May 2025 arXiv papers — page 71

Yusen Xia

In this paper, we proved that for a bounded Hopf-symmetric domain $\Omega$ in a noncompact rank one symmetric space $M$, the second Dirichlet eigenvalue $\lambda_2 (\Omega) \leq \lambda_2 (B_1)$ where $B_1$ is a geodesic ball in $M$ such that $\lambda_1 (\Omega) =\lambda_1 (B_1)$. This generalizes the work of Ashbaugh & Benguria, Benguria & Linde for bounded

Huayou Liu, Jiayuan Zhao, Jing Zhang, Huan Liu

This work demonstrates photon emission gain, i.e., emission of multiple photons per injected electron, through impact excitation in Er-doped silicon light-emitting diodes (LEDs). Conventional methods for exciting Er ions in silicon suffer from low efficiency due to mismatched energy transfer between exciton recombination and Er excitation. Here, we propose a

Talha Enes Ayranci, Florent P. Audonnet, Gerardo Aragon-Camarasa, Mireilla Bikanga Ada

Student engagement is one of the key challenges in robotics and artificial intelligence (AI) education. Tangible learning approaches, such as educational robots, provide an effective way to enhance engagement and learning by offering real-world applications to bridge the gap between theory and practice. However, existing platforms often face barriers such as

AbdelRahim Elmadany, Sang Yun Kwon, Hawau Olamide Toyin, Alcides Alcoba Inciarte

Africa's rich linguistic diversity remains significantly underrepresented in speech technologies, creating barriers to digital inclusion. To alleviate this challenge, we systematically map the continent's speech space of datasets and technologies, leading to a new comprehensive benchmark SimbaBench for downstream African speech tasks. Using SimbaBench, we in

Mohammadreza Iranpour, Mohammad Rasoul Narimani

The Optimal Power Flow (OPF) problem is central to the reliable and efficient operation of power systems, yet its non-convex nature poses significant challenges for finding globally optimal solutions. While convex relaxation techniques such as Quadratic Convex (QC) relaxation have shown promise in providing tight lower bounds, they typically do not guarantee

Yuliang Cai, Jesse Thomason, Mohammad Rostami

Vision-language models (VLMs), such as CLIP, have demonstrated strong performance across a range of downstream tasks. However, CLIP is still limited in negation understanding: the ability to recognize the absence or exclusion of a concept. Existing methods address the problem by using a large language model (LLM) to generate large-scale data of image caption

Zhiyao Zhang, Myeung Suk Oh, FNU Hairi, Ziyue Luo

Actor-critic methods for decentralized multi-agent reinforcement learning (MARL) facilitate collaborative optimal decision making without centralized coordination, thus enabling a wide range of applications in practice. To date, however, most theoretical convergence studies for existing actor-critic decentralized MARL methods are limited to the guarantee of

Semanto Mondal

As a social being, we have an intimate bond with the environment. A plethora of things in human life, such as lifestyle, health, and food are dependent on the environment and agriculture. It comes under our responsibility to support the environment as well as agriculture. However, traditional farming practices often result in inefficient resource use and env

Kweku Andoh Yamoah, Jackson Weako, Emmanuel J. Dorley

In this paper, we introduce the first publicly available English-Kpelle dataset for machine translation, comprising over 2000 sentence pairs drawn from everyday communication, religious texts, and educational materials. By fine-tuning Meta's No Language Left Behind(NLLB) model on two versions of the dataset, we achieved BLEU scores of up to 30 in the Kpe

Sotirios Touliopoulos, Nicholas M. Glykos

Atomic packing is an important metric for characterizing protein structures, as it significantly influences various features including the stability, the rate of evolution and the functional roles of proteins. Packing in protein structures is a measure of the overall proximity between the proteins' atoms and it can vary notably among different structures

Zhidong Zhang

The common feature for a nontrivial hard problem is the existence of nontrivial topological structures, non-planarity graphs, nonlocalities, or long-range spin entanglements in a model system with randomness. For instance, the Boolean satisfiability (K-SAT) problems are nontrivial, due to the existence of non-planarity graphs, nonlocalities, and the randomne

Parisa Momeni, Gabriel Laverghetta, Jay Ligatti, Lingyao Li

Background: User experience strongly influences pharmaceutical drug effectiveness. Social media platforms like X have become major spaces where people share medication-related experiences, especially for widely marketed drugs such as semaglutide. Despite high activity online, how different user groups engage in semaglutide discussions remains unclear. Object

Scott Kostyshak, Neel U. Sukhatme

How much does luck matter to a criminal defendant in a jury trial? We use rich data on jury selection to causally estimate how parties who are randomly assigned a less favorable jury (as proxied by whether their attorneys exhaust their peremptory strikes) fare at trial. Our novel identification strategy is unique in that it captures variation in juror predis

Dennis Bollweg, Xiang Gao, Swagato Mukherjee, Yong Zhao

We present the first lattice QCD calculations of the isovector helicity transverse momentum-dependent parton distribution function (TMDPDF) and the flavor-dependent unpolarized TMDPDFs for up and down quarks in the proton. Our computations utilize domain-wall fermion discretization with physical quark masses. Employing Coulomb-gauge-fixed quark correlation f

Prakhar Mishra, Amir Hossain Raj, Xuesu Xiao, Dinesh Manocha

We address the problem of agile and rapid locomotion, a key characteristic of quadrupedal and bipedal robots. We present a new algorithm that maintains stability and generates high-speed trajectories by considering the temporal aspect of locomotion. Our formulation takes into account past information based on a novel history-aware curriculum Learning (HACL)

Shou Yoshikawa

Let $k$ be a complete non-archimedean non-trivial valued field. In this paper, we investigate whether every $k$-algebra homomorphism between $k$-affinoid algebras is automatically bounded. We show that this property holds if and only if either $|k^\times|^\mathbb{Q} = \mathbb{R}_{>0}$ holds, or $k$ has positive characteristic and is $F$-finite.

James Cuin, Davide Carbone, O. Deniz Akyildiz

We utilise a sampler originating from nonequilibrium statistical mechanics, termed here Jarzynski-adjusted Langevin algorithm (JALA), to build statistical estimation methods in latent variable models. We achieve this by leveraging Jarzynski's equality and developing algorithms based on a weighted version of the unadjusted Langevin algorithm (ULA) with recurs

Khandakar Ashrafi Akbar, Md Nahiyan Uddin, Latifur Khan, Trayce Hockstad

As connected and automated transportation systems evolve, there is a growing need for federal and state authorities to revise existing laws and develop new statutes to address emerging cybersecurity and data privacy challenges. This study introduces a Retrieval-Augmented Generation (RAG) based Large Language Model (LLM) framework designed to support policyma

Menghua Wu, Yujia Bao

AI systems have increasingly become our gateways to the Internet. We argue that just as advertising has driven the monetization of web search and social media, so too will commercial incentives shape the content served by AI. Unlike traditional media, however, the outputs of these systems are dynamic, personalized, and lack clear provenance -- raising concer

Tianyi Ren, Juampablo E. Heras Rivera, Hitender Oswal, Yutong Pan

Stroke is among the top three causes of death worldwide, and accurate identification of stroke lesion boundaries is critical for diagnosis and treatment. Supervised deep learning methods have emerged as the leading solution for stroke lesion segmentation but require large, diverse, and annotated datasets. The ISLES'24 challenge addresses this need by providi

Afshin Bozorgpour, Sina Ghorbani Kolahi, Reza Azad, Ilker Hacihaliloglu

Medical image segmentation, particularly in multi-domain scenarios, requires precise preservation of anatomical structures across diverse representations. While deep learning has advanced this field, existing models often struggle with accurate boundary representation, variability in organ morphology, and information loss during downsampling, limiting their

Junyi Fan, Shuheng Chen, Li Sun, Yong Si

Aplastic anemia is a rare, life-threatening hematologic disorder characterized by pancytopenia and bone marrow failure. ICU admission in these patients often signals critical complications or disease progression, making early risk assessment crucial for clinical decision-making and resource allocation. In this study, we used the MIMIC-IV database to identify

Harsh Vardhan, Heng Zhu, Avishek Ghosh, Arya Mazumdar

In this paper, we analyze the classical $K$-means alternating-minimization algorithm, also known as Lloyd's algorithm (Lloyd, 1956), for a mixture of Gaussians in a data-distributed setting that incorporates local iteration steps. Assuming unlabeled data distributed across multiple machines, we propose an algorithm, LocalKMeans, that performs Lloyd's algorit

Qingwen Liang, Matias Carrasco Kind

This paper examines the impact of managers' non-responses (NORs) during quarterly earnings calls on analyst forecast behavior by developing a novel measure of NORs using two large language models: ChatGPT-4 and LLaMA 3.3. We adopt a three step prompting approach including identification, classification, and evaluation, to extract NORs from earnings call tran

Ali Abedi, Charlene H. Chu, Shehroz S. Khan

Agitation is one of the most common responsive behaviors in people living with dementia, particularly among those residing in community settings without continuous clinical supervision. Timely prediction of agitation can enable early intervention, reduce caregiver burden, and improve the quality of life for both patients and caregivers. This study aimed to d

Prakhar Mishra, Amir Hossain Raj, Xuesu Xiao, Dinesh Manocha

We present Morphology-Control-Aware Reinforcement Learning (McARL), a new approach to overcome challenges of hyperparameter tuning and transfer loss, enabling generalizable locomotion across robot morphologies. We use a morphology-conditioned policy by incorporating a randomized morphology vector, sampled from a defined morphology range, into both the actor

Achkan Salehi

Ballbot (i.e. Ball balancing robot) navigation usually relies on methods rooted in control theory (CT), and works that apply Reinforcement learning (RL) to the problem remain rare while generally being limited to specific subtasks (e.g. balance recovery). Unlike CT based methods, RL does not require (simplifying) assumptions about environment dynamics (e.g.

Gelareh Hajian, Ali Abedi, Bing Ye, Jennifer Campos

Driving is a key component of independence and quality of life for older adults. However, cognitive decline associated with conditions such as mild cognitive impairment and dementia can compromise driving safety and often lead to premature driving cessation. Conditionally automated vehicles, which require drivers to take over control when automation reaches

Sanjana Kamath, Laurent Talon, Meera Ramaswamy, Christopher A. Browne

We use event-driven pore network modeling to study the transport of hydrogel particles through disordered porous media -- a process that underlies diverse applications. By simulating particle advection, deformation, and clogging at the pore scale, we identify a dimensionless "squeezing parameter" that quantitatively predicts the depth to which particles pene

Chandra Kundu, Abiy Tasissa, HanQin Cai

Euclidean Distance Matrix (EDM), which consists of pairwise squared Euclidean distances of a given point configuration, finds many applications in modern machine learning. This paper considers the setting where only a set of anchor nodes is used to collect the distances between themselves and the rest. In the presence of potential outliers, it results in a s

Toshiaki Koike-Akino, Xiangyu Chen, Jing Liu, Ye Wang

Modern foundation models such as large language models (LLMs) and large multi-modal models (LMMs) require a massive amount of computational and memory resources. We propose a new framework to convert such LLMs/LMMs into a reduced-dimension latent structure. Our method extends a local activation-aware tensor decomposition to a global attention-aware joint ten

Jessica Tang, Ali Abedi, Tracey J. F. Colella, Shehroz S. Khan

Exercise-based rehabilitation improves quality of life and reduces morbidity, mortality, and rehospitalization, though transportation constraints and staff shortages lead to high dropout rates from rehabilitation programs. Virtual platforms enable patients to complete prescribed exercises at home, while AI algorithms analyze performance, deliver feedback, an

Yue Jiang, Jichu Li, Yang Liu, Dingkang Yang

We introduce DanmakuTPPBench, a comprehensive benchmark designed to advance multi-modal Temporal Point Process (TPP) modeling in the era of Large Language Models (LLMs). While TPPs have been widely studied for modeling temporal event sequences, existing datasets are predominantly unimodal, hindering progress in models that require joint reasoning over tempor

Seunghyun Lee, Yuqi Gu

This paper considers a challenging problem of identifying a causal graphical model under the presence of latent variables. While various identifiability conditions have been proposed in the literature, they often require multiple pure children per latent variable or restrictions on the latent causal graph. Furthermore, it is common for all observed variables

Ahmed Bouajjani, Constantin Enea, Enrique Román-Calvo

Concurrent accesses to databases are typically grouped in transactions which define units of work that should be isolated from other concurrent computations and resilient to failures. Modern databases provide different levels of isolation for transactions that correspond to different trade-offs between consistency and throughput. Quite often, an application

Yao Fehlis

Scientific laboratories, particularly those in pharmaceutical and biotechnology companies, encounter significant challenges in optimizing workflows due to the complexity and volume of tasks such as compound screening and assay execution. We introduce Cycle Time Reduction Agents (CTRA), a LangGraph-based agentic workflow designed to automate the analysis of l

Valérie Hayot-Sasson, Abby Stevens, Nicholson Collier, Sudershan Sridhar

The COVID-19 pandemic highlighted the need for new data infrastructure, as epidemiologists and public health workers raced to harness rapidly evolving data, analytics, and infrastructure in support of cross-sector investigations. To meet this need, we developed AERO, an automated research and data sharing platform for continuous, distributed, and multi-disci

Yizhou Zhang, Kishan Panaganti, Laixi Shi, Juba Ziani

Differential Privacy (DP) provides a rigorous framework for privacy, ensuring the outputs of data-driven algorithms remain statistically indistinguishable across datasets that differ in a single entry. While guaranteeing DP generally requires explicitly injecting noise either to the algorithm itself or to its outputs, the intrinsic randomness of existing alg

M. Kanafani, X. Fléchard, O. Naviliat-Cuncic, R. Garreau

Benchmarking simulation codes for electron transport and scattering in matter is a crucial step for estimating uncertainties in many applications. However, experimental data for electron energies of a few MeV is scarce to make such comparisons. We report here the measurement and the quantitative analysis of backscattering probabilities of electrons in the en

Debrup Das, Sam O' Nuallain, Razieh Rahimi

We propose RaDeR, a set of reasoning-based dense retrieval models trained with data derived from mathematical problem solving using large language models (LLMs). Our method leverages retrieval-augmented reasoning trajectories of an LLM and self-reflective relevance evaluation, enabling the creation of both diverse and hard-negative samples for reasoning-inte

Menghua Wu, Cai Zhou, Stephen Bates, Tommi Jaakkola

Reasoning large language models achieve impressive test-time scaling by thinking for longer, but this performance gain comes at significant compute cost. Directly limiting test-time budget hurts overall performance, but not all problems are equally difficult. We propose thought calibration to decide dynamically when thinking can be terminated. To calibrate o

Paul Bischoff, Salma Hammani, Maximilian Schiffer

In response to climate goals, growing environmental awareness, and financial incentives, municipalities increasingly seek to electrify public transportation networks. We study the problem of locating stationary and dynamic inductive charging stations for electric vehicles (EVs), allowing detours from fixed transit routes and schedules. Dynamic charging, whic

Sandeep Pirbhulal, Habtamu Abie, Martin Jullum, Didrik Nielsen

The advancements in communication technology (5G and beyond) and global connectivity Internet of Things (IoT) also come with new security problems that will need to be addressed in the next few years. The threats and vulnerabilities introduced by AI/ML based 5G and beyond IoT systems need to be investigated to avoid the amplification of attack vectors on AI/

Jiangbei Yue, He Wang

Crowd behaviour analysis is essential to numerous real-world applications, such as public safety and urban planning, and therefore has been studied for decades. In the last decade or so, the development of deep learning has significantly propelled the research on crowd behaviours. This chapter reviews recent advances in crowd behaviour analysis using deep le

Juan Garcia Nila, Todd A. Brun

We investigate continuous quantum error correction, comparing performance under a Markovian error model to two distinct non-Markovian models. The first non-Markovian model involves an interaction Hamiltonian between the system and an environmental qubit via an X-X coupling, with a "cooling" bath acting on the environment qubit. This model is known to exhibit

Lin Zhao, Yushu Wu, Xinru Jiang, Jianyang Gu

Recent deep learning models demand larger datasets, driving the need for dataset distillation to create compact, cost-efficient datasets while maintaining performance. Due to the powerful image generation capability of diffusion, it has been introduced to this field for generating distilled images. In this paper, we systematically investigate issues present

Liao Peiyuan

We introduce funion, a system providing end-to-end sender-receiver unlinkability for neural network inference. By leveraging the Pigeonhole storage protocol and BACAP (blinding-and-capability) scheme from the Echomix anonymity system, funion inherits the provable security guarantees of modern mixnets. Users can anonymously store input tensors in pseudorandom

Fangqiao Tian, An Luo, Jin Du, Xun Xian

A multi-agent AI system (MAS) is composed of multiple autonomous agents that interact, exchange information, and make decisions based on internal generative models. Recent advances in large language models and tool-using agents have made MAS increasingly practical in areas like scientific discovery and collaborative automation. However, key questions remain:

Steven Kordonowy, Hannes Leipold

The XY-mixer has widespread utilization in modern quantum computing, including in variational quantum algorithms, such as Quantum Alternating Operator Ansatz (QAOA). The XY ansatz is particularly useful for solving Cardinality Constrained Optimization tasks, a large class of important NP-hard problems. First, we give explicit decompositions of the dynamical

Ajmi Ali, Blair Jamieson, Lyndsay Green, Tapendra BC

A completely new under-water dark-room test facility (UWDTF) has been built at the University of Winnipeg during 2021-2023, for the testing of the equipments, optical components and detectors before they might be used in different underwater experiments, like the Hyper-Kamiokande (Hyper-K), and others. The Facility is designed for Research and Development ac

Neofytos Rodosthenous, Mihail Zervos

The maximality principle has been a valuable tool in identifying the free-boundary functions that are associated with the solutions to several optimal stopping problems involving one-dimensional time-homogeneous diffusions and their running maximum processes. In its original form, the maximality principle identifies an optimal stopping boundary function as t

Yi-Xuan Wang, Yuval Gefen

Lindbladian dynamics of open systems may be employed to steer a many-body system towards a non-trivial ground state of a local Hamiltonian. Such protocols provide us with tunable platforms facilitating the engineering and study of non-trivial many-body states. Steering towards a degenerate ground state manifold provides us with a protected platform to employ

Danny Reidenbach, Filipp Nikitin, Olexandr Isayev, Saee Paliwal

De novo 3D molecule generation is a pivotal task in drug discovery. However, many recent geometric generative models struggle to produce high-quality 3D structures, even if they maintain 2D validity and topological stability. To tackle this issue and enhance the learning of effective molecular generation dynamics, we present Megalodon-a family of scalable tr

Kazuyoshi Yoshimi, Yuichi Motoyama, Tatsumi Aoyama, Mitsuaki Kawamura

The Institute for Solid State Physics (ISSP) at The University of Tokyo has been carrying out a software development project named ``the Project for Advancement of Software Usability in Materials Science (PASUMS)". Since the launch of PASUMS, various open-source software programs have been developed/advanced, including ab initio calculations, effective model

Maxime Elkael, Michele Polese, Reshma Prasad, Stefano Maxenti

The evolution toward open, programmable O-RAN and AI-RAN 6G networks creates unprecedented opportunities for Intent-Based Networking (IBN) to dynamically optimize RAN[...]. However, applying IBN effectively to the RAN scheduler [...] remains a significant challenge. Current approaches predominantly rely on coarse-grained network slicing, lacking the granular

Taran Anusorn, Omar Barrera, Jack Kramer, Ian Anderson

This article presents an approach to control the operating frequency and fractional bandwidth (FBW) of miniature acoustic filters in thin-film lithium niobate (TFLN). More specifically, we used first-order antisymmetric (A1) mode lateral-field-excited bulk acoustic wave resonators (XBARs) to achieve efficient operation at 20.5 GHz. Our technique leverages th

Terence Gaffney, Thiago da Silva

In this work, we investigate the projectivized analytic spectrum of the double of a module, establishing some general properties, and we apply these results to $\mbox{Projan}(\cR((JM(X))_D))$ over the origin in $C\times C$, where $C$ is an irreducible curve in a hypersurface $X$.

Shuaishuai Liu, Gergely Biczók

The rise of online social networks, user-gene-rated content, and third-party apps made data sharing an inevitable trend, driven by both user behavior and the commercial value of personal information. As service providers amass vast amounts of data, safeguarding individual privacy has become increasingly challenging. Privacy threat modeling has emerged as a c

Jeba Rezwana, Corey Ford

Effective communication between AI and humans is essential for successful human-AI co-creation. However, many current co-creative AI systems lack effective communication, which limits their potential for collaboration. This paper presents the initial design of the Framework for AI Communication (FAICO) for co-creative AI, developed through a systematic revie

Boyi Wei, Benedikt Stroebl, Jiacen Xu, Joie Zhang

Foundation models are increasingly becoming better autonomous programmers, raising the prospect that they could also automate dangerous offensive cyber-operations. Current frontier model audits probe the cybersecurity risks of such agents, but most fail to account for the degrees of freedom available to adversaries in the real world. In particular, with stro

Abdellah El Mekki, Houdaifa Atou, Omer Nacar, Shady Shehata

Enhancing the linguistic capabilities of Large Language Models (LLMs) to include low-resource languages is a critical research area. Current research directions predominantly rely on synthetic data generated by translating English corpora, which, while demonstrating promising linguistic understanding and translation abilities, often results in models aligned

Jinbang Huang, Yixin Xiao, Zhanguang Zhang, Mark Coates

Pre-trained large language models (LLMs) show promise for robotic task planning but often struggle to guarantee correctness in long-horizon problems. Task and motion planning (TAMP) addresses this by grounding symbolic plans in low-level execution, yet it relies heavily on manually engineered planning domains. To improve long-horizon planning reliability and

Yike Zhang, Eduardo Davalos Anaya, Jack H. Noble

This paper presents a novel method for monocular patient-to-image intraoperative registration, specifically designed to operate without any external hardware tracking equipment or fiducial point markers. Leveraging a synthetic microscopy surgical scene dataset with a wide range of transformations, our approach directly maps preoperative CT scans to 2D intrao

Praphul Singh, Charlotte Dzialo, Jangwon Kim, Sumana Srivatsa

Ensuring clinical data privacy while preserving utility is critical for AI-driven healthcare and data analytics. Existing de-identification (De-ID) methods, including rule-based techniques, deep learning models, and large language models (LLMs), often suffer from recall errors, limited generalization, and inefficiencies, limiting their real-world applicabili

Ashley Davey, Harry Zheng

We prove convergence of the proximal policy gradient method for a class of constrained stochastic control problems with control in both the drift and diffusion of the state process. The problem requires either the running or terminal cost to be strongly convex, but other terms may be non-convex. The inclusion of control-dependent diffusion introduces additio

Qingshun Hu, Caroline Soubiran

Old open clusters (OCs) can constrain the chemical evolution of the Galactic disc through their metallicity gradients and age-metallicity relation but they are affected by low statistics. This work aims to determine precise and homogeneous metallicities for a number of old clusters ($\geq$ 500 Myr) from all-sky catalogues of stellar parameters leveraging Gai

Xinyan Zhao, Yi-Ching Tang, Akshita Singh, Victor J Cantu

We introduce AbBiBench (Antibody Binding Benchmarking), a benchmarking framework for antibody binding affinity maturation and design. Unlike previous strategies that evaluate antibodies in isolation, typically by comparing them to natural sequences with metrics such as amino acid recovery rate or structural RMSD, AbBiBench instead treats the antibody-antigen

Pingchuan Ma, Ziang Yin, Qi Jing, Zhengqi Gao

DONNs leverage light propagation for efficient analog AI and signal processing. Advances in nanophotonic fabrication and metasurface-based wavefront engineering have opened new pathways to realize high-capacity DONNs across various spectral regimes. Training such DONN systems to determine the metasurface structures remains challenging. Heuristic methods are

Anna J. G. O'Grady, Brendan O'Connor, Jared A. Goldberg, Meridith Joyce

The $\sim$$2100$d Long Secondary Period of Betelgeuse's optical lightcurve and radial velocity motivated the prediction of a low-mass stellar companion, expected to be at maximal apparent separation from Betelgeuse around December 2024. We carried out Director's Discretionary Time observations with the Chandra X-ray Observatory to identify any X-ray emission

Jared A. Goldberg, Anna J. G. O'Grady, Meridith Joyce, Christian I. Johnson

Recently, two independent analyses have asserted that the cause of the Long Secondary Period (LSP) observed in the variability spectrum of our nearest red supergiant, Betelgeuse ($\alpha$ Ori), is an as-yet undetected, low-mass binary companion dubbed $\alpha$ Ori B. In this paper, we present the results of a far-UV observational campaign using the STIS eche

Jarrod Ragsdale, Rajendra Boppana

Modeling of command-line interface (CLI) interaction has enabled flexible, execution-free output presentation. However, current approaches struggle to model inputs with complex compositions and inputs whose execution behavior depends on system characteristics. This is due to a lack of shell input-output (ShIO) data in the training distributions used by the m

Paul M. Riechers, Henry R. Bigelow, Eric A. Alt, Adam Shai

We argue that in-context learning (ICL) predictably arises from standard self-supervised next-token pretraining, rather than being an exotic emergent property. This work establishes the foundational principles of this emergence by focusing on in-distribution ICL, demonstrating how models necessarily adapt to context when trained on token sequences, especiall

Julien Chhor, Parker Knight

We consider the problem of detecting a community of densely connected vertices in a high-dimensional bipartite graph of size $n_1 \times n_2$. Under the null hypothesis, the observed graph is drawn from a bipartite Erd\H{o}s-Renyi distribution with connection probability $p_0$. Under the alternative hypothesis, there exists an unknown bipartite subgraph of s

Riley Simmons-Edler, Jean Dong, Paul Lushenko, Kanaka Rajan

Military weapon systems and command-and-control infrastructure augmented by artificial intelligence (AI) have seen rapid development and deployment in recent years. However, the sociotechnical impacts of AI on combat systems, military decision-making, and the norms of warfare have been understudied. We focus on a specific subset of lethal autonomous weapon s

Mahdieh Tahmasebi

Pricing of the lookback options using the Clark-Ocone formula for the underlying assets driven by stochastic L\'evy processes requires computing the Malliavin derivatives of their maximum or minimum on the Wiener-Poisson space and their distributions. In this work, we will find a generalization of the explicit representation of the Clark-Ocone formula on the

Csaba Both, Benjamin Hoover, Hendrik Strobelt, Dmitry Krotov

Multi-task learning improves generalization, but when does it reduce the model capacity required to learn? We provide a systematic study of how joint training affects the learning transition, the minimum model size at which a task can be learned, using nested arithmetic (ListOps) and permutation groups as controlled testbeds. Certain task pairings dramatical

Yike Zhang, Jack H. Noble

Cochlear implant surgery is a treatment for individuals with severe hearing loss. It involves inserting an array of electrodes inside the cochlea to electrically stimulate the auditory nerve and restore hearing sensation. A crucial step in this procedure is mastoidectomy, a surgical intervention that removes part of the mastoid region of the temporal bone, p

Hansa Meghwani, Amit Agarwal, Priyaranjan Pattnayak, Hitesh Laxmichand Patel

Enterprise search systems often struggle to retrieve accurate, domain-specific information due to semantic mismatches and overlapping terminologies. These issues can degrade the performance of downstream applications such as knowledge management, customer support, and retrieval-augmented generation agents. To address this challenge, we propose a scalable har

Zhangxing Bian, Shuwen Wei, Xiao Liang, Yuan-Chiao Lu

Magnetic resonance (MR) tagging is an imaging technique for noninvasively tracking tissue motion in vivo by creating a visible pattern of magnetization saturation (tags) that deforms with the tissue. Due to longitudinal relaxation and progression to steady-state, the tags and tissue brightnesses change over time, which makes tracking with optical flow method

Minwoo Jung, Lanke Frank Tarimo Fu, Maurice Fallon, Ayoung Kim

LiDAR Place Recognition (LPR) is a key component in robotic localization, enabling robots to align current scans with prior maps of their environment. While Visual Place Recognition (VPR) has embraced Vision Foundation Models (VFMs) to enhance descriptor robustness, LPR has relied on task-specific models with limited use of pre-trained foundation-level knowl

Leonardo Fonseca Larrubia, Pedro Alberto Morettin, Chang Chiann

We present the Maximal Overlap Discrete Wavelet Scattering Transform (MODWST), whose construction is inspired by the combination of the Maximal Overlap Discrete Wavelet Transform (MODWT) and the Scattering Wavelet Transform (WST). We also discuss the use of MODWST in classification tasks, evaluating its performance in two applications: stationary signal clas

AmirHossein Safdarian, Milad Mohammadi, Ehsan Jahanbakhsh, Mona Shahamat Naderi

Text-to-SQL systems translate natural language questions into executable SQL queries, and recent progress with large language models (LLMs) has driven substantial improvements in this task. Schema linking remains a critical component in Text-to-SQL systems, reducing prompt size for models with narrow context windows and sharpening model focus even when the e

Nathan Gaby, Xiaojing Ye

We develop a general theoretical framework for optimal probability density control on standard measure spaces, aimed at addressing large-scale multi-agent control problems. In particular, we establish a maximum principle (MP) for control problems posed on infinite-dimensional spaces of probability distributions and control vector fields. We further derive th

Trinity Chung, Yuchen Shen, Nathan C. L. