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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2504.15456 | Henry Bradford | Henry Bradford and Alessandro Sisto | Non-solutions to mixed equations in acylindrically hyperbolic groups
coming from random walks | 6 pages, comments welcome | null | null | null | math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | A mixed equation in a group $G$ is given by a non-trivial element $w (x)$ of
the free product $G \ast \mathbb{Z}$, and a solution is some $g\in G$ such that
$w(g)$ is the identity. For $G$ acylindrically hyperbolic with trivial finite
radical (e.g. torsion-free) we show that any mixed equation of length $n$ has a
non-solution of length comparable to $\log(n)$, which is the best possible
bound. Similarly, we show that there is a common non-solution of length $O(n)$
to all mixed equations of length $n$, again the best possible bound. In fact,
in both cases we show that a random walk of appropriate length yields a
non-solution with positive probability.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 21:52:27 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Bradford",
"Henry",
""
],
[
"Sisto",
"Alessandro",
""
]
] |
2504.15458 | Dustin Keller | Brandon Le and Dustin Keller | Compton Form Factor Extraction using Quantum Deep Neural Networks | null | null | null | null | cs.LG nucl-th quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Extraction tests of Compton Form Factors are performed using pseudodata based
on experimental data from Deeply Virtual Compton Scattering experiments
conducted at Jefferson Lab. The standard Belitsky, Kirchner, and Muller
formalism at twist-two is employed, along with a fitting procedure designed to
reduce model dependency similar to traditional local fits. The extraction of
the Compton Form Factors is performed using both Classical Deep Neural Networks
(CDNNs) and Quantum Deep Neural Networks (QDNNs). Comparative studies reveal
that QDNNs outperform CDNNs for this application, demonstrating improved
predictive accuracy and precision even for limited model complexity. The
results demonstrate the potential of QDNNs for future studies in which quantum
algorithms can be fully optimized.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 21:56:49 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Le",
"Brandon",
""
],
[
"Keller",
"Dustin",
""
]
] |
2504.15459 | Leo Diaz | Leopoldo Diaz, Harold P. Hjalmarson, Jesse J. Lutz, and Peter A.
Schultz | The case of the missing gallium vacancy in gallium arsenide: A
multiscale explanation | null | null | null | null | cond-mat.mtrl-sci physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | Irradiation of gallium arsenide (GaAs) produces immobile vacancies and mobile
interstitials. However, after decades of experimental investigation, the
immobile Ga vacancy eludes observation, raising the question: Where is the Ga
vacancy? Static first-principles calculations predict a Ga vacancy should be
readily observed. We find that short-time dynamical evolution of primary
defects is key to explaining this conundrum. Introducing a multiscale
Atomistically Informed Device Engineering (AIDE) method, we discover that
during the initial displacement damage, the Fermi level shifts to mid-gap
producing oppositely charged vacancies and interstitials. Driven by Coulomb
attraction, fast As interstitials preferentially annihilate Ga vacancies,
causing their population to plummet below detectable limits before being
experimentally observed. This innovative model solves the mystery of the
missing Ga vacancy and reveals the importance of a multiscale approach to
explore the dynamical chemical behavior in experimentally inaccessible
short-time regimes.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 21:58:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Diaz",
"Leopoldo",
""
],
[
"Hjalmarson",
"Harold P.",
""
],
[
"Lutz",
"Jesse J.",
""
],
[
"Schultz",
"Peter A.",
""
]
] |
2504.15460 | Leonhard H\"olscher | Leonhard H\"olscher, Lukas M\"uller, Or Samimi, Tamuz Danzig | Quantum Simulation-Based Optimization of a Cooling System | null | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Engineering processes involve iterative design evaluations requiring numerous
computationally intensive numerical simulations. Quantum algorithms promise up
to exponential speedups for specific tasks relevant to numerical simulations.
However, these advantages quickly vanish when considering data input and output
on quantum computers. The recently introduced Quantum Simulation-Based
Optimization (QuSO) algorithm circumvents this fundamental impediment by
treating simulations as subproblems within a larger optimization problem. In
this paper, we adapt and implement QuSO for a cooling system design problem. We
validate the algorithm through statevector simulations and provide a detailed
algorithmic complexity analysis indicating that the achievable speedup for this
particular example is at most polynomial. Notably, we also specify the
conditions under which QuSO could deliver exponential advantages. By providing
this comprehensive demonstration, we highlight both the promise and practical
constraints of QuSO in an engineering context, motivating further research into
quantum algorithms for classical simulations and the search for suitable
applications.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 21:58:21 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Hölscher",
"Leonhard",
""
],
[
"Müller",
"Lukas",
""
],
[
"Samimi",
"Or",
""
],
[
"Danzig",
"Tamuz",
""
]
] |
2504.15461 | Alexei Skorobogatov | Tatiana Bandman, Boris Kunyavskii, and Alexei N. Skorobogatov | Birational properties of word varieties | 17 pages | null | null | null | math.AG | http://creativecommons.org/licenses/by/4.0/ | We prove that the subvariety of $SL(2)\times SL(2)$ given by the matrix
equation $w(X,Y)=\alpha$, where $w$ is a word in two letters, is closely
related to an explicit smooth conic bundle over the associated `trace surface'
in the 3-dimensional affine space. When $w$ is the commutator word, we show
that this variety can be irrational if the ground field $k$ is not
algebraically closed, answering a question of Rapinchuk, Benyash-Krivetz, and
Chernousov. When $k$ is a number field, it satisfies weak approximation with
the Brauer--Manin obstruction.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 21:58:25 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Bandman",
"Tatiana",
""
],
[
"Kunyavskii",
"Boris",
""
],
[
"Skorobogatov",
"Alexei N.",
""
]
] |
2504.15463 | Jae-Yoon Choi | Byungjin Lee and Kiryang Kwon and Jae-yoon Choi | Double-pass rotating z-cut quartz plate as a rapidly variable waveplate | 12 pages, 11 figures | null | null | null | physics.optics | http://creativecommons.org/licenses/by/4.0/ | We demonstrate a rapidly tunable waveplate based on a rotating z-cut quartz
plate in a double-pass configuration. In contrast to previous single-pass
implementations, where angular rotation of birefringent crystals causes
significant beam path displacement, we show that the double-pass geometry
effectively suppresses beam walk-off, reducing lateral shifts to below 10
$\mu$m, which is stable enough to have a fiber coupling. We present a full
theoretical description of the polarization changes using Jones matrix
calculations and verify it through polarization-resolved measurements.
Additionally, the retardation is stable across a broad spectral range without
requiring wavelength-specific optimization. When combined with a polarizing
beam splitter, the system functions as a high-speed optical power modulator,
achieving a dynamic power conversion in 1~ms with its contrast about 1000:1.
This compact and robust design is particularly suited for atomic, molecular,
and optical (AMO) physics experiments requiring rapid and precise control of
light intensity.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 21:59:40 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Lee",
"Byungjin",
""
],
[
"Kwon",
"Kiryang",
""
],
[
"Choi",
"Jae-yoon",
""
]
] |
2504.15465 | Dimitrios Skarlatos | Patrick H. Coppock, Brian Zhang, Eliot H. Solomon, Vasilis Kypriotis,
Leon Yang, Bikash Sharma, Dan Schatzberg, Todd C. Mowry, and Dimitrios
Skarlatos | LithOS: An Operating System for Efficient Machine Learning on GPUs | null | null | null | null | cs.OS cs.LG | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The surging demand for GPUs in datacenters for machine learning (ML) has made
efficient GPU utilization crucial. However, meeting the diverse needs of ML
models while optimizing resource usage is challenging. To enable transparent,
fine-grained GPU management that maximizes utilization and energy efficiency
while maintaining strong isolation, an operating system (OS) approach is
needed. This paper introduces LithOS, a first step toward a GPU OS. LithOS
includes the following new abstractions and mechanisms for efficient GPU
resource management: (i) a novel TPC Scheduler that supports spatial scheduling
at the granularity of individual TPCs, unlocking efficient TPC stealing between
workloads; (ii) transparent kernel atomization to reduce head-of-line blocking
and enable dynamic resource reallocation mid-execution; (iii) a lightweight
hardware right-sizing mechanism that determines the minimal TPC resources
needed per atom; and (iv) a transparent power management mechanism that reduces
power consumption based on in-flight work behavior. We implement LithOS in Rust
and evaluate its performance across extensive ML environments, comparing it to
state-of-the-art solutions from NVIDIA and prior research. For inference
stacking, LithOS reduces tail latencies by 13x compared to MPS; compared to the
best SotA, it reduces tail latencies by 3x while improving aggregate throughput
by 1.6x. In hybrid inference-training stacking, LithOS reduces tail latencies
by 4.7x compared to MPS; compared to the best SotA, it reduces tail latencies
1.18x while improving aggregate throughput by 1.35x. Finally, for a modest
performance hit under 4%, LithOS's right-sizing provides a quarter of GPU
capacity savings on average, while for a 7% hit, its power management yields a
quarter of a GPU's energy savings. Overall, LithOS increases GPU efficiency,
establishing a foundation for future OS research on GPUs.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:21:39 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Coppock",
"Patrick H.",
""
],
[
"Zhang",
"Brian",
""
],
[
"Solomon",
"Eliot H.",
""
],
[
"Kypriotis",
"Vasilis",
""
],
[
"Yang",
"Leon",
""
],
[
"Sharma",
"Bikash",
""
],
[
"Schatzberg",
"Dan",
""
],
[
"Mowry",
"Todd C.",
""
],
[
"Skarlatos",
"Dimitrios",
""
]
] |
2504.15466 | Jiayi Pan | Jiayi Pan, Xiuyu Li, Long Lian, Charlie Snell, Yifei Zhou, Adam Yala,
Trevor Darrell, Kurt Keutzer and Alane Suhr | Learning Adaptive Parallel Reasoning with Language Models | Code, model, and data are available at
https://github.com/Parallel-Reasoning/APR. The first three authors
contributed equally to this work | null | null | null | cs.AI cs.CL | http://creativecommons.org/licenses/by/4.0/ | Scaling inference-time computation has substantially improved the reasoning
capabilities of language models. However, existing methods have significant
limitations: serialized chain-of-thought approaches generate overly long
outputs, leading to increased latency and exhausted context windows, while
parallel methods such as self-consistency suffer from insufficient
coordination, resulting in redundant computations and limited performance
gains. To address these shortcomings, we propose Adaptive Parallel Reasoning
(APR), a novel reasoning framework that enables language models to orchestrate
both serialized and parallel computations end-to-end. APR generalizes existing
reasoning methods by enabling adaptive multi-threaded inference using spawn()
and join() operations. A key innovation is our end-to-end reinforcement
learning strategy, optimizing both parent and child inference threads to
enhance task success rate without requiring predefined reasoning structures.
Experiments on the Countdown reasoning task demonstrate significant benefits of
APR: (1) higher performance within the same context window (83.4% vs. 60.0% at
4k context); (2) superior scalability with increased computation (80.1% vs.
66.6% at 20k total tokens); (3) improved accuracy at equivalent latency (75.2%
vs. 57.3% at approximately 5,000ms). APR represents a step towards enabling
language models to autonomously optimize their reasoning processes through
adaptive allocation of computation.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:29:02 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Pan",
"Jiayi",
""
],
[
"Li",
"Xiuyu",
""
],
[
"Lian",
"Long",
""
],
[
"Snell",
"Charlie",
""
],
[
"Zhou",
"Yifei",
""
],
[
"Yala",
"Adam",
""
],
[
"Darrell",
"Trevor",
""
],
[
"Keutzer",
"Kurt",
""
],
[
"Suhr",
"Alane",
""
]
] |
2504.15467 | Hanbin Song | Hanbin Song, Xueyue Zhang, Lukasz Komza, Niccolo Fiaschi, Yihuang
Xiong, Yiyang Zhi, Scott Dhuey, Adam Schwartzberg, Thomas Schenkel, Geoffroy
Hautier, Zi-Huai Zhang, Alp Sipahigil | Long-lived entanglement of a spin-qubit register in silicon photonics | null | null | null | null | quant-ph cond-mat.mes-hall physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Color centers provide an optical interface to quantum registers based on
electron and nuclear spin qubits in solids. The T center in silicon is an
emerging spin-photon interface that combines telecom O-band optical transitions
and a long-lived electron spin in a scalable photonics platform. In this work,
we demonstrate the initialization, coherent control, and state readout of a
three-qubit register based on the electron spin of a T center coupled to a
hydrogen and a silicon nuclear spin. The spin register exhibits long spin echo
coherence times of $0.41(2)$ ms for the electron spin, $112(12)$ ms for the
hydrogen nuclear spin, and $67(7)$ ms for the silicon nuclear spin. We use
nuclear-nuclear two-qubit gates to generate entanglement between the two
nuclear spins with a fidelity of $F=0.77(3)$ and a coherence time of
$T^*_2=2.60(8)$ ms. Our results show that T centers can realize a long-lived
multi-qubit register with an optical interface in silicon photonics.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:29:18 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Song",
"Hanbin",
""
],
[
"Zhang",
"Xueyue",
""
],
[
"Komza",
"Lukasz",
""
],
[
"Fiaschi",
"Niccolo",
""
],
[
"Xiong",
"Yihuang",
""
],
[
"Zhi",
"Yiyang",
""
],
[
"Dhuey",
"Scott",
""
],
[
"Schwartzberg",
"Adam",
""
],
[
"Schenkel",
"Thomas",
""
],
[
"Hautier",
"Geoffroy",
""
],
[
"Zhang",
"Zi-Huai",
""
],
[
"Sipahigil",
"Alp",
""
]
] |
2504.15468 | Jack Lonergan | Jack Lonergan, Andrew Benson, Daniel Gilman | Generating Dark Matter Subhalo Populations Using Normalizing Flows | 13 pages, 8 figures | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | Strong gravitational lensing is a powerful tool for probing the nature of
dark matter, as lensing signals are sensitive to the dark matter substructure
within the lensing galaxy. We present a comparative analysis of strong
gravitational lensing signatures generated by dark matter subhalo populations
using two different approaches. The first approach models subhalos using an
empirical model, while the second employs the Galacticus semi-analytic model of
subhalo evolution. To date, only empirical approaches have been practical in
the analysis of lensing systems, as incorporating fully physical models was
computationally infeasible. To circumvent this, we utilize a generative machine
learning algorithm, known as a normalizing flow, to learn and reproduce the
subhalo populations generated by Galacticus. We demonstrate that the
normalizing flow algorithm accurately reproduces the Galacticus subhalo
distribution while significantly reducing computation time compared to direct
simulation. Moreover, we find that subhalo populations from Galacticus produce
comparable results to the empirical model in replicating observed lensing
signals under the fiducial dark matter model. This work highlights the
potential of machine learning techniques in accelerating astrophysical
simulations and improving model comparisons of dark matter properties.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:30:07 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Lonergan",
"Jack",
""
],
[
"Benson",
"Andrew",
""
],
[
"Gilman",
"Daniel",
""
]
] |
2504.15469 | Sina Fazelpour | Sina Fazelpour, Meica Magnani | Aspirational Affordances of AI | null | null | null | null | cs.CY | http://creativecommons.org/licenses/by/4.0/ | As artificial intelligence systems increasingly permeate processes of
cultural and epistemic production, there are growing concerns about how their
outputs may confine individuals and groups to static or restricted narratives
about who or what they could be. In this paper, we advance the discourse
surrounding these concerns by making three contributions. First, we introduce
the concept of aspirational affordance to describe how culturally shared
interpretive resources can shape individual cognition, and in particular
exercises practical imagination. We show how this concept can ground productive
evaluations of the risks of AI-enabled representations and narratives. Second,
we provide three reasons for scrutinizing of AI's influence on aspirational
affordances: AI's influence is potentially more potent, but less public than
traditional sources; AI's influence is not simply incremental, but ecological,
transforming the entire landscape of cultural and epistemic practices that
traditionally shaped aspirational affordances; and AI's influence is highly
concentrated, with a few corporate-controlled systems mediating a growing
portion of aspirational possibilities. Third, to advance such a scrutiny, we
introduce the concept of aspirational harm, which, in the context of AI
systems, arises when AI-enabled aspirational affordances distort or diminish
available interpretive resources in ways that undermine individuals' ability to
imagine relevant practical possibilities and alternative futures. Through three
case studies, we illustrate how aspirational harms extend the existing
discourse on AI-inflicted harms beyond representational and allocative harms,
warranting separate attention. Through these conceptual resources and analyses,
this paper advances understanding of the psychological and societal stakes of
AI's role in shaping individual and collective aspirations.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:37:49 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Fazelpour",
"Sina",
""
],
[
"Magnani",
"Meica",
""
]
] |
2504.15470 | Omer Hofman | Jonathan Brokman, Amit Giloni, Omer Hofman, Roman Vainshtein, Hisashi
Kojima and Guy Gilboa | Manifold Induced Biases for Zero-shot and Few-shot Detection of
Generated Images | Accepted to ICLR 2025 (The International Conference on Learning
Representations) | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Distinguishing between real and AI-generated images, commonly referred to as
'image detection', presents a timely and significant challenge. Despite
extensive research in the (semi-)supervised regime, zero-shot and few-shot
solutions have only recently emerged as promising alternatives. Their main
advantage is in alleviating the ongoing data maintenance, which quickly becomes
outdated due to advances in generative technologies. We identify two main gaps:
(1) a lack of theoretical grounding for the methods, and (2) significant room
for performance improvements in zero-shot and few-shot regimes. Our approach is
founded on understanding and quantifying the biases inherent in generated
content, where we use these quantities as criteria for characterizing generated
images. Specifically, we explore the biases of the implicit probability
manifold, captured by a pre-trained diffusion model. Through score-function
analysis, we approximate the curvature, gradient, and bias towards points on
the probability manifold, establishing criteria for detection in the zero-shot
regime. We further extend our contribution to the few-shot setting by employing
a mixture-of-experts methodology. Empirical results across 20 generative models
demonstrate that our method outperforms current approaches in both zero-shot
and few-shot settings. This work advances the theoretical understanding and
practical usage of generated content biases through the lens of manifold
analysis.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:39:24 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Brokman",
"Jonathan",
""
],
[
"Giloni",
"Amit",
""
],
[
"Hofman",
"Omer",
""
],
[
"Vainshtein",
"Roman",
""
],
[
"Kojima",
"Hisashi",
""
],
[
"Gilboa",
"Guy",
""
]
] |
2504.15472 | Pingcheng Jian | Pingcheng Jian, Xiao Wei, Yanbaihui Liu, Samuel A. Moore, Michael M.
