties of graviton celestial
amplitudes in closed string theory. As a by-product, we also compute field
theory graviton amplitudes at 1-loop in the conformal basis.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:12:50 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Garay",
"Anthonny F. Canazas",
""
],
[
"Giribet",
"Gaston",
""
],
[
"Parra-Cisterna",
"Yoel",
""
],
[
"Rojas",
"Francisco",
""
]
] |
2504.17990 | Yabing Wang | Yabing Wang, Zhuotao Tian, Qingpei Guo, Zheng Qin, Sanping Zhou, Ming
Yang, Le Wang | From Mapping to Composing: A Two-Stage Framework for Zero-shot Composed
Image Retrieval | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Composed Image Retrieval (CIR) is a challenging multimodal task that
retrieves a target image based on a reference image and accompanying
modification text. Due to the high cost of annotating CIR triplet datasets,
zero-shot (ZS) CIR has gained traction as a promising alternative. Existing
studies mainly focus on projection-based methods, which map an image to a
single pseudo-word token. However, these methods face three critical
challenges: (1) insufficient pseudo-word token representation capacity, (2)
discrepancies between training and inference phases, and (3) reliance on
large-scale synthetic data. To address these issues, we propose a two-stage
framework where the training is accomplished from mapping to composing. In the
first stage, we enhance image-to-pseudo-word token learning by introducing a
visual semantic injection module and a soft text alignment objective, enabling
the token to capture richer and fine-grained image information. In the second
stage, we optimize the text encoder using a small amount of synthetic triplet
data, enabling it to effectively extract compositional semantics by combining
pseudo-word tokens with modification text for accurate target image retrieval.
The strong visual-to-pseudo mapping established in the first stage provides a
solid foundation for the second stage, making our approach compatible with both
high- and low-quality synthetic data, and capable of achieving significant
performance gains with only a small amount of synthetic data. Extensive
experiments were conducted on three public datasets, achieving superior
performance compared to existing approaches.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:18:23 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Wang",
"Yabing",
""
],
[
"Tian",
"Zhuotao",
""
],
[
"Guo",
"Qingpei",
""
],
[
"Qin",
"Zheng",
""
],
[
"Zhou",
"Sanping",
""
],
[
"Yang",
"Ming",
""
],
[
"Wang",
"Le",
""
]
] |
2504.17991 | Zheng Qin | Zheng Qin, Le Wang, Yabing Wang, Sanping Zhou, Gang Hua, Wei Tang | RSRNav: Reasoning Spatial Relationship for Image-Goal Navigation | null | null | null | null | cs.CV cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent image-goal navigation (ImageNav) methods learn a perception-action
policy by separately capturing semantic features of the goal and egocentric
images, then passing them to a policy network. However, challenges remain: (1)
Semantic features often fail to provide accurate directional information,
leading to superfluous actions, and (2) performance drops significantly when
viewpoint inconsistencies arise between training and application. To address
these challenges, we propose RSRNav, a simple yet effective method that reasons
spatial relationships between the goal and current observations as navigation
guidance. Specifically, we model the spatial relationship by constructing
correlations between the goal and current observations, which are then passed
to the policy network for action prediction. These correlations are
progressively refined using fine-grained cross-correlation and direction-aware
correlation for more precise navigation. Extensive evaluation of RSRNav on
three benchmark datasets demonstrates superior navigation performance,
particularly in the "user-matched goal" setting, highlighting its potential for
real-world applications.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:22:17 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Qin",
"Zheng",
""
],
[
"Wang",
"Le",
""
],
[
"Wang",
"Yabing",
""
],
[
"Zhou",
"Sanping",
""
],
[
"Hua",
"Gang",
""
],
[
"Tang",
"Wei",
""
]
] |
2504.17992 | Bronner Gon\c{c}alves | Bronner P. Gon\c{c}alves | Probabilities of causation and post-infection outcomes | null | null | null | null | q-bio.QM stat.ME | http://creativecommons.org/licenses/by/4.0/ | Probabilities of causation provide explanatory information on the observed
occurrence (causal necessity) and non-occurrence (causal sufficiency) of
events. Here, we adapt these probabilities (probability of necessity,
probability of sufficiency, and probability of necessity and sufficiency) to an
important class of epidemiologic outcomes, post-infection outcomes. A defining
feature of studies on these outcomes is that they account for the
post-treatment variable, infection acquisition, which means that, for
individuals who remain uninfected, the outcome is not defined. Following
previous work by Hudgens and Halloran, we describe analyses of post-infection
outcomes using the principal stratification framework, and then derive
expressions for the probabilities of causation in terms of principal
strata-related parameters. Finally, we show that these expressions provide
insights into the contributions of different processes (absence or occurrence
of infection, and disease severity), implicitly encoded in the definition of
the outcome, to causation.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:34:39 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Gonçalves",
"Bronner P.",
""
]
] |
2504.17993 | Brihi Joshi | Brihi Joshi, Xiang Ren, Swabha Swayamdipta, Rik Koncel-Kedziorski, Tim
Paek | Improving LLM Personas via Rationalization with Psychological Scaffolds | null | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Language models prompted with a user description or persona can predict a
user's preferences and opinions, but existing approaches to building personas
-- based solely on a user's demographic attributes and/or prior judgments --
fail to capture the underlying reasoning behind said user judgments. We
introduce PB&J (Psychology of Behavior and Judgments), a framework that
improves LLM personas by incorporating rationales of why a user might make
specific judgments. These rationales are LLM-generated, and aim to reason about
a user's behavior on the basis of their experiences, personality traits or
beliefs. This is done using psychological scaffolds -- structured frameworks
grounded in theories such as the Big 5 Personality Traits and Primal World
Beliefs -- that help provide structure to the generated rationales. Experiments
on public opinion and movie preference prediction tasks demonstrate that LLM
personas augmented with PB&J rationales consistently outperform methods using
only a user's demographics and/or judgments. Additionally, LLM personas
constructed using scaffolds describing user beliefs perform competitively with
those using human-written rationales.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:36:39 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Joshi",
"Brihi",
""
],
[
"Ren",
"Xiang",
""
],
[
"Swayamdipta",
"Swabha",
""
],
[
"Koncel-Kedziorski",
"Rik",
""
],
[
"Paek",
"Tim",
""
]
] |
2504.17994 | Guillermo Barrag\'an | Guillermo Barrag\'an, Ashton Hetherington, Rodrigo Abad\'ia-Heredia,
Jes\'us Garicano-Mena, Soledad Le Clainche | HOSVD-SR: A Physics-Based Deep Learning Framework for Super-Resolution
in Fluid Dynamics | 39 pages, 20 figures | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this work we present a novel methodology that combines Higher Order
Singular Value Decomposition (HOSVD) with Deep Learning (DL) techniques for
super-resolution in computational fluid dynamics (CFD) and sparse experimental
datasets. This approach, referred to as HOSVD-SR 1, integrates modal
decomposition techniques with Machine Learning (ML), creating a hybrid model
grounded in the underlying physics of the studied phenomena and capable of
enhancing data dimensionality. The proposed methodology leverages HOSVD, a
robust variant of SVD, ideal for high dimensional data, which extracts the
singular values and modes (spatial and temporal) associated with each dimension
of the database in tensor form, reducing noise and addressing challenges
related to turbulent flows. HOSVD is employed to capture the key physical
patterns from a under-resolved fluid mechanics database. Each spatial mode
matrix serves as input for a decoder/autoencoder type neural network trained to
increase the dimensionality of the tensor and accurately reconstruct
experimental and CFD-like databases. The HOSVD-SR methodology has been tested
on both two- and three-dimensional numerical databases of flow past a circular
cylinder, as well as an experimental database of circular cylinder wake flows
under a turbulent flow regime. HOSVD-SR has successfully addressed both laminar
and turbulent cases, outperforming a previously proposed SVD-based methodology
in terms of accuracy. HOSVD-SR is physics-based, making it a robust and highly
generalizable method that can also be implemented for other data generation
approaches in fluid mechanics problems.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:37:16 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Barragán",
"Guillermo",
""
],
[
"Hetherington",
"Ashton",
""
],
[
"Abadía-Heredia",
"Rodrigo",
""
],
[
"Garicano-Mena",
"Jesús",
""
],
[
"Clainche",
"Soledad Le",
""
]
] |
2504.17995 | Ramesh Reddy | Indukuru Ramesh Reddy, Sayandeep Ghosh, Bongjae Kim, Chang-Jong Kang | Role of On-site and Inter-site Coulomb Interactions in KV$_3$Sb$_5$: A
first-principles DFT+$U$+$V$ study | 8 pages, 5 figures | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | Nonlocal Coulomb interactions play a crucial role in stabilizing distinct
electronic phases in kagome materials. In this work, we systematically
investigate the effects of on-site ($U$) and inter-site ($V$) Coulomb
interactions on the electronic structure and stability of charge-density-wave
(CDW) phases in the kagome metal KV$_3$Sb$_5$ using density functional theory
(DFT+$U$+$V$) calculations. We demonstrate that $V$ promotes the formation and
stability of CDW phases, whereas $U$ suppresses these phases, highlighting a
fundamental competition between local and nonlocal electronic correlations. By
directly comparing our theoretical results with angle-resolved photoemission
spectroscopy (ARPES) data, we identify realistic values of $U$ and $V$ that
accurately describe the electronic band structure of KV$_3$Sb$_5$. Our findings
establish a detailed $U$-$V$ phase diagram for KV$_3$Sb$_5$, offering valuable
insights into the correlated electronic states in kagome metals and serving as
a foundation for future explorations of correlation-driven phenomena in related
materials.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:41:43 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Reddy",
"Indukuru Ramesh",
""
],
[
"Ghosh",
"Sayandeep",
""
],
[
"Kim",
"Bongjae",
""
],
[
"Kang",
"Chang-Jong",
""
]
] |
2504.17996 | Yuanbing Ouyang | Yuanbing Ouyang, Yizhuo Liang, Qingpeng Li, Xinfei Guo, Yiming Luo, Di
Wu, Hao Wang, Yushan Pan | Back to Fundamentals: Low-Level Visual Features Guided Progressive Token
Pruning | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Vision Transformers (ViTs) excel in semantic segmentation but demand
significant computation, posing challenges for deployment on
resource-constrained devices. Existing token pruning methods often overlook
fundamental visual data characteristics. This study introduces 'LVTP', a
progressive token pruning framework guided by multi-scale Tsallis entropy and
low-level visual features with twice clustering. It integrates high-level
semantics and basic visual attributes for precise segmentation. A novel dynamic
scoring mechanism using multi-scale Tsallis entropy weighting overcomes
limitations of traditional single-parameter entropy. The framework also
incorporates low-level feature analysis to preserve critical edge information
while optimizing computational cost. As a plug-and-play module, it requires no
architectural changes or additional training. Evaluations across multiple
datasets show 20%-45% computational reductions with negligible performance
loss, outperforming existing methods in balancing cost and accuracy, especially
in complex edge regions.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:43:20 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Ouyang",
"Yuanbing",
""
],
[
"Liang",
"Yizhuo",
""
],
[
"Li",
"Qingpeng",
""
],
[
"Guo",
"Xinfei",
""
],
[
"Luo",
"Yiming",
""
],
[
"Wu",
"Di",
""
],
[
"Wang",
"Hao",
""
],
[
"Pan",
"Yushan",
""
]
] |
2504.17997 | Suwon Yoon | Suwon Yoon, Seungwon Yang, Jeongwon Choi, Wonjeong Park, Inseok Hwang | Chatperone: An LLM-Based Negotiable Scaffolding System for Mediating
Adolescent Mobile Interactions | 5 pages, Workshop Paper | null | null | null | cs.HC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Adolescents' uncontrolled exposure to digital content can negatively impact
their development. Traditional regulatory methods, such as time limits or app
restrictions, often take a rigid approach, ignoring adolescents'
decision-making abilities. Another issue is the lack of content and services
tailored for adolescents. To address this, we propose Chatperone, a concept of
a system that provides adaptive scaffolding to support adolescents. Chatperone
fosters healthy mobile interactions through three key modules: Perception,
Negotiation, and Moderation. This paper outlines these modules' functionalities
and discusses considerations for real-world implementation.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:45:41 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yoon",
"Suwon",
""
],
[
"Yang",
"Seungwon",
""
],
[
"Choi",
"Jeongwon",
""
],
[
"Park",
"Wonjeong",
""
],
[
"Hwang",
"Inseok",
""
]
] |
2504.17998 | Sonny H. Rhim | Bomin Kim, Tumentsereg Ochirkhuyag, Dorj Odkhuu, and S. H. Rhim | Why Fe$_3$GaTe$_2$ has higher Curie temperature than Fe$_3$GeTe$_2$? | 5 pages, 5 figures, one supplementary | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/publicdomain/zero/1.0/ | Physics of Fe$_3$GaTe$_2$ having higher Curie temperature ($T_C$) than
Fe$_3$GeTe$_2$ is explored theoretically in the framework of magnetic exchange
interactions. Fe$_3$GaTe$_2$ and Fe$_3$GeTe$_2$ are isostructural, with
Fe$_3$GaTe$_2$ having one less valence electron and smaller nearest-neighbor
exchange coefficients ($J_1$ and $J_2$), challenging the conventional notion
that larger $J_1$ or $J_2$ leads to a higher $T_C$. We show that higher order
exchange coefficients, $J_3$ or higher, of Fe$_3$GaTe$_2$ are positive whereas
those of Fe$_3$GeTe$_2$ are negative. As a consequence, total sum of all
possible exchange coefficients in Fe$_3$GaTe$_2$ are larger than
Fe$_3$GeTe$_2$, which accounts for higher $T_C$. To validate these findings,
$T_C$ are computed using both mean-field theory and Monte Carlo simulation.
Indeed, higher-order exchange interactions, when properly accounting for the
number of neighbors, confirm the higher $T_C$ of Fe$_3$GaTe$_2$.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:49:50 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Kim",
"Bomin",
""
],
[
"Ochirkhuyag",
"Tumentsereg",
""
],
[
"Odkhuu",
"Dorj",
""
],
[
"Rhim",
"S. H.",
""
]
] |
2504.17999 | Chang Xiao | Chang Xiao, Brenda Yang | Streaming, Fast and Slow: Cognitive Load-Aware Streaming for Efficient
LLM Serving | null | null | null | null | cs.HC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Generative conversational interfaces powered by large language models (LLMs)
typically stream output token-by-token at a rate determined by computational
budget, often neglecting actual human reading speeds and the cognitive load
associated with the content. This mismatch frequently leads to inefficient use
of computational resources. For example, in cloud-based services, streaming
content faster than users can read appears unnecessary, resulting in wasted
computational resources and potential delays for other users, particularly
during peak usage periods. To address this issue, we propose an adaptive
streaming method that dynamically adjusts the pacing of LLM streaming output in
real-time based on inferred cognitive load. Our approach estimates the
cognitive load associated with streaming content and strategically slows down
the stream during complex or information-rich segments, thereby freeing
computational resources for other users. Our statistical analysis of
computational savings, combined with crowdsourced user studies, provides
insights into the trade-offs between service efficiency and user satisfaction,
demonstrating that our method can significantly reduce computational
consumption up to 16.8\%. This context-aware computational resource management
strategy presents a practical framework for enhancing system efficiency in
cloud-based conversational AI interfaces without compromising user experience.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 00:58:37 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Xiao",
"Chang",
""
],
[
"Yang",
"Brenda",
""
]
] |
2504.18000 | Yu Li | Yu Li, Shuang Liu, Hang Wang, Yao-Chuan Wang | The Impact of Inhomogeneous Perturbations of the Inflaton on the
Cosmological Primordial Magnetic Field | 13 pages, 1 figure | null | null | null | astro-ph.CO gr-qc | http://creativecommons.org/licenses/by/4.0/ | We investigate the impact of inhomogeneous inflaton perturbations on
primordial magnetic fields within the framework of generalized inflationary
magnetogenesis models. Extending the Ratra model to general spacetime
backgrounds, we analyze the constraint structure of the electromagnetic field
and demonstrate that the standard Coulomb gauge must be generalized to
accommodate spatial inhomogeneities. Instead of the vector potential, we solve
the conjugate momentum with the modified initial conditions introduced by the
coupling function, which become dominant during the late stages of inflation.
These change the conditions under which scale-invariant electromagnetic spectra
are achieved. Furthermore, we address the challenge of evaluating convolutions
between vector potentials and inflaton perturbations by employing separate
large- and small-scale approximations. The resulting influence to the electric
and magnetic power spectra are quantified using $\Delta_E$ and $\Delta_B$,
revealing a scale-dependent influence of inhomogeneities. We also find that the
spectrum index evolution is sensitive to the sign of $V_{\phi}$, with
distinctive behaviors for electric and magnetic fields under different
scale-invariance conditions. Notably, for nearly scale-invariant magnetic
fields, the perturbative effects shift the spectral index towards the red and
migrate toward smaller scales as inflation progresses, offering a potential
observational probe to differentiate between large-field and small-field
inflation scenarios.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:09:07 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Li",
"Yu",
""
],
[
"Liu",
"Shuang",
""
],
[
"Wang",
"Hang",
""
],
[
"Wang",
"Yao-Chuan",
""
]
] |
2504.18002 | Pariyakorn Maneekul | Pariyakorn Maneekul, Zelda B. Zabinsky, Giulia Pedrielli | Branching Adaptive Surrogate Search Optimization (BASSO) | null | null | null | null | math.OC | http://creativecommons.org/licenses/by/4.0/ | Global optimization of black-box functions is challenging in high dimensions.
We introduce a conceptual adaptive random search framework, Branching Adaptive
Surrogate Search Optimization (BASSO), that combines partitioning and surrogate
modeling for subregion sampling. We present a finite-time analysis of BASSO,
and establish conditions under which it is theoretically possible to scale to
high dimensions. While we do not expect that any implementation will achieve
the theoretical ideal, we experiment with several BASSO variations and discuss
implications on narrowing the gap between theory and implementation. Numerical
results on test problems are presented.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:13:05 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Maneekul",
"Pariyakorn",
""
],
[
"Zabinsky",
"Zelda B.",
""
],
[
"Pedrielli",
"Giulia",
""
]
] |
2504.18003 | Aditya Ellendula S | Aditya S Ellendula, Chandrajit Bajaj | Self-Balancing, Memory Efficient, Dynamic Metric Space Data Maintenance,
for Rapid Multi-Kernel Estimation | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We present a dynamic self-balancing octree data structure that enables
efficient neighborhood maintenance in evolving metric spaces, a key challenge
in modern machine learning systems. Many learning and generative models operate
as dynamical systems whose representations evolve during training, requiring
fast, adaptive spatial organization. Our two-parameter octree supports
logarithmic-time updates and queries, eliminating the need for costly full
rebuilds as data distributions shift. We demonstrate its effectiveness in four
areas: (1) accelerating Stein variational gradient descent by supporting more
particles with lower overhead; (2) enabling real-time, incremental KNN
classification with logarithmic complexity; (3) facilitating efficient, dynamic
indexing and retrieval for retrieval-augmented generation; and (4) improving
sample efficiency by jointly optimizing input and latent spaces. Across all
applications, our approach yields exponential speedups while preserving
accuracy, particularly in high-dimensional spaces where maintaining adaptive
spatial structure is critical.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:15:53 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Ellendula",
"Aditya S",
""
],
[
"Bajaj",
"Chandrajit",
""
]
] |
2504.18004 | Daisuke Niizumi | Daisuke Niizumi, Daiki Takeuchi, Masahiro Yasuda, Binh Thien Nguyen,
Yasunori Ohishi, and Noboru Harada | Assessing the Utility of Audio Foundation Models for Heart and
Respiratory Sound Analysis | 4 pages, 1 figure, and 4 tables. Accepted by IEEE EMBC 2025 | null | null | null | eess.AS cs.SD | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Pre-trained deep learning models, known as foundation models, have become
essential building blocks in machine learning domains such as natural language
processing and image domains. This trend has extended to respiratory and heart
sound models, which have demonstrated effectiveness as off-the-shelf feature
extractors. However, their evaluation benchmarking has been limited, resulting
in incompatibility with state-of-the-art (SOTA) performance, thus hindering
proof of their effectiveness. This study investigates the practical
effectiveness of off-the-shelf audio foundation models by comparing their
performance across four respiratory and heart sound tasks with SOTA fine-tuning
results. Experiments show that models struggled on two tasks with noisy data
but achieved SOTA performance on the other tasks with clean data. Moreover,
general-purpose audio models outperformed a respiratory sound model,
highlighting their broader applicability. With gained insights and the released
code, we contribute to future research on developing and leveraging foundation
models for respiratory and heart sounds.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:17:58 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Niizumi",
"Daisuke",
""
],
[
"Takeuchi",
"Daiki",
""
],
[
"Yasuda",
"Masahiro",
""
],
[
"Nguyen",
"Binh Thien",
""
],
[
"Ohishi",
"Yasunori",
""
],
[
"Harada",
"Noboru",
""
]
] |
2504.18005 | Yu Li | Hang Wang, Shuang Liu, Yu Li, Yao-chuan Wang | The equivalence between Einstein and Jordan frames: a study based on the
inflationary magnetogenesis model | 15 pages, no figure | null | null | null | astro-ph.CO gr-qc | http://creativecommons.org/licenses/by/4.0/ | The equivalence of the Jordan and Einstein frames has been a subject of
considerable interest in the field. In this paper, within the context of $f(R)$
gravity, we explore the inflationary magnetogenesis model, focusing on the
magnetic field energy density and its spectrum in both the Jordan and Einstein
frames to elucidate the equivalence between these two reference frames. Our
analysis reveals that during the inflationary epoch, while the magnetic field
exhibits a scale-invariant spectrum in the Einstein frame, it demonstrates a
blue spectrum in the Jordan frame. Additionally, we investigate the
post-inflationary evolution of the magnetic field's energy density in both
frames, uncovering that for scale-invariant spectra in the Einstein frame
during inflation, the magnetic field transitions to a blue spectrum, whereas in
the Jordan frame, it evolves into a red spectrum. We also establish the
conditions under which both frames may exhibit scale-invariant spectra
simultaneously during the inflationary period.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:18:00 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Wang",
"Hang",
""
],
[
"Liu",
"Shuang",
""
],
[
"Li",
"Yu",
""
],
[
"Wang",
"Yao-chuan",
""
]
] |
2504.18006 | Shengqi Yang | Shengqi Yang, Adam Lidz, Hui Li, Gergo Popping, Jorge A. Zavala, and
Guochao Sun | Do cosmological simulations reproduce the [OIII] 88 $\mu$m line emission
and properties of JWST-discovered galaxies at $z \geq 12$? | 11 pages, 4 figures | null | null | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent ALMA observations of the [OIII] 88 $\mu$m line provide spectroscopic
confirmation of two JWST photometric candidates, GS-z14 and GHZ2, at $z=14.2$
and $z=12.3$, respectively. These discoveries reveal that star formation and
chemical enrichment were already underway when the universe was merely 300 Myr
old, posing a challenge to galaxy formation models. Here we construct
post-processed models for the [OIII] emission lines from galaxies in the
state-of-the-art FIRE and IllustrisTNG simulations. Neither simulation suite
contains galaxies directly comparable to GS-z14 or GHZ2. However, one simulated
FIRE galaxy closely resembles GS-z14 in its star formation rate (SFR), stellar
mass, metallicity, [OIII] luminosity and line-width, albeit at $z=8.7$, lagging
GS-z14's formation by roughly 300 Myr. Although further investigation is
required, we argue that the lack of simulated galaxies matching GS-z14 and GHZ2
may largely be a consequence of the limited volume of the FIRE simulations and
the limited mass resolution of Illustris-TNG. We quantify the prospects for
follow-up spectroscopic detections of GS-z14 in the [OIII] 52 $\mu$m line with
ALMA, and in rest-frame optical [OIII] and Balmer lines with the MIRI
instrument on JWST.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:18:48 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yang",
"Shengqi",
""
],
[
"Lidz",
"Adam",
""
],
[
"Li",
"Hui",
""
],
[
"Popping",
"Gergo",
""
],
[
"Zavala",
"Jorge A.",
""
],
[
"Sun",
"Guochao",
""
]
] |
2504.18007 | Yazan Otoum | Yazan Otoum, Amiya Nayak | Differential Privacy-Driven Framework for Enhancing Heart Disease
Prediction | \c{opyright} 2025 IEEE. Accepted to IEEE International Conference on
Communications ICC 2025. Final version to appear in IEEE Xplore | null | null | null | cs.AI cs.CR cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the rapid digitalization of healthcare systems, there has been a
substantial increase in the generation and sharing of private health data.
Safeguarding patient information is essential for maintaining consumer trust
and ensuring compliance with legal data protection regulations. Machine
learning is critical in healthcare, supporting personalized treatment, early
disease detection, predictive analytics, image interpretation, drug discovery,
efficient operations, and patient monitoring. It enhances decision-making,
accelerates research, reduces errors, and improves patient outcomes. In this
paper, we utilize machine learning methodologies, including differential
privacy and federated learning, to develop privacy-preserving models that
enable healthcare stakeholders to extract insights without compromising
individual privacy. Differential privacy introduces noise to data to guarantee
statistical privacy, while federated learning enables collaborative model
training across decentralized datasets. We explore applying these technologies
to Heart Disease Data, demonstrating how they preserve privacy while delivering
valuable insights and comprehensive analysis. Our results show that using a
federated learning model with differential privacy achieved a test accuracy of
85%, ensuring patient data remained secure and private throughout the process.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:27:40 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Otoum",
"Yazan",
""
],
[
"Nayak",
"Amiya",
""
]
] |
2504.18008 | Nooshin Yousefzadeh Hosseini | Nooshin Yousefzadeh, Rahul Sengupta, Sanjay Ranka | TGDT: A Temporal Graph-based Digital Twin for Urban Traffic Corridors | 8 pages, 4 figures, 1 table | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Urban congestion at signalized intersections leads to significant delays,
economic losses, and increased emissions. Existing deep learning models often
lack spatial generalizability, rely on complex architectures, and struggle with
real-time deployment. To address these limitations, we propose the Temporal
Graph-based Digital Twin (TGDT), a scalable framework that integrates Temporal
Convolutional Networks and Attentional Graph Neural Networks for dynamic,
direction-aware traffic modeling and assessment at urban corridors. TGDT
estimates key Measures of Effectiveness (MOEs) for traffic flow optimization at
both the intersection level (e.g., queue length, waiting time) and the corridor
level (e.g., traffic volume, travel time). Its modular architecture and
sequential optimization scheme enable easy extension to any number of
intersections and MOEs. The model outperforms state-of-the-art baselines by
accurately producing high-dimensional, concurrent multi-output estimates. It
also demonstrates high robustness and accuracy across diverse traffic
conditions, including extreme scenarios, while relying on only a minimal set of
traffic features. Fully parallelized, TGDT can simulate over a thousand
scenarios within a matter of seconds, offering a cost-effective, interpretable,
and real-time solution for traffic signal optimization.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:28:32 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yousefzadeh",
"Nooshin",
""
],
[
"Sengupta",
"Rahul",
""
],
[
"Ranka",
"Sanjay",
""
]
] |
2504.18009 | Kim Weston | Jin Hyuk Choi and Kim Weston | Radner equilibrium with population growth | 12 pages, 4 figures, to appear in SIAM Journal on Financial
Mathematics, Short Communications | null | null | null | q-fin.MF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove the existence of a Radner equilibrium in a model with population
growth and analyze the effects on asset prices. A finite population of agents
grows indefinitely at a Poisson rate, while receiving unspanned income and
choosing between consumption and investing into an annuity with
infinitely-lived exponential preferences. After establishing the existence of
an equilibrium for a truncated number of agents, we prove that an equilibrium
exists for the model with unlimited population growth. Our numerics show that
increasing the birth rate reduces oscillations in the equilibrium annuity
price, and when younger agents prioritize the present more than older agents,
the equilibrium annuity price rises compared to a uniform demographic.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:29:28 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Choi",
"Jin Hyuk",
""
],
[
"Weston",
"Kim",
""
]
] |
2504.18010 | Zilin Huang | Zilin Huang, Zihao Sheng, Zhengyang Wan, Yansong Qu, Yuhao Luo, Boyue
Wang, Pei Li, Yen-Jung Chen, Jiancong Chen, Keke Long, Jiayi Meng, Yue Leng,
Sikai Chen | Sky-Drive: A Distributed Multi-Agent Simulation Platform for
Socially-Aware and Human-AI Collaborative Future Transportation | 15 pages, 7 figures | null | null | null | cs.RO cs.AI cs.HC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advances in autonomous system simulation platforms have significantly
enhanced the safe and scalable testing of driving policies. However, existing
simulators do not yet fully meet the needs of future transportation research,
particularly in modeling socially-aware driving agents and enabling effective
human-AI collaboration. This paper introduces Sky-Drive, a novel distributed
multi-agent simulation platform that addresses these limitations through four
key innovations: (a) a distributed architecture for synchronized simulation
across multiple terminals; (b) a multi-modal human-in-the-loop framework
integrating diverse sensors to collect rich behavioral data; (c) a human-AI
collaboration mechanism supporting continuous and adaptive knowledge exchange;
and (d) a digital twin (DT) framework for constructing high-fidelity virtual
replicas of real-world transportation environments. Sky-Drive supports diverse
applications such as autonomous vehicle (AV)-vulnerable road user (VRU)
interaction modeling, human-in-the-loop training, socially-aware reinforcement
learning, personalized driving policy, and customized scenario generation.
Future extensions will incorporate foundation models for context-aware decision
support and hardware-in-the-loop (HIL) testing for real-world validation. By
bridging scenario generation, data collection, algorithm training, and hardware
integration, Sky-Drive has the potential to become a foundational platform for
the next generation of socially-aware and human-centered autonomous
transportation research. The demo video and code are available
at:https://sky-lab-uw.github.io/Sky-Drive-website/
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:33:26 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Huang",
"Zilin",
""
],
[
"Sheng",
"Zihao",
""
],
[
"Wan",
"Zhengyang",
""
],
[
"Qu",
"Yansong",
""
],
[
"Luo",
"Yuhao",
""
],
[
"Wang",
"Boyue",
""
],
[
"Li",
"Pei",
""
],
[
"Chen",
"Yen-Jung",
""
],
[
"Chen",
"Jiancong",
""
],
[
"Long",
"Keke",
""
],
[
"Meng",
"Jiayi",
""
],
[
"Leng",
"Yue",
""
],
[
"Chen",
"Sikai",
""
]
] |
2504.18011 | Maik Gr\"oger | Mar\'ia Isabel Cortez and Maik Gr\"oger and Olga Lukina | Group actions with almost normal stabilizers | 26 pages | null | null | null | math.DS | http://creativecommons.org/licenses/by/4.0/ | In this paper, we consider minimal group actions of countable groups on
compact Hausdorff spaces by homeomorphisms. We show that the existence of a
point with finite stabilizer imposes strong restrictions on the dynamics: the
residual set of points then has stabilizers conjugate to the same almost normal
subgroup of the acting group. Under certain conditions on the acting group such
an action also has no essential holonomy. If the acting group is residually
finite, we show that every group action with finite almost normal stabilizers
with no essential holonomy arises as an almost finite-to-one factor of an
essentially free action.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:33:52 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Cortez",
"María Isabel",
""
],
[
"Gröger",
"Maik",
""
],
[
"Lukina",
"Olga",
""
]
] |
2504.18012 | Yu Zhuang | Zhuang Yu, Shiliang Sun, Jing Zhao, Tengfei Song, Hao Yang | Memory Reviving, Continuing Learning and Beyond: Evaluation of
Pre-trained Encoders and Decoders for Multimodal Machine Translation | null | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal Machine Translation (MMT) aims to improve translation quality by
leveraging auxiliary modalities such as images alongside textual input. While
recent advances in large-scale pre-trained language and vision models have
significantly benefited unimodal natural language processing tasks, their
effectiveness and role in MMT remain underexplored. In this work, we conduct a
systematic study on the impact of pre-trained encoders and decoders in
multimodal translation models. Specifically, we analyze how different training
strategies, from training from scratch to using pre-trained and partially
frozen components, affect translation performance under a unified MMT
framework. Experiments are carried out on the Multi30K and CoMMuTE dataset
across English-German and English-French translation tasks. Our results reveal
that pre-training plays a crucial yet asymmetrical role in multimodal settings:
pre-trained decoders consistently yield more fluent and accurate outputs, while
pre-trained encoders show varied effects depending on the quality of
visual-text alignment. Furthermore, we provide insights into the interplay
between modality fusion and pre-trained components, offering guidance for
future architecture design in multimodal translation systems.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:44:04 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yu",
"Zhuang",
""
],
[
"Sun",
"Shiliang",
""
],
[
"Zhao",
"Jing",
""
],
[
"Song",
"Tengfei",
""
],
[
"Yang",
"Hao",
""
]
] |
2504.18014 | Weixiang Sun | Weixiang Sun, Han Shen, Biwei Jiang, Xiaowei Liu | The Kinematic Signature of the Warp and Waves in the Milky Way Disk | 6 pages, 4 figures, 2 tables, accepted for publication in ApJ | null | null | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using over 170,000 red clump stars selected from LAMOST and APOGEE, we
conduct a detailed analysis of the stellar $V_{Z}$ as a function of $L_{Z}$ (or
$R_{g}$) across different $\phi$ bins for various disk populations. The $V_{Z}$
of the whole RC sample stars exhibits a wave-like pattern superimposed on an
exponentially increasing trend, indicating the contribution from disk warp,
disk flare and disk waves. Our results across various populations suggest that
the thin disk is similar to the whole RC sample behavior, while the thick disk
displays a wave-like pattern superimposed on a linearly increasing trend,
meaning that the features of disk warp and waves are present in both thin and
thick disks, and the disk flare feature is only present in the thin disk. These
results indicate that the disk warp is potentially driven by secular processes
like disk perturbations from intergalactic magnetic fields and a misaligned
dark halo. The line-of-node (LON) of the disk warp of various populations
displays a slight difference, with $\phi_{0}$ = 5.68 $\pm$ 2.91 degree for the
whole RC sample stars, $\phi_{0}$ = 5.78 $\pm$ 2.89 degree for the thin disk
stars, and $\phi_{0}$ = 4.10 $\pm$ 3.43 degree for the thick disk stars.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:50:09 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Sun",
"Weixiang",
""
],
[
"Shen",
"Han",
""
],
[
"Jiang",
"Biwei",
""
],
[
"Liu",
"Xiaowei",
""
]
] |
2504.18015 | Hanrui Wang | Hanrui Wang, Shuo Wang, Chun-Shien Lu, Isao Echizen | Diffusion-Driven Universal Model Inversion Attack for Face Recognition | null | null | null | null | cs.CR cs.CV cs.LG | http://creativecommons.org/publicdomain/zero/1.0/ | Facial recognition technology poses significant privacy risks, as it relies
on biometric data that is inherently sensitive and immutable if compromised. To
mitigate these concerns, face recognition systems convert raw images into
embeddings, traditionally considered privacy-preserving. However, model
inversion attacks pose a significant privacy threat by reconstructing these
private facial images, making them a crucial tool for evaluating the privacy
risks of face recognition systems. Existing methods usually require training
individual generators for each target model, a computationally expensive
process. In this paper, we propose DiffUMI, a training-free diffusion-driven
universal model inversion attack for face recognition systems. DiffUMI is the
first approach to apply a diffusion model for unconditional image generation in
model inversion. Unlike other methods, DiffUMI is universal, eliminating the
need for training target-specific generators. It operates within a fixed
framework and pretrained diffusion model while seamlessly adapting to diverse
target identities and models. DiffUMI breaches privacy-preserving face
recognition systems with state-of-the-art success, demonstrating that an
unconditional diffusion model, coupled with optimized adversarial search,
enables efficient and high-fidelity facial reconstruction. Additionally, we
introduce a novel application of out-of-domain detection (OODD), marking the
first use of model inversion to distinguish non-face inputs from face inputs
based solely on embeddings.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:53:27 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Wang",
"Hanrui",
""
],
[
"Wang",
"Shuo",
""
],
[
"Lu",
"Chun-Shien",
""
],
[
"Echizen",
"Isao",
""
]
] |
2504.18016 | Ying Zhang | Ying Zhang, Fan Liu, Tao Liu, Shi Jin | Optimal Power Allocation for OFDM-based Ranging Using Random
Communication Signals | 12 pages, 9 figures, submitted to IEEE for possible publication | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | High-precision ranging plays a crucial role in future 6G Integrated Sensing
and Communication (ISAC) systems. To improve the ranging performance while
maximizing the resource utilization efficiency, future 6G ISAC networks have to
reuse data payload signals for both communication and sensing, whose inherent
randomness may deteriorate the ranging performance. To address this issue, this
paper investigates the power allocation (PA) design for an OFDM-based ISAC
system under random signaling, aiming to reduce the ranging sidelobe level of
both periodic and aperiodic auto-correlation functions (P-ACF and A-ACF) of the
ISAC signal. Towards that end, we first derive the closed-form expressions of
the average squared P-ACF and A-ACF, and then propose to minimize the
expectation of the integrated sidelobe level (EISL) under arbitrary
constellation mapping. We then rigorously prove that the uniform PA scheme
achieves the global minimum of the EISL for both P-ACF and A-ACF. As a step
further, we show that this scheme also minimizes the P-ACF sidelobe level at
every lag. Moreover, we extend our analysis to the P-ACF case with
frequency-domain zero-padding, which is a typical approach to improve the
ranging resolution. We reveal that there exists a tradeoff between sidelobe
level and mainlobe width, and propose a project gradient descent algorithm to
seek a locally optimal PA scheme that reduces the EISL. Finally, we validate
our theoretical findings through extensive simulation results, confirming the
effectiveness of the proposed PA methods in reducing the ranging sidelobe level
for random OFDM signals.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:54:03 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Zhang",
"Ying",
""
],
[
"Liu",
"Fan",
""
],
[
"Liu",
"Tao",
""
],
[
"Jin",
"Shi",
""
]
] |
2504.18017 | Timothy Sudijono | Sourav Chatterjee, Timothy Sudijono | Non-identifiability distinguishes Neural Networks among Parametric
Models | 16 pages. Comments welcome | null | null | null | math.ST cs.LG stat.ML stat.TH | http://creativecommons.org/licenses/by/4.0/ | One of the enduring problems surrounding neural networks is to identify the
factors that differentiate them from traditional statistical models. We prove a
pair of results which distinguish feedforward neural networks among parametric
models at the population level, for regression tasks. Firstly, we prove that
for any pair of random variables $(X,Y)$, neural networks always learn a
nontrivial relationship between $X$ and $Y$, if one exists. Secondly, we prove
that for reasonable smooth parametric models, under local and global
identifiability conditions, there exists a nontrivial $(X,Y)$ pair for which
the parametric model learns the constant predictor $\mathbb{E}[Y]$. Together,
our results suggest that a lack of identifiability distinguishes neural
networks among the class of smooth parametric models.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 01:59:02 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chatterjee",
"Sourav",
""
],
[
"Sudijono",
"Timothy",
""
]
] |
2504.18019 | Min Yuan | M. Yuan and Y. Huang | A search for $\bar{B}_{s0}^{*}$ and $\bar{B}^{*}_{s1}$ through the
$\bar{K}p$ interaction | 6 pages, 6 figures | null | null | null | hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The formation and properties of hadrons, particularly those involving heavy
quarks, remain a central topic in particle physics due to the nonperturbative
nature of Quantum Chromodynamics (QCD) at low energies. Heavy-quark symmetry
(HQS) provides a valuable framework for understanding the behavior of heavy
quarks in hadronic systems, offering insights into spin and flavor symmetries.
However, discrepancies between HQS predictions and experimental observations --
particularly the absence of predicted $B^{(*)}K$ molecular states, which should
be the bottom counterparts of the observed $D^{(*)}K$ molecular states --
challenge its universality. This study investigates the production of the
$\bar{B}_{s0}^{*}(5725)$ and $\bar{B}_{s1}^{*}(5778)$ states, the antiparticles
of the $B_{s0}^{*}(5725)$ and $B_{s1}^{*}(5778)$ states, interpreted as
molecular states containing $B\bar{K}$ and $B^{*}\bar{K}$ components, through
the reactions $K^{-}p \to \Lambda_{b}^{0}\bar{B}_{s0}^{*}$ and $K^{-}p \to
\Lambda_{b}^{0}\bar{B}_{s1}^{*}$. Using effective Lagrangians, we analyze the
cross sections, incorporating $t$-channel $B^{(*)}$ meson exchanges and
initial-state $\bar{K}N$ interactions (ISI). Our results reveal significant ISI
enhancements and energy-dependent production thresholds, with
$\bar{B}_{s0}^{*}(5725)$ exhibiting higher cross sections than
$\bar{B}_{s1}^{*}(5778)$. We suggest that the $\bar{B}_{s0}^{*}(5725)$ state
can be optimally searched for at beam energies around $12.91~\text{GeV}$, while
the $\bar{B}_{s1}^{*}(5778)$ state requires significantly higher energies for
its production. These findings provide valuable guidance for experimental
searches at facilities like CERN and J-PARC, offering a pathway to test HQS and
explore the nature of exotic hadrons. Theoretical uncertainties and
implications for molecular versus conventional quark-antiquark interpretations
are also discussed.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:10:17 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yuan",
"M.",
""
],
[
"Huang",
"Y.",
""
]
] |
2504.18020 | Siyuan Song | Guyue Hu, Siyuan Song, Yukun Kang, Zhu Yin, Gangming Zhao, Chenglong
Li, and Jin Tang | Federated Client-tailored Adapter for Medical Image Segmentation | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Medical image segmentation in X-ray images is beneficial for computer-aided
diagnosis and lesion localization. Existing methods mainly fall into a
centralized learning paradigm, which is inapplicable in the practical medical
scenario that only has access to distributed data islands. Federated Learning
has the potential to offer a distributed solution but struggles with heavy
training instability due to client-wise domain heterogeneity (including
distribution diversity and class imbalance). In this paper, we propose a novel
Federated Client-tailored Adapter (FCA) framework for medical image
segmentation, which achieves stable and client-tailored adaptive segmentation
without sharing sensitive local data. Specifically, the federated adapter stirs
universal knowledge in off-the-shelf medical foundation models to stabilize the
federated training process. In addition, we develop two client-tailored
federated updating strategies that adaptively decompose the adapter into common
and individual components, then globally and independently update the parameter
groups associated with common client-invariant and individual client-specific
units, respectively. They further stabilize the heterogeneous federated
learning process and realize optimal client-tailored instead of sub-optimal
global-compromised segmentation models. Extensive experiments on three
large-scale datasets demonstrate the effectiveness and superiority of the
proposed FCA framework for federated medical segmentation.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:20:25 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Hu",
"Guyue",
""
],
[
"Song",
"Siyuan",
""
],
[
"Kang",
"Yukun",
""
],
[
"Yin",
"Zhu",
""
],
[
"Zhao",
"Gangming",
""
],
[
"Li",
"Chenglong",
""
],
[
"Tang",
"Jin",
""
]
] |
2504.18021 | Victor Blasco | Victor Blasco | On Artin algebras whose indecomposable modules are determined by
composition factors | null | null | null | null | math.RA math.AC math.CT | http://creativecommons.org/licenses/by/4.0/ | It was conjectured at the end of the book "Representation theory of Artin
algebras" by M. Auslander, I. Reiten and S. Smalo that an Artin algebra with
the property that its finitely generated indecomposable modules are up to
isomorphism completely determined by theirs composition factors is of finite
representation type. Examples of rings with this property are the semisimple
artinian rings and the rings of the form $\mathbb{Z}_n$. An affirmative answer
is obtained for some special cases, namely, the commutative, the hereditary and
the radical square zero case.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:23:24 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Blasco",
"Victor",
""
]
] |
2504.18022 | Jinnan Piao | Jinnan Piao, Dong Li, Yiming Sun, Zhibo Li, Ming Yang, and Xueting Yu | Iterative Joint Detection of Kalman Filter and Channel Decoder for
Sensor-to-Controller Link in Wireless Networked Control Systems | 5 pages, 4 figures | null | null | null | cs.IT cs.SY eess.SY math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this letter, we propose an iterative joint detection algorithm of Kalman
filter (KF) and channel decoder for the sensor-to-controller link of wireless
networked control systems, which utilizes the prior information of control
system to improve the control and communication performance. In the algorithm,
we first use the KF to estimate the probability density of the control system
outputs and calculate the prior probability of received signals to assist
decoding. Then, the possible outputs of the control system are traversed to
update the prior probability in order to implement iterative detection. The
simulation results show that the prior information can reduce the block error
rate performance of communications to improve the root mean square error
performance of controls.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:25:32 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Piao",
"Jinnan",
""
],
[
"Li",
"Dong",
""
],
[
"Sun",
"Yiming",
""
],
[
"Li",
"Zhibo",
""
],
[
"Yang",
"Ming",
""
],
[
"Yu",
"Xueting",
""
]
] |
2504.18023 | Xiaofei Shen | Xiaofei Shen, Yue-Yue Chen, Karen Z. Hatsagortsyan, and Christoph H.
Keitel | Anomalous lepton acceleration in the radiation reaction dominated
reflection regime | null | null | null | null | physics.plasm-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Relativistic electrons colliding with intense counterpropagating laser pulses
are expected to lose energy through radiation reaction. However, we reveal a
counterintuitive regime where reflected leptons (including incident electrons,
generated electrons, and positrons) gain significant energies when a relatively
loosely focused ultraintense laser interacts with counterpropagating electrons.
Because of strong radiation reaction, these particles can be halted and
reflected near the laser peak. The subsequent asymmetric laser field then
accelerates the reflected leptons to energies far exceeding their initial
values. Using three-dimensional particle-in-cell simulations, we demonstrate
the generation and acceleration of quasimonoenergetic positrons to multi-GeV
energies with a high number conversion efficiency employing 10~PW-class lasers.
These findings not only provide a single-stage solution for positron creation
and acceleration, but also offer a promising alternative explanation for the
origin of ultrahigh-energy cosmic rays as particularly those associated with
intense fast radio bursts.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:29:17 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Shen",
"Xiaofei",
""
],
[
"Chen",
"Yue-Yue",
""
],
[
"Hatsagortsyan",
"Karen Z.",
""
],
[
"Keitel",
"Christoph H.",
""
]
] |
2504.18024 | Yiwei Zha | Yiwei Zha | SMARTFinRAG: Interactive Modularized Financial RAG Benchmark | For open source github repo, see
https://github.com/JonathanZha47/SMARTFinRAG | null | null | null | cs.CE cs.CL cs.IR | http://creativecommons.org/licenses/by/4.0/ | Financial sectors are rapidly adopting language model technologies, yet
evaluating specialized RAG systems in this domain remains challenging. This
paper introduces SMARTFinRAG, addressing three critical gaps in financial RAG
assessment: (1) a fully modular architecture where components can be
dynamically interchanged during runtime; (2) a document-centric evaluation
paradigm generating domain-specific QA pairs from newly ingested financial
documents; and (3) an intuitive interface bridging research-implementation
divides. Our evaluation quantifies both retrieval efficacy and response
quality, revealing significant performance variations across configurations.
The platform's open-source architecture supports transparent, reproducible
research while addressing practical deployment challenges faced by financial
institutions implementing RAG systems.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:29:56 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Zha",
"Yiwei",
""
]
] |
2504.18025 | Shuanglin Yan | Shuanglin Yan, Neng Dong, Shuang Li, Rui Yan, Hao Tang, Jing Qin | ShapeSpeak: Body Shape-Aware Textual Alignment for Visible-Infrared
Person Re-Identification | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Visible-Infrared Person Re-identification (VIReID) aims to match visible and
infrared pedestrian images, but the modality differences and the complexity of
identity features make it challenging. Existing methods rely solely on identity
label supervision, which makes it difficult to fully extract high-level
semantic information. Recently, vision-language pre-trained models have been
introduced to VIReID, enhancing semantic information modeling by generating
textual descriptions. However, such methods do not explicitly model body shape
features, which are crucial for cross-modal matching. To address this, we
propose an effective Body Shape-aware Textual Alignment (BSaTa) framework that
explicitly models and utilizes body shape information to improve VIReID
performance. Specifically, we design a Body Shape Textual Alignment (BSTA)
module that extracts body shape information using a human parsing model and
converts it into structured text representations via CLIP. We also design a
Text-Visual Consistency Regularizer (TVCR) to ensure alignment between body
shape textual representations and visual body shape features. Furthermore, we
introduce a Shape-aware Representation Learning (SRL) mechanism that combines
Multi-text Supervision and Distribution Consistency Constraints to guide the
visual encoder to learn modality-invariant and discriminative identity
features, thus enhancing modality invariance. Experimental results demonstrate
that our method achieves superior performance on the SYSU-MM01 and RegDB
datasets, validating its effectiveness.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:37:47 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yan",
"Shuanglin",
""
],
[
"Dong",
"Neng",
""
],
[
"Li",
"Shuang",
""
],
[
"Yan",
"Rui",
""
],
[
"Tang",
"Hao",
""
],
[
"Qin",
"Jing",
""
]
] |
2504.18026 | Emiliano Penaloza | Emiliano Penaloza, Tianyue H. Zhan, Laurent Charlin, Mateo Espinosa
Zarlenga | Addressing Concept Mislabeling in Concept Bottleneck Models Through
Preference Optimization | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Concept Bottleneck Models (CBMs) propose to enhance the trustworthiness of AI
systems by constraining their decisions on a set of human understandable
concepts. However, CBMs typically assume that datasets contains accurate
concept labels an assumption often violated in practice, which we show can
significantly degrade performance (by 25% in some cases). To address this, we
introduce the Concept Preference Optimization (CPO) objective, a new loss
function based on Direct Preference Optimization, which effectively mitigates
the negative impact of concept mislabeling on CBM performance. We provide an
analysis on some key properties of the CPO objective showing it directly
optimizes for the concept's posterior distribution, and contrast it against
Binary Cross Entropy (BCE) where we show CPO is inherently less sensitive to
concept noise. We empirically confirm our analysis finding that CPO
consistently outperforms BCE in three real world datasets with and without
added label noise.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:43:10 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Penaloza",
"Emiliano",
""
],
[
"Zhan",
"Tianyue H.",
""
],
[
"Charlin",
"Laurent",
""
],
[
"Zarlenga",
"Mateo Espinosa",
""
]
] |
2504.18027 | Zhaoxiang Liu | Zezhou Chen, Zhaoxiang Liu, Kai Wang, Kohou Wang, Shiguo Lian | A Large Vision-Language Model based Environment Perception System for
Visually Impaired People | Accepted by IROS2024(9 pages, 8 figures) | null | 10.1109/IROS58592.2024.10801813 | null | cs.CV cs.AI cs.RO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | It is a challenging task for visually impaired people to perceive their
surrounding environment due to the complexity of the natural scenes. Their
personal and social activities are thus highly limited. This paper introduces a
Large Vision-Language Model(LVLM) based environment perception system which
helps them to better understand the surrounding environment, by capturing the
current scene they face with a wearable device, and then letting them retrieve
the analysis results through the device. The visually impaired people could
acquire a global description of the scene by long pressing the screen to
activate the LVLM output, retrieve the categories of the objects in the scene
resulting from a segmentation model by tapping or swiping the screen, and get a
detailed description of the objects they are interested in by double-tapping
the screen. To help visually impaired people more accurately perceive the
world, this paper proposes incorporating the segmentation result of the RGB
image as external knowledge into the input of LVLM to reduce the LVLM's
hallucination. Technical experiments on POPE, MME and LLaVA-QA90 show that the
system could provide a more accurate description of the scene compared to
Qwen-VL-Chat, exploratory experiments show that the system helps visually
impaired people to perceive the surrounding environment effectively.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:46:22 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chen",
"Zezhou",
""
],
[
"Liu",
"Zhaoxiang",
""
],
[
"Wang",
"Kai",
""
],
[
"Wang",
"Kohou",
""
],
[
"Lian",
"Shiguo",
""
]
] |
2504.18028 | Noah Schlossberger | Noah Schlossberger, Rajavardhan Talashila, Nikunjkumar Prajapati, and
Christopher L. Holloway | Angle-of-arrival detection of radio-frequency waves via Rydberg atom
fluorescence imaging of standing waves in a glass vapor cell | null | null | null | null | physics.atom-ph | http://creativecommons.org/licenses/by/4.0/ | We present a method for measuring the angle-of-arrival of 40 GHz
radio-frequency (RF) radiation by mapping the standing waves generated in a
rectangular glass vapor cell. These standing waves have regular and
well-defined structure from which we can infer the angle and sign of the
wavevector of the RF field. We map the field using spatially resolved light
sheet spectroscopy of Rydberg states of rubidium atoms in the cell. Unlike
traditional phased arrays, this detection scheme is compact and low-complexity,
has an active area of nearly 4$\pi$ steradians, and is sensitive to all RF
polarizations. For in-plane measurements ($\phi = 0$), we demonstrate
quantitative angle-of-arrival measurements with an uncertainty on the order of
one degree in an 11~s measurement, and for out-of-plane measurements (arbitrary
$\theta$,$\phi$), we demonstrate angle-of-arrival detection with uncertainty on
the order of several degrees.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:47:29 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Schlossberger",
"Noah",
""
],
[
"Talashila",
"Rajavardhan",
""
],
[
"Prajapati",
"Nikunjkumar",
""
],
[
"Holloway",
"Christopher L.",
""
]
] |
2504.18029 | Venkateswarlu Gudepu gudepu | L. Malakalapalli, V. Gudepu, B. Chirumamilla, S.J. Yadhunandan, K.
Kondepu | Integrating Explainable AI for Energy Efficient Open Radio Access
Networks | Accepted at FNFW 2024 | null | null | null | cs.NI cs.ET | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Open Radio Access Network (Open RAN) is an emerging idea -- transforming
the traditional Radio Access Networks (RAN) that are monolithic and inflexible
into more flexible and innovative. By leveraging open standard interfaces, data
collection across all RAN layers becomes feasible, paving the way for the
development of energy-efficient Open RAN architectures through Artificial
Intelligence / Machine Learning (AI/ML). However, the inherent complexity and
black-box nature of AI/ML models used for energy consumption prediction pose
challenges in interpreting their underlying factors and relationships. This
work presents an integration of eXplainable AI (XAI) to understand the key RAN
parameters that contribute to energy consumption. Furthermore, the paper delves
into the analysis of RAN parameters -- \emph{airtime}, \emph{goodput},
\emph{throughput}, \emph{buffer status report}, \emph{number of resource
blocks}, and many others -- identified by XAI techniques, highlighting their
significance in energy consumption.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:48:29 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Malakalapalli",
"L.",
""
],
[
"Gudepu",
"V.",
""
],
[
"Chirumamilla",
"B.",
""
],
[
"Yadhunandan",
"S. J.",
""
],
[
"Kondepu",
"K.",
""
]
] |
2504.18030 | Chidiebere Nwaogbo | Chidiebere I. Nwaogbo, Sanjib K. Das, Chinedu E. Ekuma | Tunable topological phase in 2D ScV$_6$Sn$_6$ kagome material | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | We investigate the topological properties of the vanadium-based 2D kagome
metal ScV$_6$Sn$_6$, a ferromagnetic material with a magnetic moment of 0.86
$\mu_B$ per atom. Using ab initio methods, we explore spin-orbit
coupling-induced gapped states and identify multiple Weyl-like crossings around
the Fermi energy, confirming a Chern number $|C| = 1$ and a large anomalous
Hall effect (AHE) of 257 $\Omega^{-1}$cm$^{-1}$. Our calculations reveal a
transition from a topological semimetal to a trivial metallic phase at an
electric field strength of $\approx$0.40 eV/{\AA}. These findings position 2D
ScV$_6$Sn$_6$ as a promising candidate for applications in modern electronic
devices, with its tunable topological phases offering the potential for future
innovations in quantum computing and material design.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:48:30 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Nwaogbo",
"Chidiebere I.",
""
],
[
"Das",
"Sanjib K.",
""
],
[
"Ekuma",
"Chinedu E.",
""
]
] |
2504.18031 | Kai Xiong | Kai Xiong, Chenxin Yang, Yujie Qin, Chau Yuen | Joint Resource Estimation and Trajectory Optimization for eVTOL-involved
CR network: A Monte Carlo Tree Search-based Approach | null | null | null | null | cs.NI eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Electric Vertical Take-Off and Landing (eVTOL) aircraft, pivotal to Advanced
Air Mobility (AAM), are emerging as a transformative transportation paradigm
with the potential to redefine urban and regional mobility. While these systems
offer unprecedented efficiency in transporting people and goods, they rely
heavily on computation capability, safety-critical operations such as real-time
navigation, environmental sensing, and trajectory tracking--necessitating
robust offboard computational support. A widely adopted solution involves
offloading these tasks to terrestrial base stations (BSs) along the flight
path. However, air-to-ground connectivity is often constrained by spectrum
conflicts with terrestrial users, which poses a significant challenge to
maintaining reliable task execution. Cognitive radio (CR) techniques offer
promising capabilities for dynamic spectrum access, making them a natural fit
for addressing this issue. Existing studies often overlook the time-varying
nature of BS resources, such as spectrum availability and CPU cycles, which
leads to inaccurate trajectory planning, suboptimal offloading success rates,
excessive energy consumption, and operational delays. To address these
challenges, we propose a trajectory optimization framework for eVTOL swarms
that maximizes task offloading success probability while minimizing both energy
consumption and resource competition (e.g., spectrum and CPU cycles) with
primary terrestrial users. The proposed algorithm integrates a Multi-Armed
Bandit (MAB) model to dynamically estimate BS resource availability and a Monte
Carlo Tree Search (MCTS) algorithm to determine optimal offloading decisions,
selecting both the BSs and access time windows that align with energy and
temporal constraints.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:48:48 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Xiong",
"Kai",
""
],
[
"Yang",
"Chenxin",
""
],
[
"Qin",
"Yujie",
""
],
[
"Yuen",
"Chau",
""
]
] |
2504.18032 | Chen Chen | Chen Chen, Daochang Liu, Mubarak Shah, Chang Xu | Enhancing Privacy-Utility Trade-offs to Mitigate Memorization in
Diffusion Models | Accepted at CVPR 2025. Project page:
https://chenchen-usyd.github.io/PRSS-Project-Page/ | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Text-to-image diffusion models have demonstrated remarkable capabilities in
creating images highly aligned with user prompts, yet their proclivity for
memorizing training set images has sparked concerns about the originality of
the generated images and privacy issues, potentially leading to legal
complications for both model owners and users, particularly when the memorized
images contain proprietary content. Although methods to mitigate these issues
have been suggested, enhancing privacy often results in a significant decrease
in the utility of the outputs, as indicated by text-alignment scores. To bridge
the research gap, we introduce a novel method, PRSS, which refines the
classifier-free guidance approach in diffusion models by integrating prompt
re-anchoring (PR) to improve privacy and incorporating semantic prompt search
(SS) to enhance utility. Extensive experiments across various privacy levels
demonstrate that our approach consistently improves the privacy-utility
trade-off, establishing a new state-of-the-art.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:51:23 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chen",
"Chen",
""
],
[
"Liu",
"Daochang",
""
],
[
"Shah",
"Mubarak",
""
],
[
"Xu",
"Chang",
""
]
] |
2504.18033 | Won-Kwang Park | Junyong Eom, Sangwoo Kang, Minyeob Lee, and Won-Kwang Park | Real-time inversion of two-dimensional Fresnel experimental database
using orthogonality sampling method with single and multiple sources: the
case of transverse electric polarized waves | 26 pages, 21 figures | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | This paper concerns an application of the orthogonality sampling method (OSM)
for a real-time identification of small objects from two-dimensional Fresnel
experimental dataset in transverse electric polarization. First, we apply the
OSM with a single source by designing an indicator function based on the
asymptotic expansion formula for the scattered field in the presence of small
objects. We demonstrate that the indicator function can be expressed by an
infinite series of Bessel functions of integer order of the first kind, the
range of the signal receiver, and the location of the emitter. Based on this,
we then investigate the applicability and limitations of the designed OSM.
Specifically, we find that the imaging performance is strongly dependent on the
source and the applied frequency. We then apply the OSM with multiple sources
to improve imaging performance. Based on the identified structure of the OSM
with a single source, we design an indicator function with multiple sources and
demonstrate that it can be expressed by an infinite series of the Bessel
function of integer order of the first kind, and we explain that objects can be
identified uniquely using the designed OSM. Numerical simulation results
obtained with the Fresnel experimental dataset demonstrate the advantages and
disadvantages of the OSM with a single source and confirm that the designed OSM
with multiple sources improves imaging performance.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 02:51:46 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Eom",
"Junyong",
""
],
[
"Kang",
"Sangwoo",
""
],
[
"Lee",
"Minyeob",
""
],
[
"Park",
"Won-Kwang",
""
]
] |
2504.18034 | Yoshinori Hara | Yoshinori Hara and Yoshiyuki Kabashima | Quantum effects in rotationally invariant spin glass models | 27 pages, 15 figures | null | null | null | cond-mat.dis-nn cond-mat.stat-mech quant-ph | http://creativecommons.org/licenses/by/4.0/ | This study investigates the quantum effects in transverse-field Ising spin
glass models with rotationally invariant random interactions. The primary aim
is to evaluate the validity of a quasi-static approximation that captures the
imaginary-time dependence of the order parameters beyond the conventional
static approximation. Using the replica method combined with the
Suzuki--Trotter decomposition, we established a stability condition for the
replica symmetric solution, which is analogous to the de Almeida--Thouless
criterion. Numerical analysis of the Sherrington--Kirkpatrick model estimates a
value of the critical transverse field, $\Gamma_\mathrm{c}$, which agrees with
previous Monte Carlo-based estimations. For the Hopfield model, it provides an
estimate of $\Gamma_\mathrm{c}$, which has not been previously evaluated. For
the random orthogonal model, our analysis suggests that quantum effects alter
the random first-order transition scenario in the low-temperature limit. This
study supports a quasi-static treatment for analyzing quantum spin glasses and
may offer useful insights into the analysis of quantum optimization algorithms.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:01:44 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Hara",
"Yoshinori",
""
],
[
"Kabashima",
"Yoshiyuki",
""
]
] |
2504.18035 | D. Bhanu Prakash | D Bhanu Prakash, D K K Vamsi | Dynamics and Control of Additional Food Provided Prey-Predator Systems
exhibiting Holling Type-III Functional Response and Intra-specific
Competition among Predators | arXiv admin note: substantial text overlap with arXiv:2504.09078 | null | null | null | math.DS | http://creativecommons.org/licenses/by/4.0/ | The dynamics of predator-prey systems influenced by intra-specific
competition and additional food resources have increasingly become a subject of
rigorous study in the realm of mathematical biology. In this study, we consider
an additional food provided prey-predator model exhibiting Holling type-III
functional response and the intra-specific competition among predators. We
prove the existence and uniqueness of global positive solutions for the
proposed model. We study the existence and stability of equilibrium points and
further explore the possible bifurcations. We numerically depict the presence
of Hysteresis loop in the system. We further study the global dynamics of the
system and discuss the consequences of providing additional food. Later, we do
the time-optimal control studies with respect to the quality and quantity of
additional food as control variables by transforming the independent variable
in the control system. We show that the findings of these dynamics and control
studies emphasises the role of additional food and intra-specific competition
in bio-control of pests.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:01:59 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Prakash",
"D Bhanu",
""
],
[
"Vamsi",
"D K K",
""
]
] |
2504.18036 | Won-Kwang Park | Won-Kwang Park | Direct sampling method to retrieve small objects from two-dimensional
limited-aperture scattered field data | 18 pages, 14 figures | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | In this study, we investigated the application of the direct sampling method
(DSM) to identify small dielectric objects in a limited-aperture inverse
scattering problem. Unlike previous studies, we consider the bistatic
measurement configuration corresponding to the transmitter location and design
indicator functions for both a single source and multiple sources, and we
convert the unknown measurement data to a fixed nonzero constant. To explain
the applicability and limitation of object detection, we demonstrate that the
indicator functions can be expressed by an infinite series of Bessel functions,
the material properties of the objects, the bistatic angle, and the converted
constant. Based on the theoretical results, we explain how the imaging
performance of the DSM is influenced by the bistatic angle and the converted
constant. In addition, the results of our analyses demonstrate that a smaller
bistatic angle enhances the imaging accuracy and that optimal selection of the
converted constant is crucial to realize reliable object detection. The results
of the numerical simulations obtained using a two-dimensional Fresnel dataset
validated the theoretical findings and illustrate the effectiveness and
limitations of the designed indicator functions for small objects.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:02:44 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Park",
"Won-Kwang",
""
]
] |
2504.18037 | Yong Chen | Yong Chen, Shi Li, Zihao Liang | Complexity and Approximation Algorithms for Fixed Charge Transportation
Problems | null | null | null | null | cs.DS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The Fixed Charge Transportation (FCT) problem models transportation scenarios
where we need to send a commodity from $n$ sources to $m$ sinks, and the cost
of sending a commodity from a source to a sink consists of a linear component
and a fixed component. Despite extensive research on exponential time exact
algorithms and heuristic algorithms for FCT and its variants, their
approximability and computational complexity are not well understood.
In this work, we initiate a systematic study of the approximability and
complexity of these problems. When there are no linear costs, we call the
problem the Pure Fixed Charge Transportation (PFCT) problem. We also
distinguish between cases with general, sink-independent, and uniform fixed
costs; we use the suffixes ``-S'' and ``-U'' to denote the latter two cases,
respectively. This gives us six variants of the FCT problem.
We give a complete characterization of the existence of $O(1)$-approximation
algorithms for these variants. In particular, we give $2$-approximation
algorithms for FCT-U and PFCT-S, and a $(6/5 + \epsilon)$-approximation for
PFCT-U. On the negative side, we prove that FCT and PFCT are NP-hard to
approximate within a factor of $O(\log^{2-\epsilon} (\max\{n, m\}))$ for any
constant $\epsilon > 0$, FCT-S is NP-hard to approximate within a factor of
$c\log (\max\{n, m\})$ for some constant $c> 0$, and PFCT-U is APX-hard.
Additionally, we design an Efficient Parameterized Approximation Scheme (EPAS)
for PFCT when parameterized by the number $n$ of sources, and an
$O(1/\epsilon)$-bicriteria approximation for the FCT problem, when we are
allowed to violate the demand constraints for sinks by a factor of $1\pm
\epsilon$.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:03:18 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chen",
"Yong",
""
],
[
"Li",
"Shi",
""
],
[
"Liang",
"Zihao",
""
]
] |
2504.18038 | Pedro Soto | Pedro Soto | Optimal Secure Coded Distributed Computation over all Fields | null | null | null | null | cs.IT cs.DC cs.DS cs.SC math.AG math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We construct optimal secure coded distributed schemes that extend the known
optimal constructions over fields of characteristic 0 to all fields. A
serendipitous result is that we can encode \emph{all} functions over finite
fields with a recovery threshold proportional to the complexity (tensor rank or
multiplicative); this is due to the well-known result that all functions over a
finite field can be represented as multivariate polynomials (or symmetric
tensors). We get that a tensor of order $\ell$ (or a multivariate polynomial of
degree $\ell$) can be computed in the faulty network of $N$ nodes setting
within a factor of $\ell$ and an additive term depending on the genus of a code
with $N$ rational points and distance covering the number of faulty servers; in
particular, we present a coding scheme for general matrix multiplication of two
$m \times m $ matrices with a recovery threshold of $2 m^{\omega } -1+g$ where
$\omega $ is the exponent of matrix multiplication which is optimal for coding
schemes using AG codes. Moreover, we give sufficient conditions for which the
Hadamard-Shur product of general linear codes gives a similar recovery
threshold, which we call \textit{log-additive codes}. Finally, we show that
evaluation codes with a \textit{curve degree} function (first defined in
[Ben-Sasson et al. (STOC '13)]) that have well-behaved zero sets are
log-additive.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:11:04 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Soto",
"Pedro",
""
]
] |
2504.18040 | Xiaoyi Liu | Xiaoyi Liu and Hao Tang | Cabbage: A Differential Growth Framework for Open Surfaces | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose Cabbage, a differential growth framework to model buckling
behavior in 3D open surfaces found in nature-like the curling of flower petals.
Cabbage creates high-quality triangular meshes free of self-intersection.
Cabbage-Shell is driven by edge subdivision which differentially increases
discretization resolution. Shell forces expands the surface, generating
buckling over time. Feature-aware smoothing and remeshing ensures mesh quality.
Corrective collision effectively prevents self-collision even in tight spaces.
We additionally provide Cabbage-Collision, and approximate alternative,
followed by CAD-ready surface generation. Cabbage is the first open-source
effort with this calibre and robustness, outperforming SOTA methods in its
morphological expressiveness, mesh quality, and stably generates large, complex
patterns over hundreds of simulation steps. It is a source not only of
computational modeling, digital fabrication, education, but also high-quality,
annotated data for geometry processing and shape analysis.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:25:13 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Liu",
"Xiaoyi",
""
],
[
"Tang",
"Hao",
""
]
] |
2504.18041 | Shiyue Zhang | Bang An, Shiyue Zhang, Mark Dredze | RAG LLMs are Not Safer: A Safety Analysis of Retrieval-Augmented
Generation for Large Language Models | NAACL 2025 | null | null | null | cs.CL cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Efforts to ensure the safety of large language models (LLMs) include safety
fine-tuning, evaluation, and red teaming. However, despite the widespread use
of the Retrieval-Augmented Generation (RAG) framework, AI safety work focuses
on standard LLMs, which means we know little about how RAG use cases change a
model's safety profile. We conduct a detailed comparative analysis of RAG and
non-RAG frameworks with eleven LLMs. We find that RAG can make models less safe
and change their safety profile. We explore the causes of this change and find
that even combinations of safe models with safe documents can cause unsafe
generations. In addition, we evaluate some existing red teaming methods for RAG
settings and show that they are less effective than when used for non-RAG
settings. Our work highlights the need for safety research and red-teaming
methods specifically tailored for RAG LLMs.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:25:18 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"An",
"Bang",
""
],
[
"Zhang",
"Shiyue",
""
],
[
"Dredze",
"Mark",
""
]
] |
2504.18043 | Gregory Czap | Gregory Czap, Christoph Wolf, Jose Reina-G\'alvez, Mark H. Sherwood,
and Christopher P. Lutz | Magnetic Resonance Imaging of Single Organic Radicals with Sub-Molecular
Resolution | 20 pages, 5 figures | null | null | null | cond-mat.mes-hall cond-mat.other physics.atom-ph physics.chem-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Interest in the magnetism of organic compounds is growing because of new
organic magnets, spin-based electronics and the diverse properties of magnetic
edge states in graphene nanoribbons. Electron spin resonance spectroscopy
combined with the scanning tunneling microscopy has recently been developed as
a powerful tool to address individual magnetic atoms and molecules at the
atomic scale. Here we demonstrate electron spin resonance and magnetic
resonance imaging of all-organic radical anions adsorbed on a protective thin
insulating film grown on a metal support. We show that using the highly
localized exchange field of the magnetic tip apex allows visualization of the
delocalized spin density with sub-molecular resolution, enabling spin-density
tomography that can distinguish similar molecular species. These results
provide new opportunities for visualizing spin density and magnetic
interactions at the atomic scale.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:25:51 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Czap",
"Gregory",
""
],
[
"Wolf",
"Christoph",
""
],
[
"Reina-Gálvez",
"Jose",
""
],
[
"Sherwood",
"Mark H.",
""
],
[
"Lutz",
"Christopher P.",
""
]
] |
2504.18044 | Omid Veisi | Omid Veisi, Sasan Bahrami, Roman Englert, Claudia M\"uller | AI Ethics and Social Norms: Exploring ChatGPT's Capabilities From What
to How | Accepted for presentation at the ACM Conference on Computer-Supported
Cooperative Work and Social Computing (CSCW) 2025. To appear in Proceedings
of the ACM on Human-Computer Interaction (PACM HCI) | null | null | null | cs.CY cs.AI cs.HC cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | Using LLMs in healthcare, Computer-Supported Cooperative Work, and Social
Computing requires the examination of ethical and social norms to ensure safe
incorporation into human life. We conducted a mixed-method study, including an
online survey with 111 participants and an interview study with 38 experts, to
investigate the AI ethics and social norms in ChatGPT as everyday life tools.
This study aims to evaluate whether ChatGPT in an empirical context operates
following ethics and social norms, which is critical for understanding actions
in industrial and academic research and achieving machine ethics. The findings
of this study provide initial insights into six important aspects of AI ethics,
including bias, trustworthiness, security, toxicology, social norms, and
ethical data. Significant obstacles related to transparency and bias in
unsupervised data collection methods are identified as ChatGPT's ethical
concerns.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:26:30 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Veisi",
"Omid",
""
],
[
"Bahrami",
"Sasan",
""
],
[
"Englert",
"Roman",
""
],
[
"Müller",
"Claudia",
""
]
] |
2504.18045 | Asmita Kumari | Asmita Kumari and Saikat Patra | Revealing hidden nonlocality and preparation contextuality for an
arbitrary input Bell inequality | 10 pages, 5 Figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | In recent years, the activation of hidden nonlocality for a mixed entangled
state, admitting a local model, has gained considerable interest. In this
paper, we study the activation of hidden nonlocality and preparation
contextuality for a class of mixed entangled states, using local filtering
operations. For our demonstration, we consider the two-party (Alice and Bob)
one-way communication game known as parity oblivious random access code
(PORAC). The quantum success probability of such $n$-bit PORAC solely depends
on a Bell functional involving $2^{n-1}$ and $n$ dichotomic measurement
settings for Alice and Bob, respectively. Such a Bell functional has two
classical bounds, the local and the preparation non-contextual. We show that
using local filtering operations on local mixed entangled state, the
nonlocality can be revealed for any non-zero value of the mixedness parameter
of the entangled state if $n\geq6$. Further, we show that the preparation
contextuality, which is a comparatively weaker quantum correlation than
nonlocality, can be revealed for any non-zero value of the mixedness parameter
of the entangled state if $n\geq 4$.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:26:37 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Kumari",
"Asmita",
""
],
[
"Patra",
"Saikat",
""
]
] |
2504.18046 | Guohao Huo | Guohao Huo, Zibo Lin, Zitong Wang, Ruiting Dai, Hao Tang | DMS-Net:Dual-Modal Multi-Scale Siamese Network for Binocular Fundus
Image Classification | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ophthalmic diseases pose a significant global health challenge, yet
traditional diagnosis methods and existing single-eye deep learning approaches
often fail to account for binocular pathological correlations. To address this,
we propose DMS-Net, a dual-modal multi-scale Siamese network for binocular
fundus image classification. Our framework leverages weight-shared Siamese
ResNet-152 backbones to extract deep semantic features from paired fundus
images. To tackle challenges such as lesion boundary ambiguity and scattered
pathological distributions, we introduce a Multi-Scale Context-Aware Module
(MSCAM) that integrates adaptive pooling and attention mechanisms for
multi-resolution feature aggregation. Additionally, a Dual-Modal Feature Fusion
(DMFF) module enhances cross-modal interaction through spatial-semantic
recalibration and bidirectional attention, effectively combining global context
and local edge features. Evaluated on the ODIR-5K dataset, DMS-Net achieves
state-of-the-art performance with 80.5% accuracy, 86.1% recall, and 83.8%
Cohen's kappa, demonstrating superior capability in detecting symmetric
pathologies and advancing clinical decision-making for ocular diseases.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:27:28 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Huo",
"Guohao",
""
],
[
"Lin",
"Zibo",
""
],
[
"Wang",
"Zitong",
""
],
[
"Dai",
"Ruiting",
""
],
[
"Tang",
"Hao",
""
]
] |
2504.18047 | Yasser Nabil | Yasser Nabil, Mahmoud Abdelhadi, Sameh Sorour, Hesham ElSawy, Sara A.
Elsayed, and Hossam S. Hassanein | Spatiotemporal Analysis of Parallelized Computing at the Extreme Edge | null | null | null | null | cs.PF cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extreme Edge Computing (EEC) pushes computing even closer to end users than
traditional Multi-access Edge Computing (MEC), harnessing the idle resources of
Extreme Edge Devices (EEDs) to enable low-latency, distributed processing.
However, EEC faces key challenges, including spatial randomness in device
distribution, limited EED computational power necessitating parallel task
execution, vulnerability to failure, and temporal randomness due to variability
in wireless communication and execution times. These challenges highlight the
need for a rigorous analytical framework to evaluate EEC performance. We
present the first spatiotemporal mathematical model for EEC over large-scale
millimeter-wave networks. Utilizing stochastic geometry and an Absorbing
Continuous-Time Markov Chain (ACTMC), the framework captures the complex
interaction between communication and computation performance, including their
temporal overlap during parallel execution. We evaluate two key metrics:
average task response delay and task completion probability. Together, they
provide a holistic view of latency and reliability. The analysis considers
fundamental offloading strategies, including randomized and location-aware
schemes, while accounting for EED failures. Results show that there exists an
optimal task segmentation that minimizes delay. Under limited EED availability,
we investigate a bias-based EEC and MEC collaboration that offloads excess
demand to MEC resources, effectively reducing congestion and improving system
responsiveness.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:30:30 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Nabil",
"Yasser",
""
],
[
"Abdelhadi",
"Mahmoud",
""
],
[
"Sorour",
"Sameh",
""
],
[
"ElSawy",
"Hesham",
""
],
[
"Elsayed",
"Sara A.",
""
],
[
"Hassanein",
"Hossam S.",
""
]
] |
2504.18048 | Daniel Murfet | Zhongtian Chen, Daniel Murfet | Modes of Sequence Models and Learning Coefficients | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | We develop a geometric account of sequence modelling that links patterns in
the data to measurable properties of the loss landscape in transformer
networks. First, we cast conditional sequence distributions into a
Hilbert-space framework and apply tensor decompositions to identify their
principal modes. Truncating the small-amplitude modes yields an effective data
distribution that preserves dominant structure while discarding statistical
detail. Second, we show theoretically that Local Learning Coefficient (LLC)
estimates are insensitive to modes below a data-dependent threshold.
Consequently, the LLC calculated in practice characterises the geometry of the
effective rather than the true distribution. This insight clarifies why
reliable LLC estimates can be obtained even when a network parameter is not a
strict minimiser of the population loss, and it highlights how the inverse
temperature in SGLD acts as a resolution dial on the landscape structure.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:38:10 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chen",
"Zhongtian",
""
],
[
"Murfet",
"Daniel",
""
]
] |
2504.18049 | Xin Li | Xin Li, Wenhui Zhu, Peijie Qiu, Oana M. Dumitrascu, Amal Youssef,
Yalin Wang | A BERT-Style Self-Supervised Learning CNN for Disease Identification
from Retinal Images | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the field of medical imaging, the advent of deep learning, especially the
application of convolutional neural networks (CNNs) has revolutionized the
analysis and interpretation of medical images. Nevertheless, deep learning
methods usually rely on large amounts of labeled data. In medical imaging
research, the acquisition of high-quality labels is both expensive and
difficult. The introduction of Vision Transformers (ViT) and self-supervised
learning provides a pre-training strategy that utilizes abundant unlabeled
data, effectively alleviating the label acquisition challenge while broadening
the breadth of data utilization. However, ViT's high computational density and
substantial demand for computing power, coupled with the lack of localization
characteristics of its operations on image patches, limit its efficiency and
applicability in many application scenarios. In this study, we employ
nn-MobileNet, a lightweight CNN framework, to implement a BERT-style
self-supervised learning approach. We pre-train the network on the unlabeled
retinal fundus images from the UK Biobank to improve downstream application
performance. We validate the results of the pre-trained model on Alzheimer's
disease (AD), Parkinson's disease (PD), and various retinal diseases
identification. The results show that our approach can significantly improve
performance in the downstream tasks. In summary, this study combines the
benefits of CNNs with the capabilities of advanced self-supervised learning in
handling large-scale unlabeled data, demonstrating the potential of CNNs in the
presence of label scarcity.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:38:55 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Li",
"Xin",
""
],
[
"Zhu",
"Wenhui",
""
],
[
"Qiu",
"Peijie",
""
],
[
"Dumitrascu",
"Oana M.",
""
],
[
"Youssef",
"Amal",
""
],
[
"Wang",
"Yalin",
""
]
] |
2504.18050 | Mingwei Zheng | Mingwei Zheng, Danning Xie, Qingkai Shi, Chengpeng Wang, Xiangyu Zhang | Validating Network Protocol Parsers with Traceable RFC Document
Interpretation | null | null | 10.1145/3728955 | null | cs.SE cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Validating the correctness of network protocol implementations is highly
challenging due to the oracle and traceability problems. The former determines
when a protocol implementation can be considered buggy, especially when the
bugs do not cause any observable symptoms. The latter allows developers to
understand how an implementation violates the protocol specification, thereby
facilitating bug fixes. Unlike existing works that rarely take both problems
into account, this work considers both and provides an effective solution using
recent advances in large language models (LLMs). Our key observation is that
network protocols are often released with structured specification documents,
a.k.a. RFC documents, which can be systematically translated to formal protocol
message specifications via LLMs. Such specifications, which may contain errors
due to the hallucination of LLMs, are used as a quasi-oracle to validate
protocol parsers, while the validation results in return gradually refine the
oracle. Since the oracle is derived from the document, any bugs we find in a
protocol implementation can be traced back to the document, thus addressing the
traceability problem. We have extensively evaluated our approach using nine
network protocols and their implementations written in C, Python, and Go. The
results show that our approach outperforms the state-of-the-art and has
detected 69 bugs, with 36 confirmed. The project also demonstrates the
potential for fully automating software validation based on natural language
specifications, a process previously considered predominantly manual due to the
need to understand specification documents and derive expected outputs for test
inputs.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:39:19 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Zheng",
"Mingwei",
""
],
[
"Xie",
"Danning",
""
],
[
"Shi",
"Qingkai",
""
],
[
"Wang",
"Chengpeng",
""
],
[
"Zhang",
"Xiangyu",
""
]
] |
2504.18051 | Ryusuke Ikeda | Yuto Yokota, Dai Nakashima, Ryusuke Ikeda | Stability and Response of Vortex Solid Formed in Second Landau Level | 18 pages, 6 figures | null | null | null | cond-mat.supr-con | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Physical properties of the vortex solid phase formed in the second Landau
level (2LL), which may be stabilized by strong paramagnetic pairbreaking (PPB),
are examined in type II limit with no magnetic screening. First, it is shown
that the spectrum of the low energy mode of this vortex solid has the same form
as that of the conventional vortex solid in the first (i.e., the lowest) Landau
level. Using this result, the melting line of the 2LL vortex solid is examined
according to the Lindemann criterion. In contrast to the properties in
equlibrium, the electromagnetic response of this vortex solid is quite unusual:
Reflecting the presence of antivortices supporting the stability of the lattice
structure, the superfluid stiffness measuring the response for a current
perpendicular to the magnetic field is found to be nonvanishing, and its sign
depends upon the applied current direction. Consequences of this response
property are briefly discussed.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:40:34 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yokota",
"Yuto",
""
],
[
"Nakashima",
"Dai",
""
],
[
"Ikeda",
"Ryusuke",
""
]
] |
2504.18052 | Chuangchuang Kang | Yaxi Jiang, Chuangchuang Kang, Jiafeng L\"u | Manin triples, bialgebras and Yang-Baxter equation of $A_3$-associative
algebras | null | null | null | null | math.RA | http://creativecommons.org/licenses/by/4.0/ | $A_3$-associative algebra is a generalization of associative algebra and is
one of the four remarkable types of Lie-admissible algebras, along with
associative algebra, left-symmetric algebra and right-symmetric algebra. This
paper develops bialgebra theory for $A_3$-associative algebras. We introduce
Manin triples and bialgebras for $A_3$-associative algebras, prove their
equivalence using matched pairs of $A_3$-associative algebras, and define the
$A_3$-associative Yang-Baxter equation and triangular $A_3$-associative
bialgebras. Additionally, we introduce relative Rota-Baxter operators to
provide skew-symmetric solutions of the $A_3$-associative Yang-Baxter equation.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:43:18 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Jiang",
"Yaxi",
""
],
[
"Kang",
"Chuangchuang",
""
],
[
"Lü",
"Jiafeng",
""
]
] |
2504.18053 | Xingyuan Bu | Jianyu Liu, Hangyu Guo, Ranjie Duan, Xingyuan Bu, Yancheng He, Shilong
Li, Hui Huang, Jiaheng Liu, Yucheng Wang, Chenchen Jing, Xingwei Qu, Xiao
Zhang, Yingshui Tan, Yanan Wu, Jihao Gu, Yangguang Li, Jianke Zhu | DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal
Large Language Models | [NAACL 2025] The first four authors contribute equally, 23 pages,
repo at https://github.com/Kizna1ver/DREAM | null | null | null | cs.CL cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal Large Language Models (MLLMs) pose unique safety challenges due to
their integration of visual and textual data, thereby introducing new
dimensions of potential attacks and complex risk combinations. In this paper,
we begin with a detailed analysis aimed at disentangling risks through
step-by-step reasoning within multimodal inputs. We find that systematic
multimodal risk disentanglement substantially enhances the risk awareness of
MLLMs. Via leveraging the strong discriminative abilities of multimodal risk
disentanglement, we further introduce \textbf{DREAM}
(\textit{\textbf{D}isentangling \textbf{R}isks to \textbf{E}nhance Safety
\textbf{A}lignment in \textbf{M}LLMs}), a novel approach that enhances safety
alignment in MLLMs through supervised fine-tuning and iterative Reinforcement
Learning from AI Feedback (RLAIF). Experimental results show that DREAM
significantly boosts safety during both inference and training phases without
compromising performance on normal tasks (namely oversafety), achieving a
16.17\% improvement in the SIUO safe\&effective score compared to GPT-4V. The
data and code are available at https://github.com/Kizna1ver/DREAM.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:54:24 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Liu",
"Jianyu",
""
],
[
"Guo",
"Hangyu",
""
],
[
"Duan",
"Ranjie",
""
],
[
"Bu",
"Xingyuan",
""
],
[
"He",
"Yancheng",
""
],
[
"Li",
"Shilong",
""
],
[
"Huang",
"Hui",
""
],
[
"Liu",
"Jiaheng",
""
],
[
"Wang",
"Yucheng",
""
],
[
"Jing",
"Chenchen",
""
],
[
"Qu",
"Xingwei",
""
],
[
"Zhang",
"Xiao",
""
],
[
"Tan",
"Yingshui",
""
],
[
"Wu",
"Yanan",
""
],
[
"Gu",
"Jihao",
""
],
[
"Li",
"Yangguang",
""
],
[
"Zhu",
"Jianke",
""
]
] |
2504.18054 | Xiang Zhong | Eric T. Chung, Changqing Ye, Xiang Zhong | A locking free multiscale method for linear elasticity in
stress-displacement formulation with high contrast coefficients | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Achieving strongly symmetric stress approximations for linear elasticity
problems in high-contrast media poses a significant computational challenge.
Conventional methods often struggle with prohibitively high computational costs
due to excessive degrees of freedom, limiting their practical applicability. To
overcome this challenge, we introduce an efficient multiscale model reduction
method and a computationally inexpensive coarse-grid simulation technique for
linear elasticity equations in highly heterogeneous, high-contrast media. We
first utilize a stable stress-displacement mixed finite element method to
discretize the linear elasticity problem and then present the construction of
multiscale basis functions for the displacement and the stress. The mixed
formulation offers several advantages such as direct stress computation without
post-processing, local momentum conservation (ensuring physical consistency),
and robustness against locking effects, even for nearly incompressible
materials. Theoretical analysis confirms that our method is inf-sup stable and
locking-free, with first-order convergence relative to the coarse mesh size.
Notably, the convergence remains independent of contrast ratios as enlarging
oversampling regions. Numerical experiments validate the method's
effectiveness, demonstrating its superior performance even under extreme
contrast conditions.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 03:55:16 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chung",
"Eric T.",
""
],
[
"Ye",
"Changqing",
""
],
[
"Zhong",
"Xiang",
""
]
] |
2504.18055 | Xiaoye Zheng | Xiaoye Zheng (1), Zhiyuan Wan (1 and 2), David Lo (3), Difan Xie (2),
and Xiaohu Yang (1) ((1) Zhejiang University, (2) Hangzhou High-Tech Zone
(Binjiang) Institute of Blockchain and Data Security, (3) Singapore
Management University) | Why Does My Transaction Fail? A First Look at Failed Transactions on the
Solana Blockchain | null | null | 10.1145/3728943 | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | Solana is an emerging blockchain platform, recognized for its high throughput
and low transaction costs, positioning it as a preferred infrastructure for
Decentralized Finance (DeFi), Non-Fungible Tokens (NFTs), and other Web 3.0
applications. In the Solana ecosystem, transaction initiators submit various
instructions to interact with a diverse range of Solana smart contracts, among
which are decentralized exchanges (DEXs) that utilize automated market makers
(AMMs), allowing users to trade cryptocurrencies directly on the blockchain
without the need for intermediaries. Despite the high throughput and low
transaction costs of Solana, the advantages have exposed Solana to bot spamming
for financial exploitation, resulting in the prevalence of failed transactions
and network congestion.
Prior work on Solana has mainly focused on the evaluation of the performance
of the Solana blockchain, particularly scalability and transaction throughput,
as well as on the improvement of smart contract security, leaving a gap in
understanding the characteristics and implications of failed transactions on
Solana. To address this gap, we conducted a large-scale empirical study of
failed transactions on Solana, using a curated dataset of over 1.5 billion
failed transactions across more than 72 million blocks. Specifically, we first
characterized the failed transactions in terms of their initiators,
failure-triggering programs, and temporal patterns, and compared their block
positions and transaction costs with those of successful transactions. We then
categorized the failed transactions by the error messages in their error logs,
and investigated how specific programs and transaction initiators are
associated with these errors...
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:05:28 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Zheng",
"Xiaoye",
"",
"1 and 2"
],
[
"Wan",
"Zhiyuan",
"",
"1 and 2"
],
[
"Lo",
"David",
""
],
[
"Xie",
"Difan",
""
],
[
"Yang",
"Xiaohu",
""
]
] |
2504.18056 | Takumi Nakao | Takumi Nakao, Kenji Koide, Aoki Takanose, Shuji Oishi, Masashi
Yokozuka, Hisashi Date | Range-based 6-DoF Monte Carlo SLAM with Gradient-guided Particle Filter
on GPU | IEEE International Conference on Robotics and Automation | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents range-based 6-DoF Monte Carlo SLAM with a gradient-guided
particle update strategy. While non-parametric state estimation methods, such
as particle filters, are robust in situations with high ambiguity, they are
known to be unsuitable for high-dimensional problems due to the curse of
dimensionality. To address this issue, we propose a particle update strategy
that improves the sampling efficiency by using the gradient information of the
likelihood function to guide particles toward its mode. Additionally, we
introduce a keyframe-based map representation that represents the global map as
a set of past frames (i.e., keyframes) to mitigate memory consumption. The
keyframe poses for each particle are corrected using a simple loop closure
method to maintain trajectory consistency. The combination of gradient
information and keyframe-based map representation significantly enhances
sampling efficiency and reduces memory usage compared to traditional RBPF
approaches. To process a large number of particles (e.g., 100,000 particles) in
real-time, the proposed framework is designed to fully exploit GPU parallel
processing. Experimental results demonstrate that the proposed method exhibits
extreme robustness to state ambiguity and can even deal with kidnapping
situations, such as when the sensor moves to different floors via an elevator,
with minimal heuristics.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:07:08 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Nakao",
"Takumi",
""
],
[
"Koide",
"Kenji",
""
],
[
"Takanose",
"Aoki",
""
],
[
"Oishi",
"Shuji",
""
],
[
"Yokozuka",
"Masashi",
""
],
[
"Date",
"Hisashi",
""
]
] |
2504.18057 | Jiayi Chen | Jiayi Chen, Shuai Wang, Guoliang Li, Wei Xu, Guangxu Zhu, Derrick Wing
Kwan Ng and Chengzhong Xu | Opportunistic Collaborative Planning with Large Vision Model Guided
Control and Joint Query-Service Optimization | null | null | null | null | cs.RO cs.AI | http://creativecommons.org/licenses/by/4.0/ | Navigating autonomous vehicles in open scenarios is a challenge due to the
difficulties in handling unseen objects. Existing solutions either rely on
small models that struggle with generalization or large models that are
resource-intensive. While collaboration between the two offers a promising
solution, the key challenge is deciding when and how to engage the large model.
To address this issue, this paper proposes opportunistic collaborative planning
(OCP), which seamlessly integrates efficient local models with powerful cloud
models through two key innovations. First, we propose large vision model guided
model predictive control (LVM-MPC), which leverages the cloud for LVM
perception and decision making. The cloud output serves as a global guidance
for a local MPC, thereby forming a closed-loop perception-to-control system.
Second, to determine the best timing for large model query and service, we
propose collaboration timing optimization (CTO), including object detection
confidence thresholding (ODCT) and cloud forward simulation (CFS), to decide
when to seek cloud assistance and when to offer cloud service. Extensive
experiments show that the proposed OCP outperforms existing methods in terms of
both navigation time and success rate.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:07:21 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chen",
"Jiayi",
""
],
[
"Wang",
"Shuai",
""
],
[
"Li",
"Guoliang",
""
],
[
"Xu",
"Wei",
""
],
[
"Zhu",
"Guangxu",
""
],
[
"Ng",
"Derrick Wing Kwan",
""
],
[
"Xu",
"Chengzhong",
""
]
] |
2504.18058 | Wen Yu Chang Morris | Sijia Cheng, Wen-Yu Chang, Yun-Nung Chen | Exploring Personality-Aware Interactions in Salesperson Dialogue Agents | Accepted by IWSDS 2025 | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | The integration of dialogue agents into the sales domain requires a deep
understanding of how these systems interact with users possessing diverse
personas. This study explores the influence of user personas, defined using the
Myers-Briggs Type Indicator (MBTI), on the interaction quality and performance
of sales-oriented dialogue agents. Through large-scale testing and analysis, we
assess the pre-trained agent's effectiveness, adaptability, and personalization
capabilities across a wide range of MBTI-defined user types. Our findings
reveal significant patterns in interaction dynamics, task completion rates, and
dialogue naturalness, underscoring the future potential for dialogue agents to
refine their strategies to better align with varying personality traits. This
work not only provides actionable insights for building more adaptive and
user-centric conversational systems in the sales domain but also contributes
broadly to the field by releasing persona-defined user simulators. These
simulators, unconstrained by domain, offer valuable tools for future research
and demonstrate the potential for scaling personalized dialogue systems across
diverse applications.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:10:25 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Cheng",
"Sijia",
""
],
[
"Chang",
"Wen-Yu",
""
],
[
"Chen",
"Yun-Nung",
""
]
] |
2504.18059 | Prachi Garg Ms | Prachi Garg, Joseph K J, Vineeth N Balasubramanian, Necati Cihan
Camgoz, Chengde Wan, Kenrick Kin, Weiguang Si, Shugao Ma, Fernando De La
Torre | POET: Prompt Offset Tuning for Continual Human Action Adaptation | ECCV 2024 (Oral), webpage
https://humansensinglab.github.io/POET-continual-action-recognition/ | ECCV 2024, Lecture Notes in Computer Science, vol. 15122,
Springer, 2025, pp. 436-455 | 10.1007/978-3-031-73039-9_25 | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As extended reality (XR) is redefining how users interact with computing
devices, research in human action recognition is gaining prominence. Typically,
models deployed on immersive computing devices are static and limited to their
default set of classes. The goal of our research is to provide users and
developers with the capability to personalize their experience by adding new
action classes to their device models continually. Importantly, a user should
be able to add new classes in a low-shot and efficient manner, while this
process should not require storing or replaying any of user's sensitive
training data. We formalize this problem as privacy-aware few-shot continual
action recognition. Towards this end, we propose POET: Prompt-Offset Tuning.
While existing prompt tuning approaches have shown great promise for continual
learning of image, text, and video modalities; they demand access to
extensively pretrained transformers. Breaking away from this assumption, POET
demonstrates the efficacy of prompt tuning a significantly lightweight
backbone, pretrained exclusively on the base class data. We propose a novel
spatio-temporal learnable prompt offset tuning approach, and are the first to
apply such prompt tuning to Graph Neural Networks. We contribute two new
benchmarks for our new problem setting in human action recognition: (i) NTU
RGB+D dataset for activity recognition, and (ii) SHREC-2017 dataset for hand
gesture recognition. We find that POET consistently outperforms comprehensive
benchmarks. Source code at
https://github.com/humansensinglab/POET-continual-action-recognition.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:11:24 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Garg",
"Prachi",
""
],
[
"J",
"Joseph K",
""
],
[
"Balasubramanian",
"Vineeth N",
""
],
[
"Camgoz",
"Necati Cihan",
""
],
[
"Wan",
"Chengde",
""
],
[
"Kin",
"Kenrick",
""
],
[
"Si",
"Weiguang",
""
],
[
"Ma",
"Shugao",
""
],
[
"De La Torre",
"Fernando",
""
]
] |
2504.18060 | Robert Osburn | Robert Osburn, Matthias Storzer | Arborescent links and modular tails | 25 pages | null | null | MPIM-Bonn-2025 | math.GT math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We prove an explicit formula for the tail of the colored Jones polynomial for
a class of arborescent links in terms of a product of theta functions and/or
false theta functions. We also provide numerical evidence towards a
classification of the modularity of tails of the colored Jones polynomial for
alternating knots.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:16:13 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Osburn",
"Robert",
""
],
[
"Storzer",
"Matthias",
""
]
] |
2504.18061 | Andrei T. Patrascu | Andrei Tudor Patrascu | Uhlmann-Higgs stabilisation of modified causal structures | null | null | null | null | physics.gen-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present a dynamical gauge theory for purified quantum fields, based on the
Uhlmann connection in the extended system ancilla Hilbert space, and
demonstrate that nontrivial Uhlmann holonomies induce global modifications of
spacetime causal structure. By coupling purified electromagnetic modes to an
emergent U(n) gauge field, we derive an effective metric in the geometric
optics limit whose curvature contributions tilt and stretch lightcones.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:16:23 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Patrascu",
"Andrei Tudor",
""
]
] |
2504.18062 | Lingyan Bao | Lingyan Bao, Sinwoong Yun, Jemin Lee, and Tony Q.S. Quek | LLM-Guided Open RAN: Empowering Hierarchical RAN Intelligent Control | null | null | null | null | cs.NI cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent advancements in large language models (LLMs) have led to a significant
interest in deploying LLMempowered algorithms for wireless communication
networks. Meanwhile, open radio access network (O-RAN) techniques offer
unprecedented flexibility, with the non-real-time (non-RT) radio access network
(RAN) intelligent controller (RIC) (non-RT RIC) and near-real-time (near-RT)
RIC (near-RT RIC) components enabling intelligent resource management across
different time scales. In this paper, we propose the LLM empowered hierarchical
RIC (LLM-hRIC) framework to improve the collaboration between RICs. This
framework integrates LLMs with reinforcement learning (RL) for efficient
network resource management. In this framework, LLMs-empowered non-RT RICs
provide strategic guidance and high-level policies based on environmental
context. Concurrently, RL-empowered near-RT RICs perform low-latency tasks
based on strategic guidance and local near-RT observation. We evaluate the
LLM-hRIC framework in an integrated access and backhaul (IAB) network setting.
Simulation results demonstrate that the proposed framework achieves superior
performance. Finally, we discuss the key future challenges in applying LLMs to
O-RAN.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:18:23 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Bao",
"Lingyan",
""
],
[
"Yun",
"Sinwoong",
""
],
[
"Lee",
"Jemin",
""
],
[
"Quek",
"Tony Q. S.",
""
]
] |
2504.18063 | Feng Mei Prof. | Yi Li, Jia-Hui Zhang, Yang Kou, Liantuan Xiao, Suotang Jia, Linhu Li,
Feng Mei | Observation of gauge field induced non-Hermitian helical spin skin
effects | 10 pages, 5 figures, | null | null | null | cond-mat.mes-hall physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Synthetic gauge fields and non-Hermitian skin effects are pivotal to
topological phases and non-Hermitian physics, each recently attracting great
interest across diverse research fields. Notably, realizing skin effects
typically require nonreciprocal couplings or on-site gain and loss. Here, we
theoretically and experimentally report that, under gauge fields, non-Hermitian
systems with reciprocal couplings and without gain and loss, can nontrivially
give rise to an unprecedented nonreciprocal skin effect, featuring helical
spin-dependent skin accumulations, dubbed as the ``helical spin skin effect".
Specifically, we construct a lattice model with reciprocal dissipative
couplings and gauge fields, implemented in circuit metamaterials. Before
introducing the gauge fields, this model exhibits localized spin edge modes and
extended bulk modes, without skin effects. As the gauge field strength is
applied from $0$ to $\pi$, we observe the emergence of two distinct spin skin
effects and their transitions: the hybrid-order and second-order helical spin
skin effects. Furthermore, we show that these effects can be harnessed to
realize helical spin-dependent nonreciprocal transmissions. Our findings not
only highlight gauge field induced and enriched non-Hermitian topology, but
also brings spin-dependent helicity into skin effects and nonreciprocality.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:24:07 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Li",
"Yi",
""
],
[
"Zhang",
"Jia-Hui",
""
],
[
"Kou",
"Yang",
""
],
[
"Xiao",
"Liantuan",
""
],
[
"Jia",
"Suotang",
""
],
[
"Li",
"Linhu",
""
],
[
"Mei",
"Feng",
""
]
] |
2504.18064 | Rui Chen | Siwei Liang, Yixuan Guan, Jing Xu, Hongyu Qian, Xiangjun Zhang, Dan
Wu, Wenbo Ding, Rui Chen | AllTact Fin Ray: A Compliant Robot Gripper with Omni-Directional Tactile
Sensing | null | null | null | null | cs.RO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Tactile sensing plays a crucial role in robot grasping and manipulation by
providing essential contact information between the robot and the environment.
In this paper, we present AllTact Fin Ray, a novel compliant gripper design
with omni-directional and local tactile sensing capabilities. The finger body
is unibody-casted using transparent elastic silicone, and a camera positioned
at the base of the finger captures the deformation of the whole body and the
contact face. Due to the global deformation of the adaptive structure, existing
vision-based tactile sensing approaches that assume constant illumination are
no longer applicable. To address this, we propose a novel sensing method where
the global deformation is first reconstructed from the image using edge
features and spatial constraints. Then, detailed contact geometry is computed
from the brightness difference against a dynamically retrieved reference image.
Extensive experiments validate the effectiveness of our proposed gripper design
and sensing method in contact detection, force estimation, object grasping, and
precise manipulation.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:32:25 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Liang",
"Siwei",
""
],
[
"Guan",
"Yixuan",
""
],
[
"Xu",
"Jing",
""
],
[
"Qian",
"Hongyu",
""
],
[
"Zhang",
"Xiangjun",
""
],
[
"Wu",
"Dan",
""
],
[
"Ding",
"Wenbo",
""
],
[
"Chen",
"Rui",
""
]
] |
2504.18065 | Mawei Wu | Mawei Wu | Mackey functors from the viewpoint of double category theories | 11 pages | null | null | null | math.GR math.CT math.RT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let $G$ be a finite group. In this paper, we first introduce a new notion,
so-called the Mackey double category of $G$. Then we prove that the category of
Mackey double categories and the category of Mackey functors of $G$ are
equivalent.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:32:28 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Wu",
"Mawei",
""
]
] |
2504.18066 | Masakazu Yamamoto | Masakazu Yamamoto | Logarithmic evolutions on the incompressible Navier--Stokes flow | 10 pages | null | null | null | math.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Through the asymptotic expansion, large-time behavior of incompressible
Navier--Stokes flow in $n$ dimensional whole space is drawn. In particular, the
logarithmic evolution included in flow velocity is focused. When the components
of velocity are ordered from major to minor according to the parabolic scale,
logarithmically evolving components appear in a certain pattern. This work
asserts that this pattern varies depending on the evenness and oddness of the
space dimension. In the preceding works, the expansion with $2n$th order was
already derived. The assertion is derived by reexamining these works in detail.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:33:40 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yamamoto",
"Masakazu",
""
]
] |
2504.18068 | Zhuohao Yan | Zhuohao Yan, Shaoquan Feng, Xingxing Li, Yuxuan Zhou, Chunxi Xia,
Shengyu Li | S3MOT: Monocular 3D Object Tracking with Selective State Space Model | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate and reliable multi-object tracking (MOT) in 3D space is essential
for advancing robotics and computer vision applications. However, it remains a
significant challenge in monocular setups due to the difficulty of mining 3D
spatiotemporal associations from 2D video streams. In this work, we present
three innovative techniques to enhance the fusion and exploitation of
heterogeneous cues for monocular 3D MOT: (1) we introduce the Hungarian State
Space Model (HSSM), a novel data association mechanism that compresses
contextual tracking cues across multiple paths, enabling efficient and
comprehensive assignment decisions with linear complexity. HSSM features a
global receptive field and dynamic weights, in contrast to traditional linear
assignment algorithms that rely on hand-crafted association costs. (2) We
propose Fully Convolutional One-stage Embedding (FCOE), which eliminates ROI
pooling by directly using dense feature maps for contrastive learning, thus
improving object re-identification accuracy under challenging conditions such
as varying viewpoints and lighting. (3) We enhance 6-DoF pose estimation
through VeloSSM, an encoder-decoder architecture that models temporal
dependencies in velocity to capture motion dynamics, overcoming the limitations
of frame-based 3D inference. Experiments on the KITTI public test benchmark
demonstrate the effectiveness of our method, achieving a new state-of-the-art
performance of 76.86~HOTA at 31~FPS. Our approach outperforms the previous best
by significant margins of +2.63~HOTA and +3.62~AssA, showcasing its robustness
and efficiency for monocular 3D MOT tasks. The code and models are available at
https://github.com/bytepioneerX/s3mot.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:45:35 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yan",
"Zhuohao",
""
],
[
"Feng",
"Shaoquan",
""
],
[
"Li",
"Xingxing",
""
],
[
"Zhou",
"Yuxuan",
""
],
[
"Xia",
"Chunxi",
""
],
[
"Li",
"Shengyu",
""
]
] |
2504.18069 | Peter Jenniskens | Stuart Pilorz, Peter Jenniskens | Sun Close-Encounter Model of Long-Period Comet and Meteoroid Orbit
Stochastic Evolution | 23 pages, 20 figures | Icarus 437 (2025) id.116559 | 10.1016/j.icarus.2025.116559 | null | astro-ph.EP | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The dynamical evolution of long-period comets (LPCs) and their meteoroid
streams is usually described with the Sun as the primary body, but over most of
their orbits the Solar System barycenter (SSB) is effectively the orbital
focus. Detailed numerical integrations show that the orbital elements in the
barycentric reference frame are nearly constant, except within the orbit of
Jupiter where the comet or meteoroid shifts to a heliocentric orbit. Here we
show that this encounter can be modeled in the barycentric frame analogously to
how planetary close encounters are treated in the heliocentric frame, with the
comet captured into an elliptic orbit about the Sun as it in turn orbits SSB.
Modeling the encounters as a two-body interaction in the SSB frame gives a
different insight into the dynamics than offered by secular perturbation
analyses, and reveals that a large portion of the stochasticity seen in the
evolution of the comet's orbit is due to the Sun's state relative to SSB at the
time of encounter. LPCs sample the Sun's state randomly at each return, so that
a statistical characterization of Sun's state is sufficient to determine the
qualitative evolution of their orbits, including stream dispersion. The
barycentric orbital elements are shown to execute random walks
well-characterized by Maxwellian distributions. This is superimposed atop a
systematic orbital precession induced by planetary torques. Planetary close
encounters add a second stochastic component, but this component does not
typically dominate the solar perturbations. Based on the statistics of Sun's
state alone, the age of a long-period comet meteoroid stream in a given orbit
can be derived to reasonable precision from the observed random dispersion of
the angular orbital elements at Earth.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:45:48 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Pilorz",
"Stuart",
""
],
[
"Jenniskens",
"Peter",
""
]
] |
2504.18070 | Jingjin Wang | Jingjin Wang | PropRAG: Guiding Retrieval with Beam Search over Proposition Paths | Code and data to be released at:
https://github.com/ReLink-Inc/PropRAG | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Retrieval Augmented Generation (RAG) has become the standard non-parametric
approach for equipping Large Language Models (LLMs) with up-to-date knowledge
and mitigating catastrophic forgetting common in continual learning. However,
standard RAG, relying on independent passage retrieval, fails to capture the
interconnected nature of human memory crucial for complex reasoning
(associativity) and contextual understanding (sense-making). While structured
RAG methods like HippoRAG utilize knowledge graphs (KGs) built from triples,
the inherent context loss limits fidelity. We introduce PropRAG, a framework
leveraging contextually rich propositions and a novel beam search algorithm
over proposition paths to explicitly discover multi-step reasoning chains.
Crucially, PropRAG's online retrieval process operates entirely without
invoking generative LLMs, relying instead on efficient graph traversal and
pre-computed embeddings. This avoids online LLM inference costs and potential
inconsistencies during evidence gathering. LLMs are used effectively offline
for high-quality proposition extraction and post-retrieval for answer
generation. PropRAG achieves state-of-the-art zero-shot Recall@5 results on
PopQA (55.3%), 2Wiki (93.7%), HotpotQA (97.0%), and MuSiQue (77.3%), alongside
top F1 scores (e.g., 52.4% on MuSiQue). By improving evidence retrieval through
richer representation and explicit, LLM-free online path finding, PropRAG
advances non-parametric continual learning.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:47:34 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Wang",
"Jingjin",
""
]
] |
2504.18071 | Yuan Ping | Chunhao Guo, Gabriele Riva, Jacopo Simoni, Junqing Xu, Yuan Ping | Phonon-Assisted Radiative Lifetimes and Exciton Dynamics from First
Principles | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Exciton-phonon interactions play a fundamental role in phonon-assisted
radiative recombination and exciton dynamics in solids. In this work, we
present a first-principles framework for computing phonon-assisted radiative
lifetimes and exciton dynamics at finite temperatures. Starting from the
solution of the Bethe-Salpeter equation, we construct an effective excitonic
Hamiltonian that incorporates both exciton-photon and exciton-phonon
interactions. Phonon-assisted radiative lifetimes in anisotropic media are
evaluated using time-dependent second-order perturbation theory. We further
analyze the temperature and phonon-mode dependence of phonon-assisted radiative
lifetime and compare our results with available experimental data. We explain
the nonmonotonic temperature dependence of the phonon-assisted radiative
lifetime by different mechanisms at low and high-temperature regimes. Finally,
we perform real-time exciton relaxation at the diagonal approximation of
Lindbladian dynamics for time-resolved exciton occupation, providing insights
into ultrafast thermalization and scattering pathways. Our ab-initio theory
offers a detailed microscopic understanding of phonon-mediated exciton
relaxation and recombination processes, and provides in-depth perspectives on
phonon-assisted many-body interactions and their influence on optical
properties for light-emitting and optoelectronic applications.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 04:48:27 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Guo",
"Chunhao",
""
],
[
"Riva",
"Gabriele",
""
],
[
"Simoni",
"Jacopo",
""
],
[
"Xu",
"Junqing",
""
],
[
"Ping",
"Yuan",
""
]
] |
2504.18072 | Konstantin Sch\"urholt | Konstantin Sch\"urholt, L\'eo Meynent, Yefan Zhou, Haiquan Lu, Yaoqing
Yang, Damian Borth | A Model Zoo on Phase Transitions in Neural Networks | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Using the weights of trained Neural Network (NN) models as data modality has
recently gained traction as a research field - dubbed Weight Space Learning
(WSL). Multiple recent works propose WSL methods to analyze models, evaluate
methods, or synthesize weights. Weight space learning methods require
populations of trained models as datasets for development and evaluation.
However, existing collections of models - called `model zoos' - are
unstructured or follow a rudimentary definition of diversity. In parallel, work
rooted in statistical physics has identified phases and phase transitions in NN
models. Models are homogeneous within the same phase but qualitatively differ
from one phase to another. We combine the idea of `model zoos' with phase
information to create a controlled notion of diversity in populations. We
introduce 12 large-scale zoos that systematically cover known phases and vary
over model architecture, size, and datasets. These datasets cover different
modalities, such as computer vision, natural language processing, and
scientific ML. For every model, we compute loss landscape metrics and validate
full coverage of the phases. With this dataset, we provide the community with a
resource with a wide range of potential applications for WSL and beyond.
Evidence suggests the loss landscape phase plays a role in applications such as
model training, analysis, or sparsification. We demonstrate this in an
exploratory study of the downstream methods like transfer learning or model
weights averaging.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:01:52 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Schürholt",
"Konstantin",
""
],
[
"Meynent",
"Léo",
""
],
[
"Zhou",
"Yefan",
""
],
[
"Lu",
"Haiquan",
""
],
[
"Yang",
"Yaoqing",
""
],
[
"Borth",
"Damian",
""
]
] |
2504.18073 | Jimin Kim | Jimin Kim, Jeremy T. Florman, Julia A. Santos, Mark J. Alkema, Eli
Shlizerman | Modular integration of neural connectomics, dynamics and biomechanics
for identification of behavioral sensorimotor pathways in Caenorhabditis
elegans | null | null | null | null | q-bio.QM q-bio.NC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Computational approaches which emulate in-vivo nervous system are needed to
investigate mechanisms of the brain to orchestrate behavior. Such approaches
must integrate a series of biophysical models encompassing the nervous system,
muscles, biomechanics to allow observing the system in its entirety while
supporting model variations. Here we develop modWorm: a modular modeling
framework for the nematode C. elegans. modWorm allows for construction of a
model as an integrated series of configurable, exchangeable modules each
describing specific biophysical processes across different modalities.
Utilizing modWorm, we propose a base neuro-mechanical model for C. elegans
built upon the complete connectome. The model integrates a series of 7 modules:
i) intra-cellular dynamics, ii) electrical and iii) chemical extra-cellular
neural dynamics, iv) translation of neural activity to muscle calcium dynamics,
v) muscle calcium dynamics to muscle forces, vi) muscle forces to body postures
and vii) proprioceptive feedback. We validate the base model by in-silico
injection of constant currents into neurons known to be associated with
locomotion behaviors and by applying external forces to the body. Applications
of in-silico neural stimuli experimentally known to modulate locomotion show
that the model can recapitulate natural behavioral responses such as forward
and backward locomotion as well as mid-locomotion responses such as avoidance
and turns. Furthermore, through in-silico ablation surveys, the model can infer
novel neural circuits involved in sensorimotor behaviors. To further dissect
mechanisms of locomotion, we utilize modWorm to introduce empirical based model
variations and model optimizations to elucidate their effects on simulated
locomotion. Our results show that modWorm can be utilized to identify neural
circuits which control, mediate and generate natural behavior.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:04:03 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Kim",
"Jimin",
""
],
[
"Florman",
"Jeremy T.",
""
],
[
"Santos",
"Julia A.",
""
],
[
"Alkema",
"Mark J.",
""
],
[
"Shlizerman",
"Eli",
""
]
] |
2504.18074 | Zhuyao Jin | Zhu-yao Jin and Jun Jing | Universal quantum control over Rydberg atoms | 12 pages, 6 figures | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | In this paper, the universal quantum control with error correction is applied
to the generation of the Greenberger-Horne-Zeilinger (GHZ) states of multiple
Rydberg atoms, in which the qubits are encoded on the hyperfine ground levels.
Our system is featured with the off-resonant driving fields rather than the
strong Rydberg interaction sufficient for blockade. It can closely follow the
designed nonadiabatic passage during the time evolution and avoid the unwanted
transition by imposing the path-dependent and fast-varying global phase. A
general GHZ state of $N$ Rydberg qubits can be prepared in $N-1$ steps, which
is found to be robust against both environmental noises and systematic errors.
Our protocol therefore provides an avenue towards large-scale entanglement,
which is essential for quantum information and quantum computation based on
neutral atoms.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:06:01 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Jin",
"Zhu-yao",
""
],
[
"Jing",
"Jun",
""
]
] |
2504.18075 | Gurdal Arslan | Jason Castiglione and G\"urdal Arslan | Fictitious Play in Extensive-Form Games of Imperfect Information | null | null | null | null | cs.GT math.OC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We study the long-term behavior of the fictitious play process in repeated
extensive-form games of imperfect information with perfect recall. Each player
maintains incorrect beliefs that the moves at all information sets, except the
one at which the player is about to make a move, are made according to fixed
random strategies, independently across all information sets. Accordingly, each
player makes his moves at any of his information sets to maximize his expected
payoff assuming that, at any other information set, the moves are made
according to the empirical frequencies of the past moves. We extend the
well-known Monderer-Shapley result [1] on the convergence of the empirical
frequencies to the set of Nash equilibria to a certain class of extensive-form
games with identical interests. We then strengthen this result by the use of
inertia and fading memory, and prove the convergence of the realized play-paths
to an essentially pure Nash equilibrium in all extensive-form games of
imperfect information with identical interests.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:07:48 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Castiglione",
"Jason",
""
],
[
"Arslan",
"Gürdal",
""
]
] |
2504.18076 | Ivan Monlineris | Marco Masera, Chiara Cicconetti, Francesca Ferrero, Salvatore Oliviero
and Ivan Molineris | 3plex Web: An Interactive Platform for RNA:DNA Triplex Prediction and
Analysis | null | null | null | null | q-bio.QM | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Summary: Long non-coding RNAs (lncRNAs) exert their functions by cooperating
with other molecules including proteins and DNA. Triplexes, formed through the
interaction between a single-stranded RNA (ssRNA) and a double-stranded DNA
(dsDNA), have been consistently described as a mechanism that allows lncRNAs to
target specific genomic sequences in vivo. Building on the computational tool
3plex, we developed 3plex Web, an accessible platform that enhances RNA:DNA
triplex prediction by integrating interactive visualization, statistical
evaluation, and user-friendly downstream analysis workflows. 3plex Web
implements new features such as input randomization for statistical
assessments, interactive profile plotting for triplex stability, and
customizable DNA Binding Domain (DBD) selection. This platform enables rapid
analysis through PATO, substantially reducing processing times compared to
previous methods, while offering Snakemake workflows to integrate gene
expression data and explore lncRNA regulatory mechanisms.
Availability and implementation: 3plex Web is freely available at
https://3plex.unito.it as an online web service. The source code for 3plex is
available at https://github.com/molinerisLab/3plex, paired with a definition
file to set up the application into a Singularity image.
Contact: [email protected]
Keywords: DNA; RNA; RNA-DNA interaction; triplex; long non-coding RNA;
lncRNA; gene regulation; web application
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:08:57 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Masera",
"Marco",
""
],
[
"Cicconetti",
"Chiara",
""
],
[
"Ferrero",
"Francesca",
""
],
[
"Oliviero",
"Salvatore",
""
],
[
"Molineris",
"Ivan",
""
]
] |
2504.18077 | Jalaja Pandya | Jalaja Pandya, Malika Singhal, Navinder Singh, Naveen Chauhan | DFT Investigations of Major Defects in Quartz Crystal: Implications for
Luminescence and ESR Dosimetry and Dating | 26 pages, 18 figures | null | null | null | cond-mat.mtrl-sci cond-mat.dis-nn cond-mat.other physics.geo-ph | http://creativecommons.org/licenses/by/4.0/ | Quartz is extensively used for luminescence and ESR dosimetry as well as
dating. These techniques use inherent defects introduced in quartz crystal
during its crystallization in nature. The defect comprises both intrinsic as
well as extrinsic defects. These defects give important luminescence properties
to quartz but are not yet well understood from a theoretical perspective.
Specifically, in case of luminescence dosimetry the nature of traps and their
involvement in luminescence production is not exactly known. Thus, present work
attempts to understand the basic physics of defects and their implication for
luminescence and ESR techniques via Density Functional Theory (DFT) modelling.
The work uses DFT to model the presence of some possible major impurities in
quartz. Several interesting novel results are obtained that will have
implications for ongoing research in Luminescence and ESR methods. The DFT
modelling suggested that Oxygen deficiency in quartz crystal results in the
formation of both electron and hole trapping centres. However, it is observed
that these centres can be passivated by the introduction of charge compensating
OH or H ions. Further, it is found that peroxy defects can be formed in the
presence of either excess Oxygen or due to the absence of Silicon (Si4+),
however, the nature of the traps formed in both cases is different. Besides
these intrinsic defects, Al and Fe are the major impurities which are observed
as defects in quartz. The modelling of these impurities suggested that
negligible change in DOS is observed for Al defect and Fe generally forms a
recombination centre or hole trap. In addition to these, there are several
interesting first-time observations that are not reported and will be helpful
for progressing luminescence and ESR dosimetry research.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:09:23 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Pandya",
"Jalaja",
""
],
[
"Singhal",
"Malika",
""
],
[
"Singh",
"Navinder",
""
],
[
"Chauhan",
"Naveen",
""
]
] |
2504.18078 | Hao Wang | Xiaolu Chen, Chenghao Huang, Yanru Zhang, Hao Wang | Privacy-Preserving Personalized Federated Learning for Distributed
Photovoltaic Disaggregation under Statistical Heterogeneity | 11 pages | IEEE Transactions on Instrumentation and Measurement IEEE
Transactions on Instrumentation and Measurement IEEE Transactions on
Instrumentation and Measurement, 2025 | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | The rapid expansion of distributed photovoltaic (PV) installations worldwide,
many being behind-the-meter systems, has significantly challenged energy
management and grid operations, as unobservable PV generation further
complicates the supply-demand balance. Therefore, estimating this generation
from net load, known as PV disaggregation, is critical. Given privacy concerns
and the need for large training datasets, federated learning becomes a
promising approach, but statistical heterogeneity, arising from geographical
and behavioral variations among prosumers, poses new challenges to PV
disaggregation. To overcome these challenges, a privacy-preserving distributed
PV disaggregation framework is proposed using Personalized Federated Learning
(PFL). The proposed method employs a two-level framework that combines local
and global modeling. At the local level, a transformer-based PV disaggregation
model is designed to generate solar irradiance embeddings for representing
local PV conditions. A novel adaptive local aggregation mechanism is adopted to
mitigate the impact of statistical heterogeneity on the local model, extracting
a portion of global information that benefits the local model. At the global
level, a central server aggregates information uploaded from multiple data
centers, preserving privacy while enabling cross-center knowledge sharing.
Experiments on real-world data demonstrate the effectiveness of this proposed
framework, showing improved accuracy and robustness compared to benchmark
methods.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:09:27 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Chen",
"Xiaolu",
""
],
[
"Huang",
"Chenghao",
""
],
[
"Zhang",
"Yanru",
""
],
[
"Wang",
"Hao",
""
]
] |
2504.18079 | Wei Wang | Limeng Lin, Wei Wang, Hao Zhang | An Upper Bound on the Number of Generalized Cospectral Mates of Oriented
Graphs | null | null | null | null | math.CO | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper examines the spectral characterizations of oriented graphs. Let
$\Sigma$ be an $n$-vertex oriented graph with skew-adjacency matrix $S$.
Previous research mainly focused on self-converse oriented graphs, proposing
arithmetic conditions for these graphs to be uniquely determined by their
generalized skew-spectrum ($\mathrm{DGSS}$). However, self-converse graphs are
extremely rare; this paper considers a more general class of oriented graphs
$\mathcal{G}_{n}$ (not limited to self-converse graphs), consisting of all
$n$-vertex oriented graphs $\Sigma$ such that $2^{-\left \lfloor \frac{n}{2}
\right \rfloor }\det W(\Sigma)$ is an odd and square-free integer, where
$W(\Sigma)=[e,Se,\dots,S^{n-1}e]$ ($e$ is the all-one vector) is the skew-walk
matrix of $\Sigma$. Given that $\Sigma$ is cospectral with its converse
$\Sigma^{\rm T}$, there always exists a unique regular rational orthogonal
$Q_0$ such that $Q_0^{\rm T}SQ_0=-S$. This study reveals that there exists a
deep relationship between the level $\ell_0$ of $Q_0$ and the number of
generalized cospectral mates of $\Sigma$. More precisely, we show, among
others, that the maximum number of generalized cospectral mates of
$\Sigma\in\mathcal{G}_{n}$ is at most $2^{t}-1$, where $t$ is the number of
prime factors of $\ell_0$. Moreover, some numerical examples are also provided
to demonstrate that the above upper bound is attainable. Finally, we also
provide a criterion for the oriented graphs $\Sigma\in\mathcal{G}_{n}$ to be
weakly determined by the generalized skew-spectrum ($\mathrm{WDGSS})$.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:11:36 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Lin",
"Limeng",
""
],
[
"Wang",
"Wei",
""
],
[
"Zhang",
"Hao",
""
]
] |
2504.18080 | Junichiro Iwasawa | Wataru Kawakami, Keita Suzuki and Junichiro Iwasawa | Stabilizing Reasoning in Medical LLMs with Continued Pretraining and
Reasoning Preference Optimization | null | null | null | null | cs.CL cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Models (LLMs) show potential in medicine, yet clinical
adoption is hindered by concerns over factual accuracy, language-specific
limitations (e.g., Japanese), and critically, their reliability when required
to generate reasoning explanations -- a prerequisite for trust. This paper
introduces Preferred-MedLLM-Qwen-72B, a 72B-parameter model optimized for the
Japanese medical domain to achieve both high accuracy and stable reasoning. We
employ a two-stage fine-tuning process on the Qwen2.5-72B base model: first,
Continued Pretraining (CPT) on a comprehensive Japanese medical corpus instills
deep domain knowledge. Second, Reasoning Preference Optimization (RPO), a
preference-based method, enhances the generation of reliable reasoning pathways
while preserving high answer accuracy. Evaluations on the Japanese Medical
Licensing Exam benchmark (IgakuQA) show Preferred-MedLLM-Qwen-72B achieves
state-of-the-art performance (0.868 accuracy), surpassing strong proprietary
models like GPT-4o (0.866). Crucially, unlike baseline or CPT-only models which
exhibit significant accuracy degradation (up to 11.5\% and 3.8\% respectively
on IgakuQA) when prompted for explanations, our model maintains its high
accuracy (0.868) under such conditions. This highlights RPO's effectiveness in
stabilizing reasoning generation. This work underscores the importance of
optimizing for reliable explanations alongside accuracy. We release the
Preferred-MedLLM-Qwen-72B model weights to foster research into trustworthy
LLMs for specialized, high-stakes applications.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:15:31 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Kawakami",
"Wataru",
""
],
[
"Suzuki",
"Keita",
""
],
[
"Iwasawa",
"Junichiro",
""
]
] |
2504.18081 | Vinh Truong Dr | Vinh Truong (RMIT University) | Hype and Adoption of Generative Artificial Intelligence Applications | null | null | null | null | cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | New technologies create opportunities while displacing others. They enhance
life by supporting entertainment, education, and social connectivity but also
replace humans in productivity and analytical tasks. Adapting to these shifts
requires technical adjustments and social readiness. For digital transformation
to succeed, organizations and their workforce must be psychologically prepared.
We are entering the era of Generative AI with tools like ChatGPT, Bing AI, and
Microsoft Office Copilot. Understanding public sentiment toward these
innovations is crucial for refining technology acceptance models and informing
market strategies. Using the Gartner Hype Cycle and Kubler-Ross Change Curve,
this study suggests that generative AI adoption is a dual-stage process. It
follows the phases of technology trigger, peak of expectations, trough of
disillusionment, slope of enlightenment, and plateau of productivity, while
also reflecting emotional stages like shock, denial, and integration. The study
used sentiment and emotion analysis on a large dataset of tweets about
generative AI, translating them into scores to track user responses over time.
Unlike prior research, which offered a snapshot of sentiment, this study
captures the dynamic evolution of attitudes, linking empirical evidence with
theoretical frameworks. It shifts the focus from information seekers to content
creators. With the release of generative AI tools, there is a significant gap
in understanding societal reception and adaptation. Policymakers face
uncertainty about guiding markets for these changes. This research validates
the applicability of the Gartner Hype Cycle and Kubler-Ross Change Curve to
generative AI. It provides insights for businesses in integrating these tools
and crafting policies to enhance readiness and resilience.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:15:57 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Truong",
"Vinh",
"",
"RMIT University"
]
] |
2504.18082 | Vignesh Balaji | Vignesh Balaji, Christos Kozyrakis, Gal Chechik, Haggai Maron | Efficient GNN Training Through Structure-Aware Randomized Mini-Batching | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Graph Neural Networks (GNNs) enable learning on realworld graphs and
mini-batch training has emerged as the de facto standard for training GNNs
because it can scale to very large graphs and improve convergence. Current
mini-batch construction policies largely ignore efficiency considerations of
GNN training. Specifically, existing mini-batching techniques employ
randomization schemes to improve accuracy and convergence. However, these
randomization schemes are often agnostic to the structural properties of the
graph (for eg. community structure), resulting in highly irregular memory
access patterns during GNN training that make suboptimal use of on-chip GPU
caches. On the other hand, while deterministic mini-batching based solely on
graph structure delivers fast runtime performance, the lack of randomness
compromises both the final model accuracy and training convergence speed. In
this paper, we present Community-structure-aware Randomized Mini-batching
(COMM-RAND), a novel methodology that bridges the gap between the above
extremes. COMM-RAND allows practitioners to explore the space between pure
randomness and pure graph structural awareness during mini-batch construction,
leading to significantly more efficient GNN training with similar accuracy. We
evaluated COMM-RAND across four popular graph learning benchmarks. COMM-RAND
cuts down GNN training time by up to 2.76x (1.8x on average) while achieving an
accuracy that is within 1.79% points (0.42% on average) compared to popular
random mini-batching approaches.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:16:53 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Balaji",
"Vignesh",
""
],
[
"Kozyrakis",
"Christos",
""
],
[
"Chechik",
"Gal",
""
],
[
"Maron",
"Haggai",
""
]
] |
2504.18083 | Yuqiao Yang | Yuqiao Yang, Yongzhao Zhang, Wenhao Liu, Jun Li, Pengtao Shi, DingYu
Zhong, Jie Yang, Ting Chen, Sheng Cao, Yuntao Ren, Yongyue Wu, Xiaosong Zhang | Automating Function-Level TARA for Automotive Full-Lifecycle Security | null | null | null | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As modern vehicles evolve into intelligent and connected systems, their
growing complexity introduces significant cybersecurity risks. Threat Analysis
and Risk Assessment (TARA) has therefore become essential for managing these
risks under mandatory regulations. However, existing TARA automation methods
rely on static threat libraries, limiting their utility in the detailed,
function-level analyses demanded by industry. This paper introduces
DefenseWeaver, the first system that automates function-level TARA using
component-specific details and large language models (LLMs). DefenseWeaver
dynamically generates attack trees and risk evaluations from system
configurations described in an extended OpenXSAM++ format, then employs a
multi-agent framework to coordinate specialized LLM roles for more robust
analysis. To further adapt to evolving threats and diverse standards,
DefenseWeaver incorporates Low-Rank Adaptation (LoRA) fine-tuning and
Retrieval-Augmented Generation (RAG) with expert-curated TARA reports. We
validated DefenseWeaver through deployment in four automotive security
projects, where it identified 11 critical attack paths, verified through
penetration testing, and subsequently reported and remediated by the relevant
automakers and suppliers. Additionally, DefenseWeaver demonstrated cross-domain
adaptability, successfully applying to unmanned aerial vehicles (UAVs) and
marine navigation systems. In comparison to human experts, DefenseWeaver
outperformed manual attack tree generation across six assessment scenarios.
Integrated into commercial cybersecurity platforms such as UAES and Xiaomi,
DefenseWeaver has generated over 8,200 attack trees. These results highlight
its ability to significantly reduce processing time, and its scalability and
transformative impact on cybersecurity across industries.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:19:02 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Yang",
"Yuqiao",
""
],
[
"Zhang",
"Yongzhao",
""
],
[
"Liu",
"Wenhao",
""
],
[
"Li",
"Jun",
""
],
[
"Shi",
"Pengtao",
""
],
[
"Zhong",
"DingYu",
""
],
[
"Yang",
"Jie",
""
],
[
"Chen",
"Ting",
""
],
[
"Cao",
"Sheng",
""
],
[
"Ren",
"Yuntao",
""
],
[
"Wu",
"Yongyue",
""
],
[
"Zhang",
"Xiaosong",
""
]
] |
2504.18084 | Atsushi Kanehira Mr. | Atsushi Kanehira, Naoki Wake, Kazuhiro Sasabuchi, Jun Takamatsu,
Katsushi Ikeuchi | RL-Driven Data Generation for Robust Vision-Based Dexterous Grasping | null | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This work presents reinforcement learning (RL)-driven data augmentation to
improve the generalization of vision-action (VA) models for dexterous grasping.
While real-to-sim-to-real frameworks, where a few real demonstrations seed
large-scale simulated data, have proven effective for VA models, applying them
to dexterous settings remains challenging: obtaining stable multi-finger
contacts is nontrivial across diverse object shapes. To address this, we
leverage RL to generate contact-rich grasping data across varied geometries. In
line with the real-to-sim-to-real paradigm, the grasp skill is formulated as a
parameterized and tunable reference trajectory refined by a residual policy
learned via RL. This modular design enables trajectory-level control that is
both consistent with real demonstrations and adaptable to diverse object
geometries. A vision-conditioned policy trained on simulation-augmented data
demonstrates strong generalization to unseen objects, highlighting the
potential of our approach to alleviate the data bottleneck in training VA
models.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:21:19 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Kanehira",
"Atsushi",
""
],
[
"Wake",
"Naoki",
""
],
[
"Sasabuchi",
"Kazuhiro",
""
],
[
"Takamatsu",
"Jun",
""
],
[
"Ikeuchi",
"Katsushi",
""
]
] |
2504.18085 | Muhammad Mubashar | Muhammad Mubashar, Shireen Kudukkil Manchingal, Fabio Cuzzolin | Random-Set Large Language Models | 16 pages, 6 figures | null | null | null | cs.CL cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large Language Models (LLMs) are known to produce very high-quality tests and
responses to our queries. But how much can we trust this generated text? In
this paper, we study the problem of uncertainty quantification in LLMs. We
propose a novel Random-Set Large Language Model (RSLLM) approach which predicts
finite random sets (belief functions) over the token space, rather than
probability vectors as in classical LLMs. In order to allow so efficiently, we
also present a methodology based on hierarchical clustering to extract and use
a budget of "focal" subsets of tokens upon which the belief prediction is
defined, rather than using all possible collections of tokens, making the
method scalable yet effective. RS-LLMs encode the epistemic uncertainty induced
in their generation process by the size and diversity of its training set via
the size of the credal sets associated with the predicted belief functions. The
proposed approach is evaluated on CoQA and OBQA datasets using Llama2-7b,
Mistral-7b and Phi-2 models and is shown to outperform the standard model in
both datasets in terms of correctness of answer while also showing potential in
estimating the second level uncertainty in its predictions and providing the
capability to detect when its hallucinating.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:25:27 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Mubashar",
"Muhammad",
""
],
[
"Manchingal",
"Shireen Kudukkil",
""
],
[
"Cuzzolin",
"Fabio",
""
]
] |
2504.18086 | Pram Milan P Robin | Priyotosh Bandyopadhyay, Pram Milan P. Robin | Fate of the scalar quartic couplings in the inert models | 34 pages, 23 figures | null | null | IITH-PH-0001/25 | hep-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this article we consider three inert models, namely the inert singlet
model (ISM), inert triplet model (ITM), and inert doublet model (IDM) as beyond
Standard Model scenarios, and look into the running of the scalar quartic
couplings at one- and two-loop levels. Interestingly, the Landau poles at
one-loop and Fixed points at two-loop have been observed for these scenarios,
very similar to the $\phi^4$ theory, which is absent in the Standard Model. The
role of Higgs portal couplings in attaining these features has been examined.
For inert singlet and triplet models, the portal couplings give rise to terms
without any residual phases, enhancing this Fixed point behaviour. In the case
of the inert doublet model, $\lambda_{4,5}$ terms which have the residual
phases, spoil this behaviour. Finally, we consider perturbative unitarity
constraints to put limits on the scalar quartic couplings for various
perturbativity scales in both one- and two-loop. Larger portal couplings, i.e.
$\gsim 2$, get strong perturbative bounds as low as $10^{2-3}$ GeV.
| [
{
"version": "v1",
"created": "Fri, 25 Apr 2025 05:27:34 GMT"
}
] | 2025-04-28T00:00:00 | [
[
"Bandyopadhyay",
"Priyotosh",
""
],
[
"Robin",
"Pram Milan P.",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.