Zavlanos, Boyuan Chen | LAPP: Large Language Model Feedback for Preference-Driven Reinforcement
Learning | null | null | null | null | cs.RO cs.LG cs.SY eess.SY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We introduce Large Language Model-Assisted Preference Prediction (LAPP), a
novel framework for robot learning that enables efficient, customizable, and
expressive behavior acquisition with minimum human effort. Unlike prior
approaches that rely heavily on reward engineering, human demonstrations,
motion capture, or expensive pairwise preference labels, LAPP leverages large
language models (LLMs) to automatically generate preference labels from raw
state-action trajectories collected during reinforcement learning (RL). These
labels are used to train an online preference predictor, which in turn guides
the policy optimization process toward satisfying high-level behavioral
specifications provided by humans. Our key technical contribution is the
integration of LLMs into the RL feedback loop through trajectory-level
preference prediction, enabling robots to acquire complex skills including
subtle control over gait patterns and rhythmic timing. We evaluate LAPP on a
diverse set of quadruped locomotion and dexterous manipulation tasks and show
that it achieves efficient learning, higher final performance, faster
adaptation, and precise control of high-level behaviors. Notably, LAPP enables
robots to master highly dynamic and expressive tasks such as quadruped
backflips, which remain out of reach for standard LLM-generated or handcrafted
rewards. Our results highlight LAPP as a promising direction for scalable
preference-driven robot learning.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:46:29 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Jian",
"Pingcheng",
""
],
[
"Wei",
"Xiao",
""
],
[
"Liu",
"Yanbaihui",
""
],
[
"Moore",
"Samuel A.",
""
],
[
"Zavlanos",
"Michael M.",
""
],
[
"Chen",
"Boyuan",
""
]
] |
2504.15473 | Berk Tinaz | Berk Tinaz, Zalan Fabian, Mahdi Soltanolkotabi | Emergence and Evolution of Interpretable Concepts in Diffusion Models | 32 pages, 32 figures, preliminary version | null | null | null | cs.CV cs.LG eess.IV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Diffusion models have become the go-to method for text-to-image generation,
producing high-quality images from noise through a process called reverse
diffusion. Understanding the dynamics of the reverse diffusion process is
crucial in steering the generation and achieving high sample quality. However,
the inner workings of diffusion models is still largely a mystery due to their
black-box nature and complex, multi-step generation process. Mechanistic
Interpretability (MI) techniques, such as Sparse Autoencoders (SAEs), aim at
uncovering the operating principles of models through granular analysis of
their internal representations. These MI techniques have been successful in
understanding and steering the behavior of large language models at scale.
However, the great potential of SAEs has not yet been applied toward gaining
insight into the intricate generative process of diffusion models. In this
work, we leverage the SAE framework to probe the inner workings of a popular
text-to-image diffusion model, and uncover a variety of human-interpretable
concepts in its activations. Interestingly, we find that even before the first
reverse diffusion step is completed, the final composition of the scene can be
predicted surprisingly well by looking at the spatial distribution of activated
concepts. Moreover, going beyond correlational analysis, we show that the
discovered concepts have a causal effect on the model output and can be
leveraged to steer the generative process. We design intervention techniques
aimed at manipulating image composition and style, and demonstrate that (1) in
early stages of diffusion image composition can be effectively controlled, (2)
in the middle stages of diffusion image composition is finalized, however
stylistic interventions are effective, and (3) in the final stages of diffusion
only minor textural details are subject to change.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:48:37 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Tinaz",
"Berk",
""
],
[
"Fabian",
"Zalan",
""
],
[
"Soltanolkotabi",
"Mahdi",
""
]
] |
2504.15474 | Sidong Feng | Sidong Feng, Changhao Du, Huaxiao Liu, Qingnan Wang, Zhengwei Lv, Gang
Huo, Xu Yang, Chunyang Chen | Agent for User: Testing Multi-User Interactive Features in TikTok | Accepted to ICSE 2025 Industry paper | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | TikTok, a widely-used social media app boasting over a billion monthly active
users, requires effective app quality assurance for its intricate features.
Feature testing is crucial in achieving this goal. However, the multi-user
interactive features within the app, such as live streaming, voice calls, etc.,
pose significant challenges for developers, who must handle simultaneous device
management and user interaction coordination. To address this, we introduce a
novel multi-agent approach, powered by the Large Language Models (LLMs), to
automate the testing of multi-user interactive app features. In detail, we
build a virtual device farm that allocates the necessary number of devices for
a given multi-user interactive task. For each device, we deploy an LLM-based
agent that simulates a user, thereby mimicking user interactions to
collaboratively automate the testing process. The evaluations on 24 multi-user
interactive tasks within the TikTok app, showcase its capability to cover 75%
of tasks with 85.9% action similarity and offer 87% time savings for
developers. Additionally, we have also integrated our approach into the
real-world TikTok testing platform, aiding in the detection of 26 multi-user
interactive bugs.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 22:50:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Feng",
"Sidong",
""
],
[
"Du",
"Changhao",
""
],
[
"Liu",
"Huaxiao",
""
],
[
"Wang",
"Qingnan",
""
],
[
"Lv",
"Zhengwei",
""
],
[
"Huo",
"Gang",
""
],
[
"Yang",
"Xu",
""
],
[
"Chen",
"Chunyang",
""
]
] |
2504.15475 | Szymon Kobus | Szymon Kobus and Deniz G\"und\"uz | Speculative Sampling via Exponential Races | null | null | null | null | cs.CL cs.IT math.IT | http://creativecommons.org/licenses/by/4.0/ | Speculative decoding accelerates large language model inference using a
smaller draft model. In this paper, we establish a surprising connection
between speculative decoding and channel simulation, which aims at simulating a
noisy channel using as few bits as possible. This connection allows us to
provide an information-theoretic analysis of the speed up that can be achieved
by speculative decoding. Leveraging this link, we derive an explicit relation
between generation speed-up and the number of tokens $k$ generated by the draft
model for large $k$, which serves as an upper bound for all $k$. We also
propose a novel speculative decoding method via exponential race ERSD that
matches state-of-the-art performance.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:02:08 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Kobus",
"Szymon",
""
],
[
"Gündüz",
"Deniz",
""
]
] |
2504.15476 | Rohan Surana | Rohan Surana, Junda Wu, Zhouhang Xie, Yu Xia, Harald Steck, Dawen
Liang, Nathan Kallus, Julian McAuley | From Reviews to Dialogues: Active Synthesis for Zero-Shot LLM-based
Conversational Recommender System | 11 pages, 2 figures | null | null | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | Conversational recommender systems (CRS) typically require extensive
domain-specific conversational datasets, yet high costs, privacy concerns, and
data-collection challenges severely limit their availability. Although Large
Language Models (LLMs) demonstrate strong zero-shot recommendation
capabilities, practical applications often favor smaller, internally managed
recommender models due to scalability, interpretability, and data privacy
constraints, especially in sensitive or rapidly evolving domains. However,
training these smaller models effectively still demands substantial
domain-specific conversational data, which remains challenging to obtain. To
address these limitations, we propose an active data augmentation framework
that synthesizes conversational training data by leveraging black-box LLMs
guided by active learning techniques. Specifically, our method utilizes
publicly available non-conversational domain data, including item metadata,
user reviews, and collaborative signals, as seed inputs. By employing active
learning strategies to select the most informative seed samples, our approach
efficiently guides LLMs to generate synthetic, semantically coherent
conversational interactions tailored explicitly to the target domain. Extensive
experiments validate that conversational data generated by our proposed
framework significantly improves the performance of LLM-based CRS models,
effectively addressing the challenges of building CRS in no- or low-resource
scenarios.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:05:47 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Surana",
"Rohan",
""
],
[
"Wu",
"Junda",
""
],
[
"Xie",
"Zhouhang",
""
],
[
"Xia",
"Yu",
""
],
[
"Steck",
"Harald",
""
],
[
"Liang",
"Dawen",
""
],
[
"Kallus",
"Nathan",
""
],
[
"McAuley",
"Julian",
""
]
] |
2504.15477 | Junda Wu | Junda Wu, Rohan Surana, Zhouhang Xie, Yiran Shen, Yu Xia, Tong Yu,
Ryan A. Rossi, Prithviraj Ammanabrolu, Julian McAuley | In-context Ranking Preference Optimization | 10 pages | null | null | null | cs.LG | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Recent developments in Direct Preference Optimization (DPO) allow large
language models (LLMs) to function as implicit ranking models by maximizing the
margin between preferred and non-preferred responses. In practice, user
feedback on such lists typically involves identifying a few relevant items in
context rather than providing detailed pairwise comparisons for every possible
item pair. Moreover, many complex information retrieval tasks, such as
conversational agents and summarization systems, critically depend on ranking
the highest-quality outputs at the top, emphasizing the need to support natural
and flexible forms of user feedback. To address the challenge of limited and
sparse pairwise feedback in the in-context setting, we propose an In-context
Ranking Preference Optimization (IRPO) framework that directly optimizes LLMs
based on ranking lists constructed during inference. To further capture
flexible forms of feedback, IRPO extends the DPO objective by incorporating
both the relevance of items and their positions in the list. Modeling these
aspects jointly is non-trivial, as ranking metrics are inherently discrete and
non-differentiable, making direct optimization difficult. To overcome this,
IRPO introduces a differentiable objective based on positional aggregation of
pairwise item preferences, enabling effective gradient-based optimization of
discrete ranking metrics. We further provide theoretical insights showing that
IRPO (i) automatically emphasizes items with greater disagreement between the
model and the reference ranking, and (ii) links its gradient to an importance
sampling estimator, yielding an unbiased estimator with reduced variance.
Empirical results show IRPO outperforms standard DPO approaches in ranking
performance, highlighting its effectiveness in aligning LLMs with direct
in-context ranking preferences.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:06:12 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wu",
"Junda",
""
],
[
"Surana",
"Rohan",
""
],
[
"Xie",
"Zhouhang",
""
],
[
"Shen",
"Yiran",
""
],
[
"Xia",
"Yu",
""
],
[
"Yu",
"Tong",
""
],
[
"Rossi",
"Ryan A.",
""
],
[
"Ammanabrolu",
"Prithviraj",
""
],
[
"McAuley",
"Julian",
""
]
] |
2504.15478 | Edoardo Provenzi | Edoardo Provenzi | Geometry of color perception. Part 1: Structures and metrics of a
homogeneous color space | null | The Journal of Mathematical Neuroscience, 10(1), 7 (2020) | null | null | q-bio.NC | http://creativecommons.org/licenses/by/4.0/ | This is the first half of a two-part paper dealing with the geometry of color
perception. Here we analyze in detail the seminal 1974 work by H.L. Resnikoff,
who showed that there are only two possible geometric structures and Riemannian
metrics on the perceived color space P compatible with the set of
Schr\"odinger's axioms completed with the hypothesis of homogeneity. We recast
Resnikoff's model into a more modern colorimetric setting, provide a much
simpler proof of the main result of the original paper and motivate the need of
psychophysical experiments to confute or confirm the linearity of background
transformations, which act transitively on P. Finally, we show that the
Riemannian metrics singled out by Resnikoff through an axiom on invariance
under background transformations are not compatibles with the crispening
effect, thus motivating the need of further research about perceptual color
metrics.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:06:23 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Provenzi",
"Edoardo",
""
]
] |
2504.15479 | Jeremy Goldwasser | Jeremy Goldwasser, Giles Hooker | Unifying Image Counterfactuals and Feature Attributions with
Latent-Space Adversarial Attacks | null | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | Counterfactuals are a popular framework for interpreting machine learning
predictions. These what if explanations are notoriously challenging to create
for computer vision models: standard gradient-based methods are prone to
produce adversarial examples, in which imperceptible modifications to image
pixels provoke large changes in predictions. We introduce a new,
easy-to-implement framework for counterfactual images that can flexibly adapt
to contemporary advances in generative modeling. Our method, Counterfactual
Attacks, resembles an adversarial attack on the representation of the image
along a low-dimensional manifold. In addition, given an auxiliary dataset of
image descriptors, we show how to accompany counterfactuals with feature
attribution that quantify the changes between the original and counterfactual
images. These importance scores can be aggregated into global counterfactual
explanations that highlight the overall features driving model predictions.
While this unification is possible for any counterfactual method, it has
particular computational efficiency for ours. We demonstrate the efficacy of
our approach with the MNIST and CelebA datasets.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:09:30 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Goldwasser",
"Jeremy",
""
],
[
"Hooker",
"Giles",
""
]
] |
2504.15480 | Antonin Brun | Antonin Brun, Gales Lucas, Bur\c{c}in Becerik-Gerber | Under Pressure: Contextualizing Workplace Stress Towards User-Centered
Interventions | CHI EA '25: Extended Abstracts of the CHI Conference on Human Factors
in Computing Systems | null | 10.1145/3706599.3719987 | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | Stress is a pervasive challenge that significantly impacts worker health and
well-being. Workplace stress is driven by various factors, ranging from
organizational changes to poor workplace design. Although individual stress
management strategies have been shown to be effective, current interventions
often overlook personal and contextual factors shaping stress experiences. In
this study, we conducted semi-structured interviews with eight office workers
to gain a deeper understanding of their personal experiences with workplace
stress. Our analysis reveals key stress triggers, coping mechanisms, and
reflections on past stressful events. We highlight the multifaceted and
individualized nature of workplace stress, emphasizing the importance of
intervention timing, modality, and recognizing that stress is not solely a
negative experience but can also have positive effects. Our findings provide
actionable insights for the design of user-centered stress management solutions
more attuned to the needs of office workers.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:22:37 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Brun",
"Antonin",
""
],
[
"Lucas",
"Gales",
""
],
[
"Becerik-Gerber",
"Burçin",
""
]
] |
2504.15481 | Edoardo Provenzi | Michel Berthier, Nicoletta Prencipe, Edoardo Provenzi | Split-quaternions for perceptual white balance | null | IEEE Signal Processing Magazine, 41(2), 42-50 (2024) | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | We propose a perceptual chromatic adaptation transform for white balance that
makes use of split-quaternions. The novelty of the present work, which is
motivated by a recently developed quantum-like model of color perception,
consists at stressing the link between the algebraic structures appearing in
this model and a certain sub-algebra of the split-quaternions. We show the
potentiality of this approach for color image processing applications by
proposing a chromatic adaptation transform, implemented via an appropriate use
of the split-quaternion multiplication. Moreover, quantitative comparisons with
the widely used state-of-the art von Kries chromatic adaptation transform are
provided.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:25:37 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Berthier",
"Michel",
""
],
[
"Prencipe",
"Nicoletta",
""
],
[
"Provenzi",
"Edoardo",
""
]
] |
2504.15482 | Yaqing Yang | Yaqing Yang, Vikram Mohanty, Nikolas Martelaro, Aniket Kittur,
Yan-Ying Chen, Matthew K. Hong | From Overload to Insight: Scaffolding Creative Ideation through
Structuring Inspiration | 2025 CHI workshop | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | Creative ideation relies on exploring diverse stimuli, but the overwhelming
abundance of information often makes it difficult to identify valuable insights
or reach the `aha' moment. Traditional methods for accessing design stimuli
lack organization and fail to support users in discovering promising
opportunities within large idea spaces. In this position paper, we explore how
AI can be leveraged to structure, organize, and surface relevant stimuli,
guiding users in both exploring idea spaces and mapping insights back to their
design challenges.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:29:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Yang",
"Yaqing",
""
],
[
"Mohanty",
"Vikram",
""
],
[
"Martelaro",
"Nikolas",
""
],
[
"Kittur",
"Aniket",
""
],
[
"Chen",
"Yan-Ying",
""
],
[
"Hong",
"Matthew K.",
""
]
] |
2504.15483 | Farnaz Fallahi | Farnaz Fallahi, Murat Yildirim, Shijia Zhao, Feng Qiu | A Sensor-Driven Optimization Framework for Asset Management in Energy
Systems: Implications for Full and Partial Digital Transformation in Hydro
Fleets | 35 pages, 8 figures | null | null | null | math.OC cs.SY eess.SY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper proposes a novel prognostics-driven approach to optimize
operations and maintenance (O&M) decisions in hydropower systems. Our approach
harnesses the insights from sensor data to accurately predict the remaining
lifetime distribution of critical generation assets in hydropower systems,
i.e., thrust bearings, and use these predictions to optimally schedule O&M
actions for a fleet of hydro generators. We consider complex interdependencies
across hydro generator failure risks, reservoir, production, and demand
management decisions. We propose a stochastic joint O&M scheduling model to
tackle the unique challenges of hydropower O&M including the interdependency of
generation capacities, the nonlinear nature of power production, operational
requirements, and uncertainties. We develop a two-level decomposition-based
solution algorithm to effectively handle large-scale cases. The algorithm
incorporates a combination of Benders optimality cuts and integer cuts to solve
the problem in an efficient manner. We design an experimental framework to
evaluate the proposed prognostics-driven O&M scheduling framework, using
real-world condition monitoring data from hydropower systems, historical market
prices, and water inflow data. The developed framework can be partially
implemented for a phased-in approach. Our experiments demonstrate the
significant benefits of the sensor-driven O&M framework in improving
reliability, availability, effective usage of resources, and system
profitability, especially when gradually shifting from traditional time-based
maintenance policies to condition-based prognostics-driven maintenance
policies.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:31:11 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Fallahi",
"Farnaz",
""
],
[
"Yildirim",
"Murat",
""
],
[
"Zhao",
"Shijia",
""
],
[
"Qiu",
"Feng",
""
]
] |
2504.15484 | Jeremy Lin J | Jeremy Lin, Tianchen Qian | Micro-randomized Trials with Categorical Treatments: Causal Effect
Estimation and Sample Size Calculation | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Micro-randomized trials (MRTs) are widely used to assess the marginal and
moderated effect of mobile health (mHealth) treatments delivered via mobile
devices. In many applications, the mHealth treatments are categorical with
multiple levels such as different types of message contents, but existing
analysis and sample size calculation methods for MRTs only focus on binary
treatment options (i.e., prompt vs. no prompt). We extended the causal
excursion effect definition and the weighted and centered least squares
estimator to MRTs with categorical treatments. Furthermore, we developed a
sample size formula for comparing categorical treatment levels, and proved the
type I error and power guarantee under working assumptions. We conducted
extensive simulations to assess type I error and power under assumption
violations, and we provided practical guidelines for using the sample size
formula to ensure adequate power in most real-world scenarios. We illustrated
the proposed estimator and sample size formula using the HeartSteps MRT.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:32:17 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Lin",
"Jeremy",
""
],
[
"Qian",
"Tianchen",
""
]
] |
2504.15485 | Jaemin Cho | Atin Pothiraj, Elias Stengel-Eskin, Jaemin Cho, Mohit Bansal | CAPTURe: Evaluating Spatial Reasoning in Vision Language Models via
Occluded Object Counting | Code and data: https://github.com/atinpothiraj/CAPTURe | null | null | null | cs.CV cs.AI cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recognizing and reasoning about occluded (partially or fully hidden) objects
is vital to understanding visual scenes, as occlusions frequently occur in
real-world environments and act as obstacles for spatial comprehension. To test
models' ability to reason about multiple occluded objects, we introduce a novel
task, Counting Amodally for Patterns Through Unseen REgions (CAPTURe), which
requires a model to count objects arranged in a pattern by inferring how the
pattern continues behind an occluder (an object which blocks parts of the
scene). CAPTURe requires both recognizing visual patterns and reasoning, making
it a useful testbed for evaluating vision-language models (VLMs) on whether
they understand occluded patterns and possess spatial understanding skills. By
requiring models to reason about occluded objects, CAPTURe also tests VLMs'
ability to form world models that would allow them to fill in missing
information. CAPTURe consists of two parts: (1) CAPTURe-real, with manually
filtered images of real objects in patterns and (2) CAPTURe-synthetic, a
controlled diagnostic with generated patterned images. We evaluate four strong
VLMs (GPT-4o, Intern-VL2, Molmo, and Qwen2-VL) on CAPTURe, finding that models
struggle to count on both occluded and unoccluded patterns. Crucially, we find
that models perform worse with occlusion, suggesting that VLMs are also
deficient in inferring unseen spatial relationships: even the strongest VLMs
like GPT-4o fail to count with occlusion. In contrast, we find that humans
achieve very little error on CAPTURe. We also find that providing auxiliary
information of occluded object locations increases performance, underscoring
that the model error comes both from an inability to handle occlusion as well
as difficulty counting in images.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:38:43 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Pothiraj",
"Atin",
""
],
[
"Stengel-Eskin",
"Elias",
""
],
[
"Cho",
"Jaemin",
""
],
[
"Bansal",
"Mohit",
""
]
] |
2504.15486 | Ho Sang (Leon) Chan | Ho-Sang Chan, Mitchell C. Begelman, Jason Dexter | Polarimetric Signatures of Bulk Comptonization from within the Plunging
Region of Accreting Black Holes | 15 pages, 14 figures, accepted for publication in APJ | null | null | null | astro-ph.HE gr-qc | http://creativecommons.org/licenses/by/4.0/ | Inverse Compton scattering by the thermal motions of electrons is believed to
produce polarized hard X-rays in active galactic nuclei and black-hole
binaries. Meanwhile, plasma within the plunging region of the black hole free
falls into the event horizon with a bulk relativistic speed, which could also
imprint polarization on up-scattered photons but has not been discussed in
detail. To examine this, we computed polarimetric signatures via general
relativistic ray-tracing of a toy model consisting of an accreting,
geometrically thin plasma with moderate optical depth, falling onto the black
hole with a bulk relativistic speed within the plunging region. We show that
the maximum spatially unresolved linear polarization could be as large as
approximately $7 - 8$ percent when the black hole is viewed near edge-on, while
the corresponding resolved linear polarization could be roughly $50$ percent.
The large discrepancy between the two is due to 1) dilution from the radiation
outside the plunging region and 2) substantial cancellations of the Stokes $Q$
and $U$ fluxes. The resultant polarization contributed by bulk Comptonization
could nevertheless exceed that of thermal electron scattering in a
Novikov-Thorne disk. Our results thus suggest a new model for imprinting
considerable polarization on the electromagnetic observables of accreting black
holes. Measurements of X-ray polarization from black-hole binaries and the
central black hole of active galactic nuclei could provide direct detection of
the plunging region and help constrain plasma properties in the immediate
vicinity of the event horizon.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:40:12 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Chan",
"Ho-Sang",
""
],
[
"Begelman",
"Mitchell C.",
""
],
[
"Dexter",
"Jason",
""
]
] |
2504.15487 | Ashesh Chattopadhyay | Moein Darman, Pedram Hassanzadeh, Laure Zanna, Ashesh Chattopadhyay | Fourier analysis of the physics of transfer learning for data-driven
subgrid-scale models of ocean turbulence | null | null | null | null | cs.LG nlin.CD physics.ao-ph physics.geo-ph | http://creativecommons.org/licenses/by/4.0/ | Transfer learning (TL) is a powerful tool for enhancing the performance of
neural networks (NNs) in applications such as weather and climate prediction
and turbulence modeling. TL enables models to generalize to out-of-distribution
data with minimal training data from the new system. In this study, we employ a
9-layer convolutional NN to predict the subgrid forcing in a two-layer ocean
quasi-geostrophic system and examine which metrics best describe its
performance and generalizability to unseen dynamical regimes. Fourier analysis
of the NN kernels reveals that they learn low-pass, Gabor, and high-pass
filters, regardless of whether the training data are isotropic or anisotropic.
By analyzing the activation spectra, we identify why NNs fail to generalize
without TL and how TL can overcome these limitations: the learned weights and
biases from one dataset underestimate the out-of-distribution sample spectra as
they pass through the network, leading to an underestimation of output spectra.
By re-training only one layer with data from the target system, this
underestimation is corrected, enabling the NN to produce predictions that match
the target spectra. These findings are broadly applicable to data-driven
parameterization of dynamical systems.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:42:19 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Darman",
"Moein",
""
],
[
"Hassanzadeh",
"Pedram",
""
],
[
"Zanna",
"Laure",
""
],
[
"Chattopadhyay",
"Ashesh",
""
]
] |
2504.15488 | Elisabeth Werner M | Carsten Schuett, Elisabeth M Werner, Diliya Yalikun | Floating bodies for ball-convex bodies | null | null | null | null | math.MG | http://creativecommons.org/licenses/by/4.0/ | We define floating bodies in the class of $n$-dimensional ball-convex bodies.
A right derivative of volume of these floating bodies leads to a surface area
measure for ball-convex bodies which we call relative affine surface area. We
show that this quantity is a rigid motion invariant, upper semi continuous
valuation.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:43:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Schuett",
"Carsten",
""
],
[
"Werner",
"Elisabeth M",
""
],
[
"Yalikun",
"Diliya",
""
]
] |
2504.15489 | Ho Sang (Leon) Chan | Ho-Sang Chan, Prasun Dhang, Jason Dexter, Mitchell C. Begelman | The Impact of Plasma Angular Momentum on Magnetically Arrested Flows and
Relativistic Jets in Hot Accretion Flows Around Black Holes | 15 pages, 14 figures, accpeted for publication in APJ | null | null | null | astro-ph.HE gr-qc | http://creativecommons.org/licenses/by/4.0/ | In certain scenarios, the accreted angular momentum of plasma onto a black
hole could be low; however, how the accretion dynamics depend on the angular
momentum content of the plasma is still not fully understood. We present
three-dimensional, general relativistic magnetohydrodynamic simulations of low
angular momentum accretion flows around rapidly spinning black holes (with spin
$a = +0.9$). The initial condition is a Fishbone-Moncrief (FM) torus threaded
by a large amount of poloidal magnetic flux, where the angular velocity is a
fraction $f$ of the standard value. For $f = 0$, the accretion flow becomes
magnetically arrested and launches relativistic jets but only for a very short
duration. After that, free-falling plasma breaks through the magnetic barrier,
loading the jet with mass and destroying the jet-disk structure. Meanwhile,
magnetic flux is lost via giant, asymmetrical magnetic bubbles that float away
from the black hole. The accretion then exits the magnetically arrested state.
For $f = 0.1$, the dimensionless magnetic flux threading the black hole
oscillates quasi-periodically. The jet-disk structure shows concurrent revival
and destruction while the gas outflow efficiency at the event horizon changes
accordingly. For $f \geq 0.3$, we find that the dynamical behavior of the
system starts to approach that of a standard accreting FM torus. Our results
thus suggest that the accreted angular momentum is an important parameter that
governs the maintenance of a magnetically arrested flow and launching of
relativistic jets around black holes.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:44:55 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Chan",
"Ho-Sang",
""
],
[
"Dhang",
"Prasun",
""
],
[
"Dexter",
"Jason",
""
],
[
"Begelman",
"Mitchell C.",
""
]
] |
2504.15490 | Bindiya Arora | Harpreet Kaur, Jyoti, Neelam Shukla, and Bindiya Arora | Two- and three-body dispersion coefficients for interaction of Cu and Ag
atoms with {Group} I, II, and XII elements | 8 pages, 3 tables, 1 figure | Chemical Physics Volume 596, 1 August 2025, 112731 | 10.1016/j.chemphys.2025.112731 | null | physics.atom-ph quant-ph | http://creativecommons.org/licenses/by/4.0/ | The mounting interest in conducting thorough analyses and studies of
long-range interactions stems from their wide-ranging applications in cold
atomic physics, making it a compelling area for research. In this work, we have
evaluated long range van der Waals dispersion (vdW) interactions of Cu and Ag
atoms with atoms of group I (Li, Na, K, Rb, Cs, and Fr), II (Be, Mg, Ca, Sr,
and {Ba}), XII (Zn, Cd, and Hg) {as well as} singly charged ions of group II
(Be$^+$, Mg$^+$, Ca$^+$, Sr$^+$, and {Ba$^+$}) and XII (Zn$^+$, Cd$^+$, and
Hg$^+$) by calculating $C_6$(two-body) and $C_9$ (three-body) vdW dispersion
coefficients. In order to obtain these $C_6$ and $C_9$ coefficients, we have
evaluated the dynamic dipole polarizability of the considered atoms using
appropriate relativistic methods and the sum-over-states approach. To ascertain
the accuracy of our results, we have compared the evaluated static dipole
polarizabilities of Cu and Ag atoms and their oscillator strengths for dominant
transitions {with} available literature. The calculated values of $C_6$
dispersion coefficients have also been compared with the previously reported
results.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:45:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Kaur",
"Harpreet",
""
],
[
"Jyoti",
"",
""
],
[
"Shukla",
"Neelam",
""
],
[
"Arora",
"Bindiya",
""
]
] |
2504.15491 | Tengda Tang | Tengda Tang, Jianhua Yao, Yixian Wang, Qiuwu Sha, Hanrui Feng, Zhen Xu | Application of Deep Generative Models for Anomaly Detection in Complex
Financial Transactions | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study proposes an algorithm for detecting suspicious behaviors in large
payment flows based on deep generative models. By combining Generative
Adversarial Networks (GAN) and Variational Autoencoders (VAE), the algorithm is
designed to detect abnormal behaviors in financial transactions. First, the GAN
is used to generate simulated data that approximates normal payment flows. The
discriminator identifies anomalous patterns in transactions, enabling the
detection of potential fraud and money laundering behaviors. Second, a VAE is
introduced to model the latent distribution of payment flows, ensuring that the
generated data more closely resembles real transaction features, thus improving
the model's detection accuracy. The method optimizes the generative
capabilities of both GAN and VAE, ensuring that the model can effectively
capture suspicious behaviors even in sparse data conditions. Experimental
results show that the proposed method significantly outperforms traditional
machine learning algorithms and other deep learning models across various
evaluation metrics, especially in detecting rare fraudulent behaviors.
Furthermore, this study provides a detailed comparison of performance in
recognizing different transaction patterns (such as normal, money laundering,
and fraud) in large payment flows, validating the advantages of generative
models in handling complex financial data.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:49:10 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Tang",
"Tengda",
""
],
[
"Yao",
"Jianhua",
""
],
[
"Wang",
"Yixian",
""
],
[
"Sha",
"Qiuwu",
""
],
[
"Feng",
"Hanrui",
""
],
[
"Xu",
"Zhen",
""
]
] |
2504.15492 | Lennart Linden Mr. | Lennart Linden and Karl A. Kalina and J\"org Brummund and Brain Riemer
and Markus K\"astner | A dual-stage constitutive modeling framework based on finite strain
data-driven identification and physics-augmented neural networks | null | null | null | null | cs.CE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this contribution, we present a novel consistent dual-stage approach for
the automated generation of hyperelastic constitutive models which only
requires experimentally measurable data. To generate input data for our
approach, an experiment with full-field measurement has to be conducted to
gather testing force and corresponding displacement field of the sample. Then,
in the first step of the dual-stage framework, a new finite strain Data-Driven
Identification (DDI) formulation is applied. This method enables to identify
tuples consisting of stresses and strains by only prescribing the applied
boundary conditions and the measured displacement field. In the second step,
the data set is used to calibrate a Physics-Augmented Neural Network (PANN),
which fulfills all common conditions of hyperelasticity by construction and is
very flexible at the same time. We demonstrate the applicability of our
approach by several descriptive examples. Two-dimensional synthetic data are
exemplarily generated in virtual experiments by using a reference constitutive
model. The calibrated PANN is then applied in 3D Finite Element simulations. In
addition, a real experiment including noisy data is mimicked.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:50:38 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Linden",
"Lennart",
""
],
[
"Kalina",
"Karl A.",
""
],
[
"Brummund",
"Jörg",
""
],
[
"Riemer",
"Brain",
""
],
[
"Kästner",
"Markus",
""
]
] |
2504.15493 | Bindiya Arora | Jyoti, A. Chakraborty, Zhiyang Wang, Jia Zhang, Jingbiao Chen, Bindiya
Arora, and B. K. Sahoo | Investigating the $4D_{3/2}|3,\pm2\rangle$--$4D_{5/2}|3,\pm2\rangle$
transition in Nb$^{4+}$ for a THz atomic clock | 11 pages, 4 Tables, 1 figure | Phys. Rev. A 111, 022813 (2025) | 10.1103/PhysRevA.111.022813 | null | physics.atom-ph quant-ph | http://creativecommons.org/licenses/by/4.0/ | In this work, the $4D_{3/2}|3,\pm2\rangle \rightarrow 4D_{5/2}|3,\pm2\rangle$
transition in the Nb$^{4+}$ ion is identified as a promising candidate for a
terahertz (THz) atomic clock, with the transition frequency occurring at
56.0224 THz. This transition is primarily driven by the magnetic dipole decay
channel, which can easily be accessed by a laser. We focus on the stable
$^{93}$Nb isotope, which has 100\% natural abundance and a nuclear spin of
$I=9/2$ for experimental advantage. Our data analysis allows us to estimate
potential systematic shifts in the proposed clock system, including those due
to blackbody radiation, electric quadrupole, second-order Zeeman, and
second-order Doppler {shifts}. {The scheme presented in this study can help
suppress the AC Stark and electric quadrupole shifts in the clock frequency
measurement.} {All these analyses} suggest that the proposed THz atomic clock
using Nb$^{4+}$ could be valuable in both quantum thermometry and frequency
metrology.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:50:52 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Jyoti",
"",
""
],
[
"Chakraborty",
"A.",
""
],
[
"Wang",
"Zhiyang",
""
],
[
"Zhang",
"Jia",
""
],
[
"Chen",
"Jingbiao",
""
],
[
"Arora",
"Bindiya",
""
],
[
"Sahoo",
"B. K.",
""
]
] |
2504.15494 | Jinghui Cheng | Jazlyn Hellman, Itai Epstein, Jinghui Cheng, Jin L.C. Guo | "Ohhh, He's the Boss!": Unpacking Power Dynamics Among Developers,
Designers, and End-Users in FLOSS Usability | 30 pages, 4 figures, Accepted to ACM CSCW 2025 | null | null | null | cs.HC cs.SE | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Addressing usability in free, libre, and open-source software (FLOSS) is a
challenging issue, particularly due to a long-existing "by developer, for
developer" mentality. Engaging designers and end-users to work with developers
can help improve its usability, but unequal power dynamics among those
stakeholder roles must be mitigated. To explore how the power of different
FLOSS stakeholders manifests and can be mediated during collaboration, we
conducted eight design workshops with different combinations of key FLOSS
stakeholders (i.e., developers, designers, and end-users). Leveraging existing
theories on Dimensions of Power, we revealed how participants navigate existing
role-based power structures through resource utilization, knowledge gap
management, and experience referencing. We also observed that participants
exhibited diverse behaviors confirming and challenging the status quo of FLOSS
usability. Overall, our results contribute to a comprehensive understanding of
the power dynamics among FLOSS stakeholders, providing valuable insights into
ways to balance their power to improve FLOSS usability. Our work also serves as
an exemplar of using design workshops as a research method to study power
dynamics during collaboration that are usually hidden in the field.
| [
{
"version": "v1",
"created": "Mon, 21 Apr 2025 23:52:03 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Hellman",
"Jazlyn",
""
],
[
"Epstein",
"Itai",
""
],
[
"Cheng",
"Jinghui",
""
],
[
"Guo",
"Jin L. C.",
""
]
] |
2504.15495 | David G. Grier | Mia C. Morrell, Leela Elliott and David G. Grier | Nonreciprocal wave-mediated interactions power a classical time crystal | 6 pages, 3 figures | null | null | null | cond-mat.soft | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | An acoustic standing wave acts as a lattice of evenly spaced potential energy
wells for sub-wavelength-scale objects. Trapped particles interact with each
other by exchanging waves that they scatter from the standing wave. Unless the
particles have identical scattering properties, their wave-mediated
interactions are nonreciprocal. Pairs of particles can use this nonreciprocity
to harvest energy from the wave to sustain steady-state oscillations despite
viscous drag and the absence of periodic driving. We show in theory and
experiment that a minimal system composed of two acoustically levitated
particles can access five distinct dynamical states, two of which are
emergently active steady states. Under some circumstances, these emergently
active steady states break spatiotemporal symmetry and therefore constitute a
classical time crystal.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:01:45 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Morrell",
"Mia C.",
""
],
[
"Elliott",
"Leela",
""
],
[
"Grier",
"David G.",
""
]
] |
2504.15496 | Ethan Phillip M. LaRochelle PhD | Eammon A. Littler, Emmanuel A. Mannoh, Ethan P. M. LaRochelle | Fluorescence Reference Target Quantitative Analysis Library | 12 pages, 1 table, 4 figures. Code available:
https://github.com/QUEL-Imaging/quel-qal), PyPi: quel-qal | null | null | null | physics.med-ph cs.CV eess.IV q-bio.QM | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Standardized performance evaluation of fluorescence imaging systems remains a
critical unmet need in the field of fluorescence-guided surgery (FGS). While
the American Association of Physicists in Medicine (AAPM) TG311 report and
recent FDA draft guidance provide recommended metrics for system
characterization, practical tools for extracting these metrics remain limited,
inconsistent, and often inaccessible. We present QUEL-QAL, an open-source
Python library designed to streamline and standardize the quantitative analysis
of fluorescence images using solid reference targets. The library provides a
modular, reproducible workflow that includes region of interest (ROI)
detection, statistical analysis, and visualization capabilities. QUEL-QAL
supports key metrics such as response linearity, limit of detection, depth
sensitivity, and spatial resolution, in alignment with regulatory and academic
guidance. Built on widely adopted Python packages, the library is designed to
be extensible, enabling users to adapt it to novel target designs and analysis
protocols. By promoting transparency, reproducibility, and regulatory
alignment, QUEL-QAL offers a foundational tool to support standardized
benchmarking and accelerate the development and evaluation of fluorescence
imaging systems.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:03:55 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Littler",
"Eammon A.",
""
],
[
"Mannoh",
"Emmanuel A.",
""
],
[
"LaRochelle",
"Ethan P. M.",
""
]
] |
2504.15497 | Noah Subedar | Noah Subedar, Taeui Kim, Saathwick Venkataramalingam | Scalable APT Malware Classification via Parallel Feature Extraction and
GPU-Accelerated Learning | 26 pages, 54 figures, 14 tables | null | null | null | cs.CR cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents an underlying framework for both automating and
accelerating malware classification, more specifically, mapping malicious
executables to known Advanced Persistent Threat (APT) groups. The main feature
of this analysis is the assembly-level instructions present in executables
which are also known as opcodes. The collection of such opcodes on many
malicious samples is a lengthy process; hence, open-source reverse engineering
tools are used in tandem with scripts that leverage parallel computing to
analyze multiple files at once. Traditional and deep learning models are
applied to create models capable of classifying malware samples. One-gram and
two-gram datasets are constructed and used to train models such as SVM, KNN,
and Decision Tree; however, they struggle to provide adequate results without
relying on metadata to support n-gram sequences. The computational limitations
of such models are overcome with convolutional neural networks (CNNs) and
heavily accelerated using graphical compute unit (GPU) resources.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:05:05 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Subedar",
"Noah",
""
],
[
"Kim",
"Taeui",
""
],
[
"Venkataramalingam",
"Saathwick",
""
]
] |
2504.15498 | T \`Em \'It \'Op \'E Gb\'ol\'ah\`an Jaiy\'eol\'a | K. G. Ilori, T. G. Jaiyeola, O.O. Oyebola, O. B. Ogunfolu and E. A.
Alhassan | Polygroupoid, Polyquasigroup, Polyloop and their Nuclei | 23 pages | Journal of the Nigerian Mathematical Society, Vol. 44, No. 1
(2025), 115 -137 | null | null | math.GR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this paper, new hyper-algebraic structures called polygroupoid,
polyquasigroup and polyloop were introduced with concrete examples given. The
first, second, third and fourth left (middle, right) nuclei of polygroupoid
were introduced and studied. It was shown that first left (middle, right)
nuclei of a polygroupoid is contained in second, third and fourth left (middle,
right) nuclei of the polygroupoid. Hence, the second, third and fourth left
(middle, right) nuclei of a polygroupoid generalize the first left (middle,
right) nuclei of the polygroupoid. Examples of polyquasigroups and polyloops
that satisfy the above results were provided.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:07:15 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ilori",
"K. G.",
""
],
[
"Jaiyeola",
"T. G.",
""
],
[
"Oyebola",
"O. O.",
""
],
[
"Ogunfolu",
"O. B.",
""
],
[
"Alhassan",
"E. A.",
""
]
] |
2504.15499 | James Mickens | James Mickens and Sarah Radway and Ravi Netravali | Guillotine: Hypervisors for Isolating Malicious AIs | To be published in the ACM SIGOPS 2025 Workshop on Hot Topics in
Operating Systems | null | null | null | cs.CR cs.AI cs.OS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As AI models become more embedded in critical sectors like finance,
healthcare, and the military, their inscrutable behavior poses ever-greater
risks to society. To mitigate this risk, we propose Guillotine, a hypervisor
architecture for sandboxing powerful AI models -- models that, by accident or
malice, can generate existential threats to humanity. Although Guillotine
borrows some well-known virtualization techniques, Guillotine must also
introduce fundamentally new isolation mechanisms to handle the unique threat
model posed by existential-risk AIs. For example, a rogue AI may try to
introspect upon hypervisor software or the underlying hardware substrate to
enable later subversion of that control plane; thus, a Guillotine hypervisor
requires careful co-design of the hypervisor software and the CPUs, RAM, NIC,
and storage devices that support the hypervisor software, to thwart side
channel leakage and more generally eliminate mechanisms for AI to exploit
reflection-based vulnerabilities. Beyond such isolation at the software,
network, and microarchitectural layers, a Guillotine hypervisor must also
provide physical fail-safes more commonly associated with nuclear power plants,
avionic platforms, and other types of mission critical systems. Physical
fail-safes, e.g., involving electromechanical disconnection of network cables,
or the flooding of a datacenter which holds a rogue AI, provide defense in
depth if software, network, and microarchitectural isolation is compromised and
a rogue AI must be temporarily shut down or permanently destroyed.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:29:18 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Mickens",
"James",
""
],
[
"Radway",
"Sarah",
""
],
[
"Netravali",
"Ravi",
""
]
] |
2504.15500 | Shengyong Pan | Peizheng Guo and Shengyong Pan | Relative derived equivalences and relative Igusa-Todorov dimensions | All comments are welcome | null | null | null | math.RT | http://creativecommons.org/licenses/by/4.0/ | Let $A$ be an Artin algebra and $F$ a non-zero subfunctor of
$\Ext_A^{1}(-,-)$. In this paper, we characterize the relative $\phi$-dimension
of $A$ by the bi-functor $\Ext_F^1(-,-)$. Furthermore, we show that the
finiteness of relative $\phi$-dimension of an Artin algebra is invariant under
relative derived equivalence. More precisely, for an Artin algebra $A$, assume
that $F$ has enough projectives and injectives, such that there exists $G\in
\modcat{A}$ such that $\add G=\mathcal {P}(F)$, where $\mathcal {P}(F)$ is the
category of all $F$-projecitve $A$-modules. If $\cpx{T}$ is a relative tilting
complex over $A$ with term length $t(\cpx{T})$ such that $B=\End(\cpx{T})$,
then we have $\phd_{F}(A)-t(T^{\bullet})\leq
\phd(B)\leq\phd_{F}(A)+t(T^{\bullet})+2$.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:30:45 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Guo",
"Peizheng",
""
],
[
"Pan",
"Shengyong",
""
]
] |
2504.15502 | Jose Daniel Castro-Cisneros | Jose Daniel Castro-Cisneros, Renu Malhotra and Aaron J. Rosengren | Lunar impact ejecta flux on the Earth | null | null | null | null | astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | The transfer of material between planetary bodies due to impact events is
important for understanding planetary evolution, meteoroid impact fluxes, the
formation of near-Earth objects (NEOs), and even the provenance of volatile and
organic materials at Earth. This study investigates the dynamics and fate of
lunar ejecta reaching Earth. We employ the high-accuracy IAS15 integrator
withing the REBOUND package to track for 100,000 years the trajectories of
6,000 test particles launched from various lunar latitudes and longitudes. Our
model incorporates a realistic velocity distribution for ejecta fragments (tens
of meters in size), derived form large lunar cratering events. Our results show
that 22.6% of lunar ejecta collide with Earth, following a power-law $C(t)
\propto t^{0.315}$ with half of the impacts occurring within ~10,000 years. We
also confirm that impact events on the Moon's trailing hemisphere serve as a
dominant source of Earth-bound ejecta, consistent with previous studies.
Additionally, a small fraction of ejecta remains transiently in near-Earth
space, providing evidence that lunar ejecta may contribute to the NEO
population. This aligns with recent discoveries of Earth co-orbitals such as
Kamo'oalewa (469219, 2016 HO3) and 2024 PT5, both exhibiting spectral
properties consistent with lunar material. These findings enhance our
understanding of the lunar ejecta flux to Earth, providing insights into the
spatial and temporal patterns of this flux and its broader influence on the
near-Earth environment.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:46:46 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Castro-Cisneros",
"Jose Daniel",
""
],
[
"Malhotra",
"Renu",
""
],
[
"Rosengren",
"Aaron J.",
""
]
] |
2504.15503 | Chen Yang | Chen Yang, M\'arcio A. Diniz, Deukwoo Kwon, Madhu Mazumdar | Intra-Class Correlation Coefficient Ignorable Clustered Randomized
Trials for Detecting Treatment Effect Heterogeneity | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Accurately estimating the intra-class correlation coefficient (ICC) is
crucial for adequately powering clustered randomized trials (CRTs). Challenges
arise due to limited prior data on the specific outcome within the target
population, making accurate ICC estimation difficult. Furthermore, ICC can vary
significantly across studies, even for the same outcome, influenced by factors
like study design, participant characteristics, and the specific intervention.
Power calculations are extremely sensitive to ICC assumptions. Minor variation
in the assumed ICC can lead to large differences in the number of clusters
needed, potentially impacting trial feasibility and cost.
This paper identifies a special class of CRTs aiming to detect the treatment
effect heterogeneity, wherein the ICC can be completely disregarded in
calculation of power and sample size. This result offers a solution for
research projects lacking preliminary estimates of the ICC or facing challenges
in their estimate. Moreover, this design facilitates power improvement through
increasing the cluster sizes rather than the number of clusters, making it
particular advantageous in the situations where expanding the number of
clusters is difficult or costly.
This paper provides a rigorous theoretical foundation for this class of
ICC-ignorable CRTs, including mathematical proofs and practical guidance for
implementation. We also present illustrative examples to demonstrate the
practical implications of this approach in various research contexts in
healthcare delivery.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:49:16 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Yang",
"Chen",
""
],
[
"Diniz",
"Márcio A.",
""
],
[
"Kwon",
"Deukwoo",
""
],
[
"Mazumdar",
"Madhu",
""
]
] |
2504.15504 | Shahan Ali Memon | Shahan Ali Memon, Jevin D. West, Cailin O'Connor | Are Widely Known Findings Easier to Retract? | 13 pages, 2 figures, Under review | null | null | null | cs.DL cs.SI | http://creativecommons.org/licenses/by/4.0/ | Failures of retraction are common in science. Why do these failures occur?
And, relatedly, what makes findings harder or easier to retract? We use data
from Microsoft Academic Graph, Retraction Watch, and Altmetric -- including
retracted papers, citation records, and Altmetric scores and mentions -- to
test recently proposed answers to these questions. A recent previous study by
LaCroix et al. employ simple network models to argue that the social spread of
scientific information helps explain failures of retraction. One prediction of
their models is that widely known or well established results, surprisingly,
should be easier to retract, since their retraction is more relevant to more
scientists. Our results support this conclusion. We find that highly cited
papers show more significant reductions in citation after retraction and garner
more attention to their retractions as they occur.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:51:22 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Memon",
"Shahan Ali",
""
],
[
"West",
"Jevin D.",
""
],
[
"O'Connor",
"Cailin",
""
]
] |
2504.15505 | Bogdan Nica | Bogdan Nica | Variance of point-counts for families of cubic curves over
$\mathbb{F}_p$ and Jacobsthal sums | 21 pages. Comments are welcome | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We give explicit computations for the variance of the number of points along
one-parameter families of cubic curves. We highlight evaluations of variances
that involve Jacobsthal sums.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:52:12 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Nica",
"Bogdan",
""
]
] |
2504.15508 | Jianxiong Li | Jianxiong Li, Beining Zhang, Mingzhen Li, Siyu Hu, Jinzhe Zeng, Lijun
Liu, Guojun Yuan, Zhan Wang, Guangming Tan, Weile Jia | Scaling Neural-Network-Based Molecular Dynamics with Long-Range
Electrostatic Interactions to 51 Nanoseconds per Day | 12 pages, 10 figures | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Neural network-based molecular dynamics (NNMD) simulations incorporating
long-range electrostatic interactions have significantly extended the
applicability to heterogeneous and ionic systems, enabling effective modeling
critical physical phenomena such as protein folding and dipolar surface and
maintaining ab initio accuracy. However, neural network inference and
long-range force computation remain the major bottlenecks, severely limiting
simulation speed. In this paper, we target DPLR, a state-of-the-art NNMD
package that supports long-range electrostatics, and propose a set of
comprehensive optimizations to enhance computational efficiency. We introduce
(1) a hardware-offloaded FFT method to reduce the communication overhead; (2)
an overlapping strategy that hides long-range force computations using a single
core per node, and (3) a ring-based load balancing method that enables
atom-level task evenly redistribution with minimal communication overhead.
Experimental results on the Fugaku supercomputer show that our work achieves a
37x performance improvement, reaching a maximum simulation speed of 51 ns/day.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 00:58:32 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Li",
"Jianxiong",
""
],
[
"Zhang",
"Beining",
""
],
[
"Li",
"Mingzhen",
""
],
[
"Hu",
"Siyu",
""
],
[
"Zeng",
"Jinzhe",
""
],
[
"Liu",
"Lijun",
""
],
[
"Yuan",
"Guojun",
""
],
[
"Wang",
"Zhan",
""
],
[
"Tan",
"Guangming",
""
],
[
"Jia",
"Weile",
""
]
] |
2504.15509 | Keqi Deng | Keqi Deng, Wenxi Chen, Xie Chen, Philip C. Woodland | SimulS2S-LLM: Unlocking Simultaneous Inference of Speech LLMs for
Speech-to-Speech Translation | null | null | null | null | cs.CL cs.SD eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Simultaneous speech translation (SST) outputs translations in parallel with
streaming speech input, balancing translation quality and latency. While large
language models (LLMs) have been extended to handle the speech modality,
streaming remains challenging as speech is prepended as a prompt for the entire
generation process. To unlock LLM streaming capability, this paper proposes
SimulS2S-LLM, which trains speech LLMs offline and employs a test-time policy
to guide simultaneous inference. SimulS2S-LLM alleviates the mismatch between
training and inference by extracting boundary-aware speech prompts that allows
it to be better matched with text input data. SimulS2S-LLM achieves
simultaneous speech-to-speech translation (Simul-S2ST) by predicting discrete
output speech tokens and then synthesising output speech using a pre-trained
vocoder. An incremental beam search is designed to expand the search space of
speech token prediction without increasing latency. Experiments on the CVSS
speech data show that SimulS2S-LLM offers a better translation quality-latency
trade-off than existing methods that use the same training data, such as
improving ASR-BLEU scores by 3 points at similar latency.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:05:32 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Deng",
"Keqi",
""
],
[
"Chen",
"Wenxi",
""
],
[
"Chen",
"Xie",
""
],
[
"Woodland",
"Philip C.",
""
]
] |
2504.15510 | Haoran Li | Haoran Li | Ridge-Regularized Largest Root Test for General High-dimensional Double
Wishart Problems | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | In multivariate analysis, many core problems involve the eigen-analysis of an
\(F\)-matrix, \(\bF = \bW_1\bW_2^{-1}\), constructed from two Wishart matrices,
\(\bW_1\) and \(\bW_2\). These so-called \textit{Double Wishart problems} arise
in contexts such as MANOVA, covariance matrix equality testing, and hypothesis
testing in multivariate linear regression. A prominent classical approach,
Roy's largest root test, relies on the largest eigenvalue of \(\bF\) for
inference. However, in high-dimensional settings, this test becomes impractical
due to the singularity or near-singularity of \(\bW_2\). To address this
challenge, we propose a ridge-regularization framework by introducing a ridge
term to \(\bW_2\). Specifically, we develop a family of ridge-regularized
largest root tests, leveraging the largest eigenvalue of \(\bF_\lambda =
\bW_1(\bW_2 + \lambda I)^{-1}\), where \(\lambda > 0\) is the regularization
parameter. Under mild assumptions, we establish the asymptotic Tracy-Widom
distribution of the largest eigenvalue of \(\bF_\lambda\) after appropriate
scaling. An efficient method for estimating the scaling parameters is proposed
using the Mar\v{c}enko-Pastur equation, and the consistency of these estimators
is proven. The proposed framework is applied to illustrative Double Wishart
problems, and simulation studies are conducted to evaluate the numerical
performance of the methods. Finally, the proposed method is applied to the
\emph{Human Connectome Project} data to test for the presence of associations
between volumetric measurements of human brain and behavioral variables.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:10:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Li",
"Haoran",
""
]
] |
2504.15511 | Isaac Dobes | Isaac Dobes and Naihuan Jing | Cayley's First Hyperdeterminant is an Entanglement Measure | 13 pages | null | null | null | quant-ph math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | Previously, it was shown that both the concurrence and n-tangle on
$2n$-qubits can be expressed in terms of Cayley's first hyperdeterminant,
indicating that Cayley's first hyperdeterminant captures at least some aspects
of a quantum state's entanglement. In this paper, we rigorously prove that
Cayley's first hyperdeterminant is an entanglement measure on $2n$-qudits, and
thus a legitimate generalization of the concurrence and n-tangle. In
particular, we show that the modulus of the hyperdeterminant and its square
both vanish on separable $2n$-qudits, are LU invariants, and are non-increasing
on average under LOCC. Lastly, we then consider their convex roof extensions
and show that they may in fact be viewed as entanglement measures on mixed
states of $2n$-qudits.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:10:36 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Dobes",
"Isaac",
""
],
[
"Jing",
"Naihuan",
""
]
] |
2504.15513 | Yixuan Zhu | Yixuan Zhu, Haolin Wang, Ao Li, Wenliang Zhao, Yansong Tang, Jingxuan
Niu, Lei Chen, Jie Zhou, Jiwen Lu | InstaRevive: One-Step Image Enhancement via Dynamic Score Matching | Accepted by ICLR 2025 | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Image enhancement finds wide-ranging applications in real-world scenarios due
to complex environments and the inherent limitations of imaging devices. Recent
diffusion-based methods yield promising outcomes but necessitate prolonged and
computationally intensive iterative sampling. In response, we propose
InstaRevive, a straightforward yet powerful image enhancement framework that
employs score-based diffusion distillation to harness potent generative
capability and minimize the sampling steps. To fully exploit the potential of
the pre-trained diffusion model, we devise a practical and effective diffusion
distillation pipeline using dynamic control to address inaccuracies in updating
direction during score matching. Our control strategy enables a dynamic
diffusing scope, facilitating precise learning of denoising trajectories within
the diffusion model and ensuring accurate distribution matching gradients
during training. Additionally, to enrich guidance for the generative power, we
incorporate textual prompts via image captioning as auxiliary conditions,
fostering further exploration of the diffusion model. Extensive experiments
substantiate the efficacy of our framework across a diverse array of
challenging tasks and datasets, unveiling the compelling efficacy and
efficiency of InstaRevive in delivering high-quality and visually appealing
results. Code is available at https://github.com/EternalEvan/InstaRevive.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:19:53 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhu",
"Yixuan",
""
],
[
"Wang",
"Haolin",
""
],
[
"Li",
"Ao",
""
],
[
"Zhao",
"Wenliang",
""
],
[
"Tang",
"Yansong",
""
],
[
"Niu",
"Jingxuan",
""
],
[
"Chen",
"Lei",
""
],
[
"Zhou",
"Jie",
""
],
[
"Lu",
"Jiwen",
""
]
] |
2504.15514 | David Nickel | David R. Nickel and Anindya Bijoy Das and David J. Love and
Christopher G. Brinton | Learning-Based Two-Way Communications: Algorithmic Framework and
Comparative Analysis | Currently under review for IEEE Communications Letters. 5 pages | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Machine learning (ML)-based feedback channel coding has garnered significant
research interest in the past few years. However, there has been limited
research exploring ML approaches in the so-called "two-way" setting where two
users jointly encode messages and feedback for each other over a shared
channel. In this work, we present a general architecture for ML-based two-way
feedback coding, and show how several popular one-way schemes can be converted
to the two-way setting through our framework. We compare such schemes against
their one-way counterparts, revealing error-rate benefits of ML-based two-way
coding in certain signal-to-noise ratio (SNR) regimes. We then analyze the
tradeoffs between error performance and computational overhead for three
state-of-the-art neural network coding models instantiated in the two-way
paradigm.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:22:13 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Nickel",
"David R.",
""
],
[
"Das",
"Anindya Bijoy",
""
],
[
"Love",
"David J.",
""
],
[
"Brinton",
"Christopher G.",
""
]
] |
2504.15516 | Jizu Huang | Junyuan He, Zhonghao Sun, and Jizu Huang | Derivatives of tree tensor networks and its applications in Runge--Kutta
methods | null | null | null | null | math.NA cs.NA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Tree tensor networks (TTNs) provide a compact and structured representation
of high-dimensional data, making them valuable in various areas of
computational mathematics and physics. In this paper, we present a rigorous
mathematical framework for expressing high-order derivatives of functional
TTNs, both with or without constraints. Our framework decomposes the total
derivative of a given TTN into a summation of TTNs, each corresponding to the
partial derivatives of the original TTN. Using this decomposition, we derive
the Taylor expansion of vector-valued functions subject to ordinary
differential equation constraints or algebraic constraints imposed by
Runge--Kutta (RK) methods. As a concrete application, we employ this framework
to construct order conditions for RK methods. Due to the intrinsic tensor
properties of partial derivatives and the separable tensor structure in RK
methods, the Taylor expansion of numerical solutions can be obtained in a
manner analogous to that of exact solutions using tensor operators. This
enables the order conditions of RK methods to be established by directly
comparing the Taylor expansions of the exact and numerical solutions,
eliminating the need for mathematical induction. For a given function
$\vector{f}$, we derive sharper order conditions that go beyond the classical
ones, enabling the identification of situations where a standard RK scheme of
order {\it p} achieves unexpectedly higher convergence order for the particular
function. These results establish new connections between tensor network theory
and classical numerical methods, potentially opening new avenues for both
analytical exploration and practical computation.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:26:51 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"He",
"Junyuan",
""
],
[
"Sun",
"Zhonghao",
""
],
[
"Huang",
"Jizu",
""
]
] |
2504.15517 | Mingchen Song | Mingchen Song, Xiang Deng, Guoqiang Zhong, Qi Lv, Jia Wan, Yinchuan
Li, Jianye Hao, Weili Guan | Few-Shot Vision-Language Action-Incremental Policy Learning | null | null | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, Transformer-based robotic manipulation methods utilize multi-view
spatial representations and language instructions to learn robot motion
trajectories by leveraging numerous robot demonstrations. However, the
collection of robot data is extremely challenging, and existing methods lack
the capability for continuous learning on new tasks with only a few
demonstrations. In this paper, we formulate these challenges as the Few-Shot
Action-Incremental Learning (FSAIL) task, and accordingly design a Task-prOmpt
graPh evolutIon poliCy (TOPIC) to address these issues. Specifically, to
address the data scarcity issue in robotic imitation learning, TOPIC learns
Task-Specific Prompts (TSP) through the deep interaction of multi-modal
information within few-shot demonstrations, thereby effectively extracting the
task-specific discriminative information. On the other hand, to enhance the
capability for continual learning on new tasks and mitigate the issue of
catastrophic forgetting, TOPIC adopts a Continuous Evolution Strategy (CES).
CES leverages the intrinsic relationships between tasks to construct a task
relation graph, which effectively facilitates the adaptation of new tasks by
reusing skills learned from previous tasks. TOPIC pioneers few-shot continual
learning in the robotic manipulation task, and extensive experimental results
demonstrate that TOPIC outperforms state-of-the-art baselines by over 26$\%$ in
success rate, significantly enhancing the continual learning capabilities of
existing Transformer-based policies.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:30:47 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Song",
"Mingchen",
""
],
[
"Deng",
"Xiang",
""
],
[
"Zhong",
"Guoqiang",
""
],
[
"Lv",
"Qi",
""
],
[
"Wan",
"Jia",
""
],
[
"Li",
"Yinchuan",
""
],
[
"Hao",
"Jianye",
""
],
[
"Guan",
"Weili",
""
]
] |
2504.15518 | Galen Dorpalen-Barry | Anastasia Chavez, Galen Dorpalen-Barry, Luis Ferroni, Fu Liu, Felipe
Rinc\'on, Andr\'es R. Vindas-Mel\'endez | The Ehrhart polynomial of a matroid specializes to the beta invariant | 7 pages | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | We show that the linear coefficient of the Ehrhart polynomial of a matroid
base polytope evaluated at $t-1$ is equal to, up to normalization, the
$\beta$-invariant of the matroid. This yields a lattice-point counting formula
for the $\beta$-invariant and establishes a new and unexpected positivity
property of Ehrhart polynomials of matroid polytopes.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:33:20 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Chavez",
"Anastasia",
""
],
[
"Dorpalen-Barry",
"Galen",
""
],
[
"Ferroni",
"Luis",
""
],
[
"Liu",
"Fu",
""
],
[
"Rincón",
"Felipe",
""
],
[
"Vindas-Meléndez",
"Andrés R.",
""
]
] |
2504.15519 | Luis F. Rodriguez | Luis F. Rodriguez, Laurent Loinard, Luis Zapata, Gisela N. Ortiz-Leon | Upper Limits to the Proper Motions of JuMBO 24, a Jupiter-Mass Binary
Object Candidate in Orion | 7 pages, 2 figures. To appear in ApJ | null | null | null | astro-ph.EP astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | Using JWST near-infrared data of the inner Orion Nebula,
\citet{Pearson_McCaughrean_2023} detected 40 binary systems they proposed to be
Jupiter-Mass Binary Objects (JuMBOs) -- although their actual nature is still
in debate. Only one of the objects, JuMBO\,24, was detected in the radio
continuum. Here, we report on new radio continuum (10 GHz) Karl G. Jansky Very
Large Array (VLA) detections of the radio counterpart to JuMBO\,24, and on an
unsuccessful search for 5 GHz continuum emission with the High Sensitivity
Array (HSA). From our new VLA detections and adopting a distance to the region,
we set an upper limit of $\simeq 6$~km~s$^{-1}$ to the velocity of the radio
source in the plane of the sky. This upper limit favors an origin for this
source similar to that of stars, that is, from a stationary contracting core.
The nature of the radio emission remains uncertain but the lack of strong
variability (all VLA observations are consistent with a steady flux of $\sim$50
$\mu$Jy), of detection on long HSA baseline, and of detectable circular
polarization in VLA data do not favor a non-thermal origin.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:34:18 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Rodriguez",
"Luis F.",
""
],
[
"Loinard",
"Laurent",
""
],
[
"Zapata",
"Luis",
""
],
[
"Ortiz-Leon",
"Gisela N.",
""
]
] |
2504.15520 | Shengsheng Zhang | Shengsheng Zhang, Taotao Ji, Meng Hua, Yongming Huang and Luxi Yang | Element-Grouping Strategy for Intelligent Reflecting Surface:
Performance Analysis and Algorithm Optimization | null | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As a revolutionary paradigm for intelligently controlling wireless channels,
intelligent reflecting surface (IRS) has emerged as a promising technology for
future sixth-generation (6G) wireless communications. While IRS-aided
communication systems can achieve attractive high channel gains, existing
schemes require plenty of IRS elements to mitigate the ``multiplicative
fading'' effect in cascaded channels, leading to high complexity for real-time
beamforming and high signaling overhead for channel estimation. In this paper,
the concept of sustainable intelligent element-grouping IRS (IEG-IRS) is
proposed to overcome those fundamental bottlenecks. Specifically, based on the
statistical channel state information (S-CSI), the proposed grouping strategy
intelligently pre-divide the IEG-IRS elements into multiple groups based on the
beam-domain grouping method, with each group sharing the common reflection
coefficient and being optimized in real time using the instantaneous channel
state information (I-CSI). Then, we further analyze the asymptotic performance
of the IEG-IRS to reveal the substantial capacity gain in an extremely
large-scale IRS (XL-IRS) aided single-user single-input single-output (SU-SISO)
system. In particular, when a line-of-sight (LoS) component exists, it
demonstrates that the combined cascaded link can be considered as a
``deterministic virtual LoS'' channel, resulting in a sustainable squared array
gain achieved by the IEG-IRS. Finally, we formulate a weighted-sum-rate (WSR)
maximization problem for an IEG-IRS-aided multiuser multiple-input
single-output (MU-MISO) system and a two-stage algorithm for optimizing the
beam-domain grouping strategy and the multi-user active-passive beamforming is
proposed.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:46:07 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhang",
"Shengsheng",
""
],
[
"Ji",
"Taotao",
""
],
[
"Hua",
"Meng",
""
],
[
"Huang",
"Yongming",
""
],
[
"Yang",
"Luxi",
""
]
] |
2504.15521 | Minghao Wu | Minghao Wu, Weixuan Wang, Sinuo Liu, Huifeng Yin, Xintong Wang, Yu
Zhao, Chenyang Lyu, Longyue Wang, Weihua Luo, Kaifu Zhang | The Bitter Lesson Learned from 2,000+ Multilingual Benchmarks | work in progress; 22 pages, 8 figures, 3 tables; | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As large language models (LLMs) continue to advance in linguistic
capabilities, robust multilingual evaluation has become essential for promoting
equitable technological progress. This position paper examines over 2,000
multilingual (non-English) benchmarks from 148 countries, published between
2021 and 2024, to evaluate past, present, and future practices in multilingual
benchmarking. Our findings reveal that, despite significant investments
amounting to tens of millions of dollars, English remains significantly
overrepresented in these benchmarks. Additionally, most benchmarks rely on
original language content rather than translations, with the majority sourced
from high-resource countries such as China, India, Germany, the UK, and the
USA. Furthermore, a comparison of benchmark performance with human judgments
highlights notable disparities. STEM-related tasks exhibit strong correlations
with human evaluations (0.70 to 0.85), while traditional NLP tasks like
question answering (e.g., XQuAD) show much weaker correlations (0.11 to 0.30).
Moreover, translating English benchmarks into other languages proves
insufficient, as localized benchmarks demonstrate significantly higher
alignment with local human judgments (0.68) than their translated counterparts
(0.47). This underscores the importance of creating culturally and
linguistically tailored benchmarks rather than relying solely on translations.
Through this comprehensive analysis, we highlight six key limitations in
current multilingual evaluation practices, propose the guiding principles
accordingly for effective multilingual benchmarking, and outline five critical
research directions to drive progress in the field. Finally, we call for a
global collaborative effort to develop human-aligned benchmarks that prioritize
real-world applications.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:47:37 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wu",
"Minghao",
""
],
[
"Wang",
"Weixuan",
""
],
[
"Liu",
"Sinuo",
""
],
[
"Yin",
"Huifeng",
""
],
[
"Wang",
"Xintong",
""
],
[
"Zhao",
"Yu",
""
],
[
"Lyu",
"Chenyang",
""
],
[
"Wang",
"Longyue",
""
],
[
"Luo",
"Weihua",
""
],
[
"Zhang",
"Kaifu",
""
]
] |
2504.15522 | Andrea Ceretani | Andrea Ceretani, Weiwei Hu, Lin Mu, and Carlos Rautenberg | A function space approach to the shape optimization of the Boussinesq
system | null | null | null | null | math.AP math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate a shape optimization problem for a heat-conducting fluid
governed by a Boussinesq system. The main goal is to determine an optimal
domain shape that yields a temperature distribution as uniform as possible.
Initially, we analyze the state problem, prove its well-posedness and establish
a local boundary regularity result for the weak solution. We then demonstrate
the existence of an optimal shape and derive a first-order optimality
condition. This requires the derivation and analysis of the adjoint system
associated with the Boussinesq model, as well as a rigorous treatment of the
directional derivatives of the objective functional under appropriate domain
perturbations. Finally, we present numerical experiments that illustrate and
support the theoretical findings.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 01:58:46 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ceretani",
"Andrea",
""
],
[
"Hu",
"Weiwei",
""
],
[
"Mu",
"Lin",
""
],
[
"Rautenberg",
"Carlos",
""
]
] |
2504.15523 | Nicolo Maria Della Ventura | Nicol\`o M. della Ventura, James D. Lamb, William C. Lenthe, McLean P.
Echlin, Julia T. P\"urstl, Emily S. Trageser, Alejandro M. Quevedo, Matthew
R. Begley, Tresa M. Pollock, Daniel S. Gianola and Marc De Graef | Orientation-Adaptive Virtual Imaging of Defects using EBSD | Affiliations: University of California Santa Barbara, Gatan + EDAX
Inc., Carnegie Mellon University | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by-nc-nd/4.0/ | EBSD is a foundational technique for characterizing crystallographic
orientation, phase distribution, and lattice strain. Embedded within EBSD
patterns lies latent information on dislocation structures, subtly encoded due
to their deviation from perfect crystallinity - a feature often underutilized.
Here, a novel framework termed orientation-adaptive virtual apertures (OAVA) is
introduced. OAVAs enable the generation of virtual images tied to specific
diffraction conditions, allowing the direct visualization of individual
dislocations from a single EBSD map. By dynamically aligning virtual apertures
in reciprocal space with the local crystallographic orientation, the method
enhances contrast from defect-related strain fields, mirroring the principles
of diffraction-contrast imaging in TEM, but without sample tilting. The
approach capitalizes on the extensive diffraction space captured in a single
high-quality EBSD scan, with its effectiveness enhanced by modern direct
electron detectors that offer high-sensitivity at low accelerating voltages,
reducing interaction volume and improving spatial resolution. We demonstrate
that using OAVAs, identical imaging conditions can be applied across a
polycrystalline field-of-view, enabling uniform contrast in differently
oriented grains. Furthermore, in single-crystal GaN, threading dislocations are
consistently resolved. Algorithms for the automated detection of dislocation
contrast are presented, advancing defect characterization. By using OAVAs
across a wide range of diffraction conditions in GaN, the
visibility/invisibility of defects, owing to the anisotropy of the elastic
strain field, is assessed and linked to candidate Burgers vectors. Altogether,
OAVA offers a new and high-throughput pathway for orientation-specific defect
characterization with the potential for automated, large-area defect analysis
in single and polycrystalline materials.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:00:03 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"della Ventura",
"Nicolò M.",
""
],
[
"Lamb",
"James D.",
""
],
[
"Lenthe",
"William C.",
""
],
[
"Echlin",
"McLean P.",
""
],
[
"Pürstl",
"Julia T.",
""
],
[
"Trageser",
"Emily S.",
""
],
[
"Quevedo",
"Alejandro M.",
""
],
[
"Begley",
"Matthew R.",
""
],
[
"Pollock",
"Tresa M.",
""
],
[
"Gianola",
"Daniel S.",
""
],
[
"De Graef",
"Marc",
""
]
] |
2504.15524 | Qiyao Wang | Qiyao Wang, Guhong Chen, Hongbo Wang, Huaren Liu, Minghui Zhu, Zhifei
Qin, Linwei Li, Yilin Yue, Shiqiang Wang, Jiayan Li, Yihang Wu, Ziqiang Liu,
Longze Chen, Run Luo, Liyang Fan, Jiaming Li, Lei Zhang, Kan Xu, Hongfei Lin,
Hamid Alinejad-Rokny, Shiwen Ni, Yuan Lin, Min Yang | IPBench: Benchmarking the Knowledge of Large Language Models in
Intellectual Property | 89 pages, 75 figures, 55 tables | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Intellectual Property (IP) is a unique domain that integrates technical and
legal knowledge, making it inherently complex and knowledge-intensive. As large
language models (LLMs) continue to advance, they show great potential for
processing IP tasks, enabling more efficient analysis, understanding, and
generation of IP-related content. However, existing datasets and benchmarks
either focus narrowly on patents or cover limited aspects of the IP field,
lacking alignment with real-world scenarios. To bridge this gap, we introduce
the first comprehensive IP task taxonomy and a large, diverse bilingual
benchmark, IPBench, covering 8 IP mechanisms and 20 tasks. This benchmark is
designed to evaluate LLMs in real-world intellectual property applications,
encompassing both understanding and generation. We benchmark 16 LLMs, ranging
from general-purpose to domain-specific models, and find that even the
best-performing model achieves only 75.8% accuracy, revealing substantial room
for improvement. Notably, open-source IP and law-oriented models lag behind
closed-source general-purpose models. We publicly release all data and code of
IPBench and will continue to update it with additional IP-related tasks to
better reflect real-world challenges in the intellectual property domain.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:00:41 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Wang",
"Qiyao",
""
],
[
"Chen",
"Guhong",
""
],
[
"Wang",
"Hongbo",
""
],
[
"Liu",
"Huaren",
""
],
[
"Zhu",
"Minghui",
""
],
[
"Qin",
"Zhifei",
""
],
[
"Li",
"Linwei",
""
],
[
"Yue",
"Yilin",
""
],
[
"Wang",
"Shiqiang",
""
],
[
"Li",
"Jiayan",
""
],
[
"Wu",
"Yihang",
""
],
[
"Liu",
"Ziqiang",
""
],
[
"Chen",
"Longze",
""
],
[
"Luo",
"Run",
""
],
[
"Fan",
"Liyang",
""
],
[
"Li",
"Jiaming",
""
],
[
"Zhang",
"Lei",
""
],
[
"Xu",
"Kan",
""
],
[
"Lin",
"Hongfei",
""
],
[
"Alinejad-Rokny",
"Hamid",
""
],
[
"Ni",
"Shiwen",
""
],
[
"Lin",
"Yuan",
""
],
[
"Yang",
"Min",
""
]
] |
2504.15525 | Yu Chengjun | Chengjun Yu, Yixin Ran, Yangyi Xia, Jia Wu, Xiaojing Liu | Federated Latent Factor Learning for Recovering Wireless Sensor Networks
Signal with Privacy-Preserving | Accepted By ICAIS&ISAS 2025 | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Wireless Sensor Networks (WSNs) are a cutting-edge domain in the field of
intelligent sensing. Due to sensor failures and energy-saving strategies, the
collected data often have massive missing data, hindering subsequent analysis
and decision-making. Although Latent Factor Learning (LFL) has been proven
effective in recovering missing data, it fails to sufficiently consider data
privacy protection. To address this issue, this paper innovatively proposes a
federated latent factor learning (FLFL) based spatial signal recovery (SSR)
model, named FLFL-SSR. Its main idea is two-fold: 1) it designs a sensor-level
federated learning framework, where each sensor uploads only gradient updates
instead of raw data to optimize the global model, and 2) it proposes a local
spatial sharing strategy, allowing sensors within the same spatial region to
share their latent feature vectors, capturing spatial correlations and
enhancing recovery accuracy. Experimental results on two real-world WSNs
datasets demonstrate that the proposed model outperforms existing federated
methods in terms of recovery performance.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:01:19 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Yu",
"Chengjun",
""
],
[
"Ran",
"Yixin",
""
],
[
"Xia",
"Yangyi",
""
],
[
"Wu",
"Jia",
""
],
[
"Liu",
"Xiaojing",
""
]
] |
2504.15526 | Nicolas Garcia | Nicolas Garcia, Ronnie Sircar, and H. Mete Soner | Mean Field Games of Control and Cryptocurrency Mining | 2 Figures | null | null | null | math.OC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper studies Mean Field Games (MFGs) in which agent dynamics are given
by jump processes of controlled intensity, with mean-field interaction via the
controls and affecting the jump intensities. We establish the existence of MFG
equilibria in a general discrete-time setting, and prove a limit theorem as the
time discretization goes to zero, establishing equilibria in the
continuous-time setting for a class of MFGs of intensity control. This
motivates numerical schemes that involve directly solving discrete-time games
as opposed to coupled Hamilton-Jacobi-Bellman and Kolmogorov equations. As an
example of the general theory, we consider cryptocurrency mining competition,
modeled as an MFG both in continuous and discrete time, and illustrate the
effectiveness of the discrete-time algorithm to solve it.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:05:29 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Garcia",
"Nicolas",
""
],
[
"Sircar",
"Ronnie",
""
],
[
"Soner",
"H. Mete",
""
]
] |
2504.15527 | Sophia Maria | Sophia Maria | Compass-V2 Technical Report | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Predominant LLMs focus on high-resource languages while leaving low-resource
languages, particularly those in Southeast Asia (SEA), underrepresented. In
addition, those models are general-purpose and pay limited attention to the
e-commerce domain. To overcome these limitations, we introduce Compass-v2, a
lightweight Mixture-of-Experts (MoE) model specifically designed for Southeast
Asian languages and e-commerce applications. To balance model performance and
inference cost, the model is designed with 30B total parameters and 5B active
parameters, incorporating both fine-grained and shared expert modules. To
enhance multilingual performance, we curated and constructed a high-quality,
industry-leading SEA dataset, to the best of our knowledge. To boost
performance in the e-commerce domain, we built a dataset comprising hundreds of
billions of tokens, sourced through external data mining and internal platform
collection. Besides, we pioneered a hybrid reasoning model that supports both
fast thinking and deep thinking within a unified framework to enhance the
reasoning capabilities, diverging from the conventional industry practice of
deploying two separate models. Through extensive experimental evaluations, our
model demonstrates state-of-the-art SEA multilingual and e-commerce performance
among sub-30B models, while maintaining significantly lower inference cost.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:08:45 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Maria",
"Sophia",
""
]
] |
2504.15528 | Yu-Bo Sheng | Jia-Wei Ying, Shi-Pu Gu, Xing-Fu Wang, Lan Zhou, Yu-Bo Sheng | Fully passive reference frame independent quantum key distribution | null | null | null | null | quant-ph | http://creativecommons.org/licenses/by/4.0/ | Reference-frame-independent quantum key distribution (RFI QKD) significantly
alleviates alignment requirements for reference frame in practical quantum
communication systems. While the original protocol requires Alice to prepare
six quantum states in $Z$, $X$, and $Y$ bases, its reliance on active
modulation introduces inherent side-channel vulnerabilities from device
imperfections. We address this security limitation by integrating a fully
passive source into the RFI framework. In this paper, we propose a fully
passive RFI QKD protocol. Our protocol avoids active modulation entirely,
suppressing side-channel risks through passive quantum state generation.
Moreover, by making full utilization of the quantum states generated by the
fully passive source, we enhance the secure key rate of fully passive protocol.
We establish a system model to analyze the performance of the protocol. Through
the optimization of post-selection intervals and intensities, we obtain the
maximum secure key transmission rate of the protocol. Under ideal
circumstances, the secure key transmission rate of our protocol can reach more
than 50% of that of the ideal QKD. Under practical conditions, we have
considered the finite-length effect. When pulse number generated by the source
reaches $10^{12}$, the maximum communication distances of the protocol can
reach 167 km and 136 km with a reference frame misalignment of $0 $ and
$45^{\circ} $ respectively. We believe that our protocol can contribute to the
development of practical QKD systems.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:08:47 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ying",
"Jia-Wei",
""
],
[
"Gu",
"Shi-Pu",
""
],
[
"Wang",
"Xing-Fu",
""
],
[
"Zhou",
"Lan",
""
],
[
"Sheng",
"Yu-Bo",
""
]
] |
2504.15531 | Osvaldo Mendez | Mohamed Khamsi, Jan Lang, Osvaldo Mendez | Modular topologies on vector spaces | null | null | null | null | math.FA math.AP math.GN | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper addresses the topological structures induced on vector spaces by
convex modulars that do not satisfy the $\Delta_2$ condition, with particular
focus on their applications to variable exponent spaces such as \( \ell^{(p_n)}
\) and \( L^{p(\cdot)} \). The motivation behind this investigation is its
applicability to the study of boundary value problems involving the variable
exponent $p(x)$-Laplacian when $p(x)$ is unbounded, a line of research recently
opened by the authors. Fundamental topological properties are analyzed,
including separation axioms, countability axioms, and the relationship between
modular convergence and classical topological concepts such as continuity.
Attention is given to the relation between modular and norm topologies. Special
emphasis is placed on the openness of modular balls, the impact of the
\(\Delta_2\)-condition, and duality with respect to modular topologies.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:21:51 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Khamsi",
"Mohamed",
""
],
[
"Lang",
"Jan",
""
],
[
"Mendez",
"Osvaldo",
""
]
] |
2504.15532 | Shayan Mortazavi | Shayan Mortazavi, Mehrnoosh Neshatian, Laurent Bozec, Hadis Zarrin,
Mahshid Kalani | Functional Amyloid Fibrils as Versatile Tools for Novel Biomaterials | null | null | null | null | q-bio.BM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Functional amyloid fibrils, once primarily associated with amyloidosis, are
now recognized for their exceptional potential as biomaterials due to their
unique structural features, including remarkable mechanical strength, high
stability, and self-assembly capabilities. This review highlights their
transformative applications across a wide range of industries, from
cutting-edge drug delivery systems and next-generation biosensors to tissue
engineering, surface technologies, energy storage, and environmental solutions.
Their versatility extends into innovative sectors like information transfer
systems, cell adhesion, protein fusion, food technology, and novel catalytic
systems. Despite significant progress, critical gaps remain in the research.
This review not only consolidates current applications but also underscores the
vast potential for future advancements, positioning functional amyloids as key
players in emerging biomaterials technologies.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:23:38 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Mortazavi",
"Shayan",
""
],
[
"Neshatian",
"Mehrnoosh",
""
],
[
"Bozec",
"Laurent",
""
],
[
"Zarrin",
"Hadis",
""
],
[
"Kalani",
"Mahshid",
""
]
] |
2504.15533 | Mamtaj Akter | Jinkyung Park, Mamtaj Akter, Naima Samreen Ali, Zainab Agha, Ashwaq
Alsoubai, Pamela Wisniewski | Towards Resilience and Autonomy-based Approaches for Adolescents Online
Safety | null | SOUPS Worksop position paper 2023 | null | null | cs.HC cs.CY | http://creativecommons.org/licenses/by/4.0/ | In this position paper, we discuss the paradigm shift that has emerged in the
literature, suggesting to move away from restrictive and authoritarian parental
mediation approaches to move toward resilient-based and privacy-preserving
solutions to promote adolescents' online safety. We highlight the limitations
of restrictive mediation strategies, which often induce a trade-off between
teens' privacy and online safety, and call for more teen-centric frameworks
that can empower teens to self-regulate while using the technology in
meaningful ways. We also present an overview of empirical studies that
conceptualized and examined resilience-based approaches to promoting the
digital well-being of teens in a way to empower teens to be more resilient.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:23:48 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Park",
"Jinkyung",
""
],
[
"Akter",
"Mamtaj",
""
],
[
"Ali",
"Naima Samreen",
""
],
[
"Agha",
"Zainab",
""
],
[
"Alsoubai",
"Ashwaq",
""
],
[
"Wisniewski",
"Pamela",
""
]
] |
2504.15534 | Jun He Prof. | Xiao-Xiao Chen, Zuo-Ming Ding, Jun He | Pole trajectories from $S$- and $P$-wave $D\bar{D}^*$ interactions | 8 pages, 5 figures | null | null | null | hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In this work, we investigate the $S$- and $P$-wave interactions of the
$D\bar{D}^*$ system within the framework of the quasipotential Bethe-Salpeter
equation, with the aim of exploring possible molecular states and their
corresponding pole trajectories. The interaction potentials are constructed
using the one-boson-exchange model, incorporating the exchanges of $\pi$,
$\eta$, $\rho$, $\omega$, $\sigma$, and $J/\psi$ mesons, based on heavy quark
effective Lagrangians. The poles of the scattering amplitude are analyzed and
their evolution on two Riemann sheets is systematically traced as the cutoff
parameter increases up to 5 GeV. We identify four molecular states arising from
the $S$-wave $D\bar{D}^*$ interaction. Among them, the bound state with quantum
numbers $I^G(J^{PC}) = 0^+(1^{++})$ corresponds well to the experimentally
observed $X(3872)$, while its isovector partner with $I^G(J^{PC}) =
1^-(1^{++})$ is found to exist only as a virtual state. Additionally, a
$0^-(1^{+-})$ state appears as a bound state. The isovector $1^{+-}$ state,
which may be associated with the $Z_c(3900)$, is observed to evolve from a
bound state to a virtual state as the interaction strength decreases. For the
$P$-wave $D\bar{D}^*$ interaction, the structure $G(3900)$ recently observed at
BESIII is likely connected to a $0^-(1^{--})$ state. A $0^+(0^{-+})$ state is
also predicted in this channel. Both can appear as either resonance or
bound/virtual state depending on the interaction strength.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:24:37 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Chen",
"Xiao-Xiao",
""
],
[
"Ding",
"Zuo-Ming",
""
],
[
"He",
"Jun",
""
]
] |
2504.15535 | Eric Chang | Kaidi Zhang, Do-Gon Kim, Eric T. Chang, Hua-Hsuan Liang, Zhanpeng He,
Kathryn Lampo, Philippe Wu, Ioannis Kymissis, Matei Ciocarlie | VibeCheck: Using Active Acoustic Tactile Sensing for Contact-Rich
Manipulation | 8 pages, 7 figures | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The acoustic response of an object can reveal a lot about its global state,
for example its material properties or the extrinsic contacts it is making with
the world. In this work, we build an active acoustic sensing gripper equipped
with two piezoelectric fingers: one for generating signals, the other for
receiving them. By sending an acoustic vibration from one finger to the other
through an object, we gain insight into an object's acoustic properties and
contact state. We use this system to classify objects, estimate grasping
position, estimate poses of internal structures, and classify the types of
extrinsic contacts an object is making with the environment. Using our contact
type classification model, we tackle a standard long-horizon manipulation
problem: peg insertion. We use a simple simulated transition model based on the
performance of our sensor to train an imitation learning policy that is robust
to imperfect predictions from the classifier. We finally demonstrate the policy
on a UR5 robot with active acoustic sensing as the only feedback.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:25:01 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhang",
"Kaidi",
""
],
[
"Kim",
"Do-Gon",
""
],
[
"Chang",
"Eric T.",
""
],
[
"Liang",
"Hua-Hsuan",
""
],
[
"He",
"Zhanpeng",
""
],
[
"Lampo",
"Kathryn",
""
],
[
"Wu",
"Philippe",
""
],
[
"Kymissis",
"Ioannis",
""
],
[
"Ciocarlie",
"Matei",
""
]
] |
2504.15536 | Maciej Matys | Maciej Matys, Atsushi Yamada, Toshihiro Ohki, and Kouji Tsunoda | Transition from positive to negative photoconductivity in AlGaN/GaN
quantum-well heterostructures | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | The AlGaN/GaN quantum-well heterostructures typically exhibit a positive
photoconductivity (PPC) during the light illumination. Surprisingly, we found
that introducing the GaN/AlN superlattice (SL) back barrier into N-polar
AlGaN/GaN quantum-well heterostructures induces a transition in these
heterostructures from PPC to negativie photoconductivity (NPC) as the SL period
number increased at room temperature. This transition occurred under an
infrared light illumination and can be well explained in terms of the
excitation of hot electrons from the two-dimensional electron gas and
subsequent trapping them in a SL structure. The NPC effect observed in N-polar
AlGaN/GaN heterostructures with SL back barrier exhibits photoconductivity
yield exceeding 85 % and thus is the largest ones reported so far for
semiconductors. In addition, NPC signal remains relatively stable at high
temperatures up to 400 K. The obtained results can be interesting for the
development of NPC related devices such as photoelectric logic gates,
photoelectronic memory and infrared photodetectors.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:25:46 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Matys",
"Maciej",
""
],
[
"Yamada",
"Atsushi",
""
],
[
"Ohki",
"Toshihiro",
""
],
[
"Tsunoda",
"Kouji",
""
]
] |
2504.15537 | Luze Xu | Grigoriy Blekherman, Jes\'us A. De Loera, Luze Xu, Shixuan Zhang | Semigroups of Integer Points in Convex Cones | null | null | null | null | math.NT math.OC | http://creativecommons.org/licenses/by/4.0/ | We study the question whether the affine semigroup of integer points in a
convex cone can be finitely generated up to symmetries of the cone. We
establish general properties of finite generation up to symmetry, and then
concentrate on the case of irrational polyhedral cones.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:27:17 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Blekherman",
"Grigoriy",
""
],
[
"De Loera",
"Jesús A.",
""
],
[
"Xu",
"Luze",
""
],
[
"Zhang",
"Shixuan",
""
]
] |
2504.15538 | James Beattie | James R. Beattie and Amitava Bhattacharjee | Scale-dependent alignment in compressible magnetohydrodynamic turbulence | 5 pages main text. 2 figures. 9 pages total. Submitted to PRL | null | null | null | physics.plasm-ph astro-ph.GA nlin.CD physics.flu-dyn | http://creativecommons.org/licenses/by/4.0/ | Using $10,\!080^3$ grid simulations, we analyze scale-dependent alignment in
driven, compressible, no net-flux magnetohydrodynamic turbulence. The plasma
self-organizes into localized, strongly aligned regions. Alignment spans all
primitive variables and their curls. Contrary to incompressible theory,
velocity-magnetic alignment scales as $\theta(\lambda) \sim \lambda^{1/8}$,
where $\lambda$ is the scale, suggesting a distinct three-dimensional eddy
anisotropy and a much higher critical transition scale toward a
reconnection-mediated cascade.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:30:28 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Beattie",
"James R.",
""
],
[
"Bhattacharjee",
"Amitava",
""
]
] |
2504.15539 | Ryan Miller | Ryan J. Miller, Alexander E. Dashuta, Brayden Rudisill, David Van
Vranken and Pierre Baldi | Interpretable Deep Learning for Polar Mechanistic Reaction Prediction | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurately predicting chemical reactions is essential for driving innovation
in synthetic chemistry, with broad applications in medicine, manufacturing, and
agriculture. At the same time, reaction prediction is a complex problem which
can be both time-consuming and resource-intensive for chemists to solve. Deep
learning methods offer an appealing solution by enabling high-throughput
reaction prediction. However, many existing models are trained on the US Patent
Office dataset and treat reactions as overall transformations: mapping
reactants directly to products with limited interpretability or mechanistic
insight. To address this, we introduce PMechRP (Polar Mechanistic Reaction
Predictor), a system that trains machine learning models on the PMechDB
dataset, which represents reactions as polar elementary steps that capture
electron flow and mechanistic detail. To further expand model coverage and
improve generalization, we augment PMechDB with a diverse set of
combinatorially generated reactions. We train and compare a range of machine
learning models, including transformer-based, graph-based, and two-step siamese
architectures. Our best-performing approach was a hybrid model, which combines
a 5-ensemble of Chemformer models with a two-step Siamese framework to leverage
the accuracy of transformer architectures, while filtering away "alchemical"
products using the two-step network predictions. For evaluation, we use a test
split of the PMechDB dataset and additionally curate a human benchmark dataset
consisting of complete mechanistic pathways extracted from an organic chemistry
textbook. Our hybrid model achieves a top-10 accuracy of 94.9% on the PMechDB
test set and a target recovery rate of 84.9% on the pathway dataset.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:31:23 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Miller",
"Ryan J.",
""
],
[
"Dashuta",
"Alexander E.",
""
],
[
"Rudisill",
"Brayden",
""
],
[
"Van Vranken",
"David",
""
],
[
"Baldi",
"Pierre",
""
]
] |
2504.15540 | Takayuki Ishizaki | Takayuki Ishizaki, Takahiro Kawaguchi, Yuichiro Yano, Yuko Hanado | Explicit Ensemble Mean Clock Synchronization for Optimal Atomic Time
Scale Generation | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | This paper presents a novel theoretical framework for atomic time scale
generation, called explicit ensemble mean synchronization, which unifies clock
synchronization and time scale generation within a control-theoretic paradigm.
By exploiting an observable canonical decomposition of a standard atomic clock
ensemble model, the system is decomposed into two complementary components: the
observable part, which represents the synchronization deviation, and the
unobservable part, which captures the synchronization destination. Within this
structure, we mathematically prove that standard Kalman filtering, widely used
in current time scale generation, can be interpreted as a special case of the
proposed framework that optimizes long-term frequency stability in terms of the
Allan variance. Furthermore, by applying appropriate state feedback control to
each component based on the Kalman filtering, both clock synchronization and
optimal time scale generation are achieved within a unified framework. This
framework provides a principled basis for robust timekeeping systems that goes
beyond conventional approaches in both scope and performance.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:33:23 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Ishizaki",
"Takayuki",
""
],
[
"Kawaguchi",
"Takahiro",
""
],
[
"Yano",
"Yuichiro",
""
],
[
"Hanado",
"Yuko",
""
]
] |
2504.15541 | Heye Huang | Qichao Liu, Heye Huang, Shiyue Zhao, Lei Shi, Soyoung Ahn, Xiaopeng Li | RiskNet: Interaction-Aware Risk Forecasting for Autonomous Driving in
Long-Tail Scenarios | 24 pages, 14 figures | null | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ensuring the safety of autonomous vehicles (AVs) in long-tail scenarios
remains a critical challenge, particularly under high uncertainty and complex
multi-agent interactions. To address this, we propose RiskNet, an
interaction-aware risk forecasting framework, which integrates deterministic
risk modeling with probabilistic behavior prediction for comprehensive risk
assessment. At its core, RiskNet employs a field-theoretic model that captures
interactions among ego vehicle, surrounding agents, and infrastructure via
interaction fields and force. This model supports multidimensional risk
evaluation across diverse scenarios (highways, intersections, and roundabouts),
and shows robustness under high-risk and long-tail settings. To capture the
behavioral uncertainty, we incorporate a graph neural network (GNN)-based
trajectory prediction module, which learns multi-modal future motion
distributions. Coupled with the deterministic risk field, it enables dynamic,
probabilistic risk inference across time, enabling proactive safety assessment
under uncertainty. Evaluations on the highD, inD, and rounD datasets, spanning
lane changes, turns, and complex merges, demonstrate that our method
significantly outperforms traditional approaches (e.g., TTC, THW, RSS, NC
Field) in terms of accuracy, responsiveness, and directional sensitivity, while
maintaining strong generalization across scenarios. This framework supports
real-time, scenario-adaptive risk forecasting and demonstrates strong
generalization across uncertain driving environments. It offers a unified
foundation for safety-critical decision-making in long-tail scenarios.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:36:54 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Liu",
"Qichao",
""
],
[
"Huang",
"Heye",
""
],
[
"Zhao",
"Shiyue",
""
],
[
"Shi",
"Lei",
""
],
[
"Ahn",
"Soyoung",
""
],
[
"Li",
"Xiaopeng",
""
]
] |
2504.15542 | Dongryul Kim | Dongryul Kim | Uniqueness and functoriality of Igusa stacks | null | null | null | null | math.NT math.AG | http://creativecommons.org/licenses/by/4.0/ | We provide an axiomatic definition of an Igusa stack associated to an
arbitrary Shimura datum. We then prove that Igusa stacks are unique and
automatically functorial with respect to morphisms of Shimura data, assuming
their existence. Using the same techniques, we also prove that the existence of
the Igusa stack passes to a Shimura subdatum.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:42:33 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Kim",
"Dongryul",
""
]
] |
2504.15543 | Joshua Chang | Joshua C. Chang and Edison Choe | Bayesian information theoretic model-averaging stochastic item selection
for computer adaptive testing: compromise-free item exposure | Under review | null | null | null | stat.ME cs.IT math.IT stat.ML | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The goal of Computer Adaptive Testing (CAT) is to reliably estimate an
individual's ability as modeled by an item response theory (IRT) instrument
using only a subset of the instrument's items. A secondary goal is to vary the
items presented across different testing sessions so that the sequence of items
does not become overly stereotypical -- we want all items to have an exposure
rate sufficiently far from zero. We formulate the optimization problem for CAT
in terms of Bayesian information theory, where one chooses the item at each
step based on the criterion of the ability model discrepancy -- the statistical
distance between the ability estimate at the next step and the full-test
ability estimate. This viewpoint of CAT naturally motivates a stochastic
selection procedure that equates choosing the next item to sampling from a
model-averaging ensemble ability model. Using the NIH Work Disability
Functional Assessment Battery (WD-FAB), we evaluate our new methods in
comparison to pre-existing methods found in the literature. We find that our
stochastic selector has superior properties in terms of both item exposure and
test accuracy/efficiency.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:45:16 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Chang",
"Joshua C.",
""
],
[
"Choe",
"Edison",
""
]
] |
2504.15544 | Issa Sugiura | Issa Sugiura, Kouta Nakayama, Yusuke Oda | llm-jp-modernbert: A ModernBERT Model Trained on a Large-Scale Japanese
Corpus with Long Context Length | 9 pages, 5 figures | null | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Encoder-only transformer models like BERT are widely adopted as a pre-trained
backbone for tasks like sentence classification and retrieval. However,
pretraining of encoder models with large-scale corpora and long contexts has
been relatively underexplored compared to decoder-only transformers. In this
work, we present llm-jp-modernbert, a ModernBERT model trained on a publicly
available, massive Japanese corpus with a context length of 8192 tokens. While
our model does not surpass existing baselines on downstream tasks, it achieves
good results on fill-mask test evaluations. We also analyze the effect of
context length expansion through pseudo-perplexity experiments. Furthermore, we
investigate sentence embeddings in detail, analyzing their transitions during
training and comparing them with those from other existing models, confirming
similar trends with models sharing the same architecture. To support
reproducibility and foster the development of long-context BERT, we release our
model, along with the training and evaluation code.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:45:19 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Sugiura",
"Issa",
""
],
[
"Nakayama",
"Kouta",
""
],
[
"Oda",
"Yusuke",
""
]
] |
2504.15545 | Zizhi Chen | Zizhi Chen, Xinyu Zhang, Minghao Han, Yizhou Liu, Ziyun Qian, Weifeng
Zhang, Xukun Zhang, Jingwei Wei, Lihua Zhang | VLM-based Prompts as the Optimal Assistant for Unpaired Histopathology
Virtual Staining | null | null | null | null | eess.IV cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In histopathology, tissue sections are typically stained using common H&E
staining or special stains (MAS, PAS, PASM, etc.) to clearly visualize specific
tissue structures. The rapid advancement of deep learning offers an effective
solution for generating virtually stained images, significantly reducing the
time and labor costs associated with traditional histochemical staining.
However, a new challenge arises in separating the fundamental visual
characteristics of tissue sections from the visual differences induced by
staining agents. Additionally, virtual staining often overlooks essential
pathological knowledge and the physical properties of staining, resulting in
only style-level transfer. To address these issues, we introduce, for the first
time in virtual staining tasks, a pathological vision-language large model
(VLM) as an auxiliary tool. We integrate contrastive learnable prompts,
foundational concept anchors for tissue sections, and staining-specific concept
anchors to leverage the extensive knowledge of the pathological VLM. This
approach is designed to describe, frame, and enhance the direction of virtual
staining. Furthermore, we have developed a data augmentation method based on
the constraints of the VLM. This method utilizes the VLM's powerful image
interpretation capabilities to further integrate image style and structural
information, proving beneficial in high-precision pathological diagnostics.
Extensive evaluations on publicly available multi-domain unpaired staining
datasets demonstrate that our method can generate highly realistic images and
enhance the accuracy of downstream tasks, such as glomerular detection and
segmentation. Our code is available at:
https://github.com/CZZZZZZZZZZZZZZZZZ/VPGAN-HARBOR
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:46:13 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Chen",
"Zizhi",
""
],
[
"Zhang",
"Xinyu",
""
],
[
"Han",
"Minghao",
""
],
[
"Liu",
"Yizhou",
""
],
[
"Qian",
"Ziyun",
""
],
[
"Zhang",
"Weifeng",
""
],
[
"Zhang",
"Xukun",
""
],
[
"Wei",
"Jingwei",
""
],
[
"Zhang",
"Lihua",
""
]
] |
2504.15547 | Yinchen Liu | Jian Li, Yinchen Liu, Yiran Zhang | Adaptivity Gaps for Stochastic Probing with Subadditive Functions | 49 pages, 7 figures | null | null | null | cs.DS | http://creativecommons.org/licenses/by-nc-sa/4.0/ | In this paper, we study the stochastic probing problem under a general
monotone norm objective. Given a ground set $U = [n]$, each element $i \in U$
has an independent nonnegative random variable $X_i$ with known distribution.
Probing an element reveals its value, and the sequence of probed elements must
satisfy a prefix-closed feasibility constraint $\mathcal{F}$. A monotone norm
$f: \mathbb{R}_{\geq 0}^n \to \mathbb{R}_{\geq 0}$ determines the reward
$f(X_P)$, where $P$ is the set of probed elements and $X_P$ is the vector with
$X_i$ for $i \in P$ and 0 otherwise. The goal is to design a probing strategy
maximizing the expected reward $\mathbb{E}[f(X_P)]$. We focus on the adaptivity
gap: the ratio between the expected rewards of optimal adaptive and optimal
non-adaptive strategies. We resolve an open question posed in [GNS17, KMS24],
showing that for general monotone norms, the adaptivity gap is $O(\log^2 n)$. A
refined analysis yields an improved bound of $O(\log r \log n / \log\log n)$,
where $r$ is the maximum size of a feasible probing sequence. As a by-product,
we derive an asymptotically tight adaptivity gap $\Theta(\log n / \log\log n)$
for Bernoulli probing with binary-XOS objectives, matching the known lower
bound. Additionally, we show an $O(\log^3 n)$ upper bound for Bernoulli probing
with general subadditive objectives. For monotone symmetric norms, we prove the
adaptivity gap is $O(1)$, improving the previous $O(\log n)$ bound from
[PRS23].
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:57:57 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Li",
"Jian",
""
],
[
"Liu",
"Yinchen",
""
],
[
"Zhang",
"Yiran",
""
]
] |
2504.15548 | Elyas Meguellati | Elyas Meguellati, Assaad Zeghina, Shazia Sadiq, Gianluca Demartini | LLM-based Semantic Augmentation for Harmful Content Detection | null | null | null | null | cs.CL cs.CY | http://creativecommons.org/licenses/by/4.0/ | Recent advances in large language models (LLMs) have demonstrated strong
performance on simple text classification tasks, frequently under zero-shot
settings. However, their efficacy declines when tackling complex social media
challenges such as propaganda detection, hateful meme classification, and
toxicity identification. Much of the existing work has focused on using LLMs to
generate synthetic training data, overlooking the potential of LLM-based text
preprocessing and semantic augmentation. In this paper, we introduce an
approach that prompts LLMs to clean noisy text and provide context-rich
explanations, thereby enhancing training sets without substantial increases in
data volume. We systematically evaluate on the SemEval 2024 multi-label
Persuasive Meme dataset and further validate on the Google Jigsaw toxic
comments and Facebook hateful memes datasets to assess generalizability. Our
results reveal that zero-shot LLM classification underperforms on these
high-context tasks compared to supervised models. In contrast, integrating
LLM-based semantic augmentation yields performance on par with approaches that
rely on human-annotated data, at a fraction of the cost. These findings
underscore the importance of strategically incorporating LLMs into machine
learning (ML) pipeline for social media classification tasks, offering broad
implications for combating harmful content online.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 02:59:03 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Meguellati",
"Elyas",
""
],
[
"Zeghina",
"Assaad",
""
],
[
"Sadiq",
"Shazia",
""
],
[
"Demartini",
"Gianluca",
""
]
] |
2504.15549 | Anjali Khurana | Anjali Khurana, Xiaotian Su, April Yi Wang, Parmit K Chilana | Do It For Me vs. Do It With Me: Investigating User Perceptions of
Different Paradigms of Automation in Copilots for Feature-Rich Software | Accepted for publication in the CHI Conference on Human Factors in
Computing Systems (CHI 2025), April 26 - May 1, 2025, Yokohama, Japan | null | 10.1145/3706598.3713431 | null | cs.HC cs.AI cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Model (LLM)-based in-application assistants, or copilots, can
automate software tasks, but users often prefer learning by doing, raising
questions about the optimal level of automation for an effective user
experience. We investigated two automation paradigms by designing and
implementing a fully automated copilot (AutoCopilot) and a semi-automated
copilot (GuidedCopilot) that automates trivial steps while offering
step-by-step visual guidance. In a user study (N=20) across data analysis and
visual design tasks, GuidedCopilot outperformed AutoCopilot in user control,
software utility, and learnability, especially for exploratory and creative
tasks, while AutoCopilot saved time for simpler visual tasks. A follow-up
design exploration (N=10) enhanced GuidedCopilot with task-and state-aware
features, including in-context preview clips and adaptive instructions. Our
findings highlight the critical role of user control and tailored guidance in
designing the next generation of copilots that enhance productivity, support
diverse skill levels, and foster deeper software engagement.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:11:10 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Khurana",
"Anjali",
""
],
[
"Su",
"Xiaotian",
""
],
[
"Wang",
"April Yi",
""
],
[
"Chilana",
"Parmit K",
""
]
] |
2504.15550 | Owen Conoly | Owen Conoly, Andres Erbsen, Adam Chlipala | Smooth, Integrated Proofs of Cryptographic Constant Time for
Nondeterministic Programs and Compilers | 34 pages, 1 table, 0 figures. to be published in PLDI 2025
proceedings | null | null | null | cs.PL | http://creativecommons.org/licenses/by/4.0/ | Formal verification of software and compilers has been used to rule out large
classes of security-critical issues, but risk of unintentional information
leakage has received much less consideration. It is a key requirement for
formal specifications to leave some details of a system's behavior unspecified
so that future implementation changes can be accommodated, and yet it is
nonetheless expected that these choices would not be made based on confidential
information the system handles. This paper formalizes that notion using
omnisemantics and plain single-copy assertions, giving for the first time a
specification of what it means for a nondeterministic program to be
constant-time or more generally to avoid leaking (a part of) its inputs. We use
this theory to prove data-leak-free execution of core cryptographic routines
compiled from Bedrock2 C to RISC-V machine code, showing that the smooth
specification and proof experience omnisemantics provides for nondeterminism
extends to constant-time properties in the same setting. We also study variants
of the key program-compiler contract, highlighting pitfalls of tempting
simplifications and subtle consequences of how inputs to nondeterministic
choices are constrained. Our results are backed by modular program-logic and
compiler-correctness theorems, and they integrate into a neat end-to-end
theorem in the Coq proof assistant.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:12:31 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Conoly",
"Owen",
""
],
[
"Erbsen",
"Andres",
""
],
[
"Chlipala",
"Adam",
""
]
] |
2504.15551 | Junrong Yan | Xianzhe Dai and Junrong Yan | Heat Kernel Approach to the Weyl Law for Schr\"odinger Operators on
Non-compact Manifolds | 20 pages. Comments welcome | null | null | null | math.DG math.AP math.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In a previous work by the authors, the heat kernel expansion for
Schr\"odinger-type operators on noncompact manifolds was studied. Motivated by
that work, this paper studies Weyl's law and the semiclassical Weyl's law for
Schr\"odinger operators of the form $\Delta + V$ and $\hbar^2\Delta + V$ on
noncompact complete manifolds with bounded geometry. We assume that the
potential $V(x)$ satisfies certain H\"older continuity conditions and tends to
$+\infty$ as $x$ approaches infinity.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:14:28 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Dai",
"Xianzhe",
""
],
[
"Yan",
"Junrong",
""
]
] |
2504.15552 | Fengyuan Li | Gengxian Cao, Fengyuan Li, Hong Duan, Ye Yang, Bofeng Wang, Donghe Li | A Multi-Agent Framework for Automated Qinqiang Opera Script Generation
Using Large Language Models | 17 pages,7 figures,1 tables | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | This paper introduces a novel multi-Agent framework that automates the end to
end production of Qinqiang opera by integrating Large Language Models , visual
generation, and Text to Speech synthesis. Three specialized agents collaborate
in sequence: Agent1 uses an LLM to craft coherent, culturally grounded
scripts;Agent2 employs visual generation models to render contextually accurate
stage scenes; and Agent3 leverages TTS to produce synchronized, emotionally
expressive vocal performances. In a case study on Dou E Yuan, the system
achieved expert ratings of 3.8 for script fidelity, 3.5 for visual coherence,
and 3.8 for speech accuracy-culminating in an overall score of 3.6, a 0.3 point
improvement over a Single Agent baseline. Ablation experiments demonstrate that
removing Agent2 or Agent3 leads to drops of 0.4 and 0.5 points, respectively,
underscoring the value of modular collaboration. This work showcases how AI
driven pipelines can streamline and scale the preservation of traditional
performing arts, and points toward future enhancements in cross modal
alignment, richer emotional nuance, and support for additional opera genres.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:14:29 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Cao",
"Gengxian",
""
],
[
"Li",
"Fengyuan",
""
],
[
"Duan",
"Hong",
""
],
[
"Yang",
"Ye",
""
],
[
"Wang",
"Bofeng",
""
],
[
"Li",
"Donghe",
""
]
] |
2504.15554 | Wei Zhao | Yueqiang Zhu, Chen Zhang, Ce Zhang, Lijing Zhong, Baiqiang Yang,
Jianrong Qiu, Kaige Wang, Jintao Bai, Wei Zhao | Partition laser assembling technique | null | null | null | null | physics.optics physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The advancement of micro/nanofabrication techniques with high throughput,
efficiency, and flexibility is critical for fields like integrated photonics,
biosensing, and medical diagnostics. This study presents Partition Laser
Assembling (PLA), a novel laser technique for fabricating complex
micro/nanostructures akin to puzzle pieces. By dividing the target patterns
described by scalable vector graphics into partitions, any structures in each
partition can be fabricated via structured lights as "light stamp" through
spatial light modulation. Unlike traditional direct laser writing, PLA
eliminates reliance on mechanical components, avoiding step-like artifacts and
ensuring smoother fabrication of complex micro/nanostructures. By seamlessly
assembling basic shapes, PLA achieves intricate structures like micro artworks
and metalenses with unmatched precision and resolution. Leveraging two-photon
fabrication, PLA guarantees high resolution and structural integrity,
positioning it as a transformative tool for nanoscale 3D printing. With
applications spanning research and industry, PLA paves the way for advanced
optical devices, micro/nanofabrications, and next-gen manufacturing
technologies.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:18:54 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhu",
"Yueqiang",
""
],
[
"Zhang",
"Chen",
""
],
[
"Zhang",
"Ce",
""
],
[
"Zhong",
"Lijing",
""
],
[
"Yang",
"Baiqiang",
""
],
[
"Qiu",
"Jianrong",
""
],
[
"Wang",
"Kaige",
""
],
[
"Bai",
"Jintao",
""
],
[
"Zhao",
"Wei",
""
]
] |
2504.15555 | Kun Zhang | Kun Zhang | Optimal Procurement Design: A Reduced Form Approach | null | null | null | null | econ.TH | http://creativecommons.org/licenses/by/4.0/ | Standard procurement models assume that the buyer knows the quality of the
good at the time of procurement; however, in many settings, the quality is
learned only long after the transaction. We study procurement problems in which
the buyer's valuation of the supplied good depends directly on its quality,
which is unverifiable and unobservable to the buyer. For a broad class of
procurement problems, we identify procurement mechanisms maximizing any
weighted average of the buyer's expected payoff and social surplus. The optimal
mechanism can be implemented by an auction that restricts sellers to submit
bids within specific intervals.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:19:02 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Zhang",
"Kun",
""
]
] |
2504.15556 | Yandi Shen | Zhou Fan, Justin Ko, Bruno Loureiro, Yue M. Lu, Yandi Shen | Dynamical mean-field analysis of adaptive Langevin diffusions:
Propagation-of-chaos and convergence of the linear response | null | null | null | null | math.ST math.PR stat.TH | http://creativecommons.org/publicdomain/zero/1.0/ | Motivated by an application to empirical Bayes learning in high-dimensional
regression, we study a class of Langevin diffusions in a system with random
disorder, where the drift coefficient is driven by a parameter that
continuously adapts to the empirical distribution of the realized process up to
the current time. The resulting dynamics take the form of a stochastic
interacting particle system having both a McKean-Vlasov type interaction and a
pairwise interaction defined by the random disorder. We prove a
propagation-of-chaos result, showing that in the large system limit over
dimension-independent time horizons, the empirical distribution of sample paths
of the Langevin process converges to a deterministic limit law that is
described by dynamical mean-field theory. This law is characterized by a system
of dynamical fixed-point equations for the limit of the drift parameter and for
the correlation and response kernels of the limiting dynamics. Using a
dynamical cavity argument, we verify that these correlation and response
kernels arise as the asymptotic limits of the averaged correlation and linear
response functions of single coordinates of the system. These results enable an
asymptotic analysis of an empirical Bayes Langevin dynamics procedure for
learning an unknown prior parameter in a linear regression model, which we
develop in a companion paper.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:20:58 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Fan",
"Zhou",
""
],
[
"Ko",
"Justin",
""
],
[
"Loureiro",
"Bruno",
""
],
[
"Lu",
"Yue M.",
""
],
[
"Shen",
"Yandi",
""
]
] |
2504.15557 | Zhangting Chen | Zhangting Chen, Bingjie Wang, Dongya Cheng | Asymptotics of higher-order conditional tail moments for
convolution-equivalently distributed losses | null | null | null | null | math.PR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper investigates the asymptotic behavior of higher-order conditional
tail moments, which quantify the contribution of individual losses in the event
of systemic collapse. The study is conducted within a framework comprising two
investment portfolios experiencing dependent losses that follow
convolution-equivalent distributions. The main results are encapsulated in two
theorems: one addressing light-tailed losses with convolution-equivalent
distributions and the other focusing on heavy-tailed losses with regularly
varying distributions. Both results reveal that the asymptotic behavior remains
robust regardless of the strength of dependence. Additionally, numerical
simulations are performed under specific scenarios to validate the theoretical
results.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:24:07 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Chen",
"Zhangting",
""
],
[
"Wang",
"Bingjie",
""
],
[
"Cheng",
"Dongya",
""
]
] |
2504.15558 | Yandi Shen | Zhou Fan, Justin Ko, Bruno Loureiro, Yue M. Lu, Yandi Shen | Dynamical mean-field analysis of adaptive Langevin diffusions:
Replica-symmetric fixed point and empirical Bayes | null | null | null | null | math.ST stat.TH | http://creativecommons.org/publicdomain/zero/1.0/ | In many applications of statistical estimation via sampling, one may wish to
sample from a high-dimensional target distribution that is adaptively evolving
to the samples already seen. We study an example of such dynamics, given by a
Langevin diffusion for posterior sampling in a Bayesian linear regression model
with i.i.d. regression design, whose prior continuously adapts to the Langevin
trajectory via a maximum marginal-likelihood scheme. Results of dynamical
mean-field theory (DMFT) developed in our companion paper establish a precise
high-dimensional asymptotic limit for the joint evolution of the prior
parameter and law of the Langevin sample. In this work, we carry out an
analysis of the equations that describe this DMFT limit, under conditions of
approximate time-translation-invariance which include, in particular, settings
where the posterior law satisfies a log-Sobolev inequality. In such settings,
we show that this adaptive Langevin trajectory converges on a
dimension-independent time horizon to an equilibrium state that is
characterized by a system of scalar fixed-point equations, and the associated
prior parameter converges to a critical point of a replica-symmetric limit for
the model free energy. As a by-product of our analyses, we obtain a new
dynamical proof that this replica-symmetric limit for the free energy is exact,
in models having a possibly misspecified prior and where a log-Sobolev
inequality holds for the posterior law.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:24:09 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Fan",
"Zhou",
""
],
[
"Ko",
"Justin",
""
],
[
"Loureiro",
"Bruno",
""
],
[
"Lu",
"Yue M.",
""
],
[
"Shen",
"Yandi",
""
]
] |
2504.15559 | ZengXing Liu | Zeng-Xing Liu, Yan-Hua Wu, Jing-Hua Sun | Dispersive-induced magnon blockade with a superconducting qubit | null | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We investigate the magnon blockade effect in a quantum magnonic system
operating in the strong dispersive regime, where a superconducting qubit
interacts dispersively with a magnonic mode in a yttrium-iron-garnet sphere.By
solving the quantum master equation, we demonstrate that the magnon blockade,
characterized by the second-order correlation function $g^{(2)}(0) \rightarrow
0.04$, emerges under optimal dispersive coupling and driving detuning.The
mechanism is attributed to suppressed two-magnon transitions as a result of
qubit-induced anharmonicity. Notably, our study identifies the critical role of
dispersive interaction strength and environmental temperature, showing that
magnon blockade remains observable under experimentally achievable cryogenic
conditions.This work extends the magnon blockade effect into the dispersive
regime, offering a robust platform for single-magnon manipulation and advancing
applications in quantum sensing and information processing.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:28:50 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Liu",
"Zeng-Xing",
""
],
[
"Wu",
"Yan-Hua",
""
],
[
"Sun",
"Jing-Hua",
""
]
] |
2504.15560 | T. K. Sridharan | T. K. Sridharan, R. Lehmensiek, S. Schwarz and D. P. Marrone | Antenna Technology Readiness for the Black Hole Explorer (BHEX) Mission | 5 pages, 3 figures, Proc. Int. Conf. Electromagn. Adv. Appl. (ICEAA),
Palermo, Italy, Sep. 2025, submitted | null | null | null | astro-ph.IM | http://creativecommons.org/publicdomain/zero/1.0/ | The Black Hole Explorer (BHEX) will be the first sub-mm wavelength Space
Very-Long-Baseline Interferometry (VLBI) mission. It targets astronomical
imaging with the highest ever spatial resolution to enable detection of the
photon ring of a supermassive black hole. BHEX is being proposed for launch in
2031 as a NASA Small Explorers mission. BHEX science goals and mission
opportunity require a high precision lightweight spaceborne antenna. A survey
of the technology landscape for realizing such an antenna is presented.
Technology readiness (TRL) for the antenna is discussed and assessed to be at
TRL 5. An update on our technology maturation efforts is provided. Design
studies leading to the conceptual design of a metallized carbon fiber
reinforced plastic (CFRP) technology based antenna with a mass of only $\dim
50$ kg, incorporating a 3.4 m primary reflector with a surface precision of <
40 $\mu$m to allow efficient operation up to 320 GHz are outlined. Current
plans anticipate attaining TRL6 in 2026 for the BHEX antenna. Completed design
studies point to a large margin in surface precision which opens up
opportunities for applications beyond BHEX, at significantly higher THz
frequencies.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:30:32 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Sridharan",
"T. K.",
""
],
[
"Lehmensiek",
"R.",
""
],
[
"Schwarz",
"S.",
""
],
[
"Marrone",
"D. P.",
""
]
] |
2504.15561 | Jingkai Xu | Jingkai Xu, Xiangli Nie | SPECI: Skill Prompts based Hierarchical Continual Imitation Learning for
Robot Manipulation | null | null | null | null | cs.RO cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Real-world robot manipulation in dynamic unstructured environments requires
lifelong adaptability to evolving objects, scenes and tasks. Traditional
imitation learning relies on static training paradigms, which are ill-suited
for lifelong adaptation. Although Continual Imitation Learnin (CIL) enables
incremental task adaptation while preserving learned knowledge, current CIL
methods primarily overlook the intrinsic skill characteristics of robot
manipulation or depend on manually defined and rigid skills, leading to
suboptimal cross-task knowledge transfer. To address these issues, we propose
Skill Prompts-based HiErarchical Continual Imitation Learning (SPECI), a novel
end-to-end hierarchical CIL policy architecture for robot manipulation. The
SPECI framework consists of a multimodal perception and fusion module for
heterogeneous sensory information encoding, a high-level skill inference module
for dynamic skill extraction and selection, and a low-level action execution
module for precise action generation. To enable efficient knowledge transfer on
both skill and task levels, SPECI performs continual implicit skill acquisition
and reuse via an expandable skill codebook and an attention-driven skill
selection mechanism. Furthermore, we introduce mode approximation to augment
the last two modules with task-specific and task-sharing parameters, thereby
enhancing task-level knowledge transfer. Extensive experiments on diverse
manipulation task suites demonstrate that SPECI consistently outperforms
state-of-the-art CIL methods across all evaluated metrics, revealing
exceptional bidirectional knowledge transfer and superior overall performance.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:30:38 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Xu",
"Jingkai",
""
],
[
"Nie",
"Xiangli",
""
]
] |
2504.15562 | Dip Roy | Dip Roy | Bayesian Autoencoder for Medical Anomaly Detection: Uncertainty-Aware
Approach for Brain 2 MRI Analysis | 16 pages, 6 figures | null | null | null | cs.LG cs.CV | http://creativecommons.org/licenses/by/4.0/ | In medical imaging, anomaly detection is a vital element of healthcare
diagnostics, especially for neurological conditions which can be
life-threatening. Conventional deterministic methods often fall short when it
comes to capturing the inherent uncertainty of anomaly detection tasks. This
paper introduces a Bayesian Variational Autoencoder (VAE) equipped with
multi-head attention mechanisms for detecting anomalies in brain magnetic
resonance imaging (MRI). For the purpose of improving anomaly detection
performance, we incorporate both epistemic and aleatoric uncertainty estimation
through Bayesian inference. The model was tested on the BraTS2020 dataset, and
the findings were a 0.83 ROC AUC and a 0.83 PR AUC. The data in our paper
suggests that modeling uncertainty is an essential component of anomaly
detection, enhancing both performance and interpretability and providing
confidence estimates, as well as anomaly predictions, for clinicians to
leverage in making medical decisions.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:30:42 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Roy",
"Dip",
""
]
] |
2504.15563 | Er-Jia Guo | Yiyan Fan, Qinghua Zhang, Jingdi Lu, Chuanrui Huo, Tianyang Wang, Qiao
Jin, Ting Cui, Qianying Wang, Dongke Rong, Shiqing Deng, Lingfei Wang,
Kuijuan Jin, Jun Chen, and Er-Jia Guo | Designing Optimal Distorted-Octahedra Superlattices for Strong
Topological Hall Effect | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Topologically protected spin states hold great promise for applications in
next generation of memory circuits and spintronic devices. These intriguing
textures typically emerge in bulk materials or heterostructures with broken
inversion symmetry, accompanied by an enhanced Dzyaloshinskii-Moriya
interaction (DMI). In this study, we successfully induced the topological Hall
effect (THE) in atomically designed (DyScO3)n/(SrRuO3)n (DnSn) superlattices
over a significant range of temperatures (10~120K) and thicknesses (16~40nm).
Using magnetic force microscopy (MFM), we observed the formation and stability
of magnetic domains, such as topological skyrmions. By precisely controlling
the interlayer thickness (n) and biaxial strain, we elucidated the mechanisms
underlying the modulation and induction of magnetic topological states.
Supporting evidence was provided by scanning transmission electron microscopy
(STEM) and X-ray absorption spectroscopy (XAS), thereby lending further
credence to our conclusions. These heterostructures offer a universal method
for exploring topological phenomena driven by distorted octahedra, while
enhancing the integrability and addressability of topologically protected
functional devices.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:32:05 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Fan",
"Yiyan",
""
],
[
"Zhang",
"Qinghua",
""
],
[
"Lu",
"Jingdi",
""
],
[
"Huo",
"Chuanrui",
""
],
[
"Wang",
"Tianyang",
""
],
[
"Jin",
"Qiao",
""
],
[
"Cui",
"Ting",
""
],
[
"Wang",
"Qianying",
""
],
[
"Rong",
"Dongke",
""
],
[
"Deng",
"Shiqing",
""
],
[
"Wang",
"Lingfei",
""
],
[
"Jin",
"Kuijuan",
""
],
[
"Chen",
"Jun",
""
],
[
"Guo",
"Er-Jia",
""
]
] |
2504.15564 | Musfiqur Rahman | Musfiqur Rahman, SayedHassan Khatoonabadi, Emad Shihab | A Large-scale Class-level Benchmark Dataset for Code Generation with
LLMs | This paper was submitted to the 29th International Conference on
Evaluation and Assessment in Software Engineering (EASE 2025) AI models/data
track | null | null | null | cs.SE cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent advancements in large language models (LLMs) have demonstrated
promising capabilities in code generation tasks. However, most existing
benchmarks focus on isolated functions and fail to capture the complexity of
real-world, class-level software structures. To address this gap, we introduce
a large-scale, Python class-level dataset curated from $13{,}174$ real-world
open-source projects. The dataset contains over 842,000 class skeletons, each
including class and method signatures, along with associated docstrings when
available. We preserve structural and contextual dependencies critical to
realistic software development scenarios and enrich the dataset with static
code metrics to support downstream analysis. To evaluate the usefulness of this
dataset, we use extracted class skeletons as prompts for GPT-4 to generate full
class implementations. Results show that the LLM-generated classes exhibit
strong lexical and structural similarity to human-written counterparts, with
average ROUGE@L, BLEU, and TSED scores of 0.80, 0.59, and 0.73, respectively.
These findings confirm that well-structured prompts derived from real-world
class skeletons significantly enhance LLM performance in class-level code
generation. This dataset offers a valuable resource for benchmarking, training,
and improving LLMs in realistic software engineering contexts.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:33:57 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Rahman",
"Musfiqur",
""
],
[
"Khatoonabadi",
"SayedHassan",
""
],
[
"Shihab",
"Emad",
""
]
] |
2504.15565 | Zheyuan Gu | Zheyuan Gu, Chang Liu, Xiyuan Zhang, Chen Yang, Gaopeng Gou, Gang
Xiong, Zhen Li, Sijia Li | DecETT: Accurate App Fingerprinting Under Encrypted Tunnels via Dual
Decouple-based Semantic Enhancement | Accepted to WWW 2025 | null | 10.1145/3696410.3714643 | null | cs.CR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Due to the growing demand for privacy protection, encrypted tunnels have
become increasingly popular among mobile app users, which brings new challenges
to app fingerprinting (AF)-based network management. Existing methods primarily
transfer traditional AF methods to encrypted tunnels directly, ignoring the
core obfuscation and re-encapsulation mechanism of encrypted tunnels, thus
resulting in unsatisfactory performance. In this paper, we propose DecETT, a
dual decouple-based semantic enhancement method for accurate AF under encrypted
tunnels. Specifically, DecETT improves AF under encrypted tunnels from two
perspectives: app-specific feature enhancement and irrelevant tunnel feature
decoupling.Considering the obfuscated app-specific information in encrypted
tunnel traffic, DecETT introduces TLS traffic with stronger app-specific
information as a semantic anchor to guide and enhance the fingerprint
generation for tunnel traffic. Furthermore, to address the app-irrelevant
tunnel feature introduced by the re-encapsulation mechanism, DecETT is designed
with a dual decouple-based fingerprint enhancement module, which decouples the
tunnel feature and app semantic feature from tunnel traffic separately, thereby
minimizing the impact of tunnel features on accurate app fingerprint
extraction. Evaluation under five prevalent encrypted tunnels indicates that
DecETT outperforms state-of-the-art methods in accurate AF under encrypted
tunnels, and further demonstrates its superiority under tunnels with more
complicated obfuscation. \textit{Project page:
\href{https://github.com/DecETT/DecETT}{https://github.com/DecETT/DecETT}}
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:34:22 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Gu",
"Zheyuan",
""
],
[
"Liu",
"Chang",
""
],
[
"Zhang",
"Xiyuan",
""
],
[
"Yang",
"Chen",
""
],
[
"Gou",
"Gaopeng",
""
],
[
"Xiong",
"Gang",
""
],
[
"Li",
"Zhen",
""
],
[
"Li",
"Sijia",
""
]
] |
2504.15566 | Akira Kitaoka | Akira Kitaoka | Minimization of Curve Length through Energy Minimization using Finite
Difference and Numerical Integration in Real Coordinate Space | 24 pages | null | null | null | math.NA cs.NA math.DG math.OC | http://creativecommons.org/licenses/by-nc-sa/4.0/ | The problem of determining the minimal length is garnering attention in
various fields such as computer vision, robotics, and machine learning. One
solution to this problem involves linearly interpolating the solution of a
nonlinear optimization problem that approximates the curve's energy
minimization problem using finite differences and numerical integration. This
method tends to be easier to implement compared to others. However, it was
previously unknown whether this approach successfully minimizes the curve's
length under the Riemannian metric in real coordinate spaces. In this paper, we
prove that the length of a curve obtained by linear interpolation of the
solution to an optimization problem, where the energy of the curve is
approximated using finite differences and the trapezoidal rule, converges to
the minimal curve length at a rate of $1/2$ in terms of the number of points
used in the numerical integration. Similarly, we prove that when using the
left-point rule, the approximated curve's length likewise converges to the
minimal curve length at a rate of $1/2$ in terms of the number of points used
in the numerical integration.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:35:08 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Kitaoka",
"Akira",
""
]
] |
2504.15567 | Xiao-tao He | Xiao-Tao He, Kun Huang, T. M. Shneidman, N. V. Antonenko, Jun Zhang,
and Hong-qiang You | Study of neutron-deficient nucleus 224Np and its {\alpha} decay by
particle-number-conserving method in the framework of deformed shell model | null | null | null | null | nucl-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The particle-number-conserving (PNC) method in the framework of the deformed
shell model (DSM) is employed to study the properties of the newly discovered
short-lived neutron-deficient nucleus 224Np and its {\alpha}-decay. The
calculated energy of {\alpha}-particle lies within 300 keV of the experimental
data. This is the first application of the PNC method to the region of
neutron-deficient nuclei. This work provides the first attempt to combine the
microscopic PNC theory with empirical formulas to study the nuclear {\alpha}
decay. The configurations of ground states are assigned as
{\pi}5/2-[523]{\otimes}{\nu}5/2+[633] for 224Np,
{\pi}1/2-[530]{\otimes}{\nu}3/2+[642] for 220Pa,
{\pi}3/2+[651]{\otimes}{\nu}1/2-[501] for 216Ac, and
{\pi}7/2-[514]{\otimes}{\nu}3/2-[501] for 212Fr. The absence of the Z = 92
subshell closure in 224Np is explained by analyzing the proton single-particle
levels. Low-lying excited statesare predicted for nuclei along the
{\alpha}-decay chain by the PNC method. Based on the PNC predicted
{\alpha}-decay energy and the assigned configurations, the {\alpha}-decay
half-lives are calculated by the empirical formulas, in which the angular
momentum taken away by the {\alpha} particle is taken into account. The angular
momentum have an important effect on the {\alpha}-decay half-life. The errors
of the empirical formulas calculation are in two orders of magnitude with the
experimental data.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:41:19 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"He",
"Xiao-Tao",
""
],
[
"Huang",
"Kun",
""
],
[
"Shneidman",
"T. M.",
""
],
[
"Antonenko",
"N. V.",
""
],
[
"Zhang",
"Jun",
""
],
[
"You",
"Hong-qiang",
""
]
] |
2504.15568 | Minbiao Han | Michael Albert, Quinlan Dawkins, Minbiao Han, Haifeng Xu | Is Learning Effective in Dynamic Strategic Interactions? Evidence from
Stackelberg Games | null | null | null | null | cs.GT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In many settings of interest, a policy is set by one party, the leader, in
order to influence the action of another party, the follower, where the
follower's response is determined by some private information. A natural
question to ask is, can the leader improve their strategy by learning about the
unknown follower through repeated interactions? A well known folk theorem from
dynamic pricing, a special case of this leader-follower setting, would suggest
that the leader cannot learn effectively from the follower when the follower is
fully strategic, leading to a large literature on learning in strategic
settings that relies on limiting the strategic space of the follower in order
to provide positive results. In this paper, we study dynamic Bayesian
Stackelberg games, where a leader and a \emph{fully strategic} follower
interact repeatedly, with the follower's type unknown. Contrary to existing
results, we show that the leader can improve their utility through learning in
repeated play. Using a novel average-case analysis, we demonstrate that
learning is effective in these settings, without needing to weaken the
follower's strategic space. Importantly, this improvement is not solely due to
the leader's ability to commit, nor does learning simply substitute for
communication between the parties. We provide an algorithm, based on a
mixed-integer linear program, to compute the optimal leader policy in these
games and develop heuristic algorithms to approximate the optimal dynamic
policy more efficiently. Through simulations, we compare the efficiency and
runtime of these algorithms against static policies.
| [
{
"version": "v1",
"created": "Tue, 22 Apr 2025 03:44:05 GMT"
}
] | 2025-04-23T00:00:00 | [
[
"Albert",
"Michael",
""
],
[
"Dawkins",
"Quinlan",
""
],
[
"Han",
"Minbiao",
""
],
[
"Xu",
"Haifeng",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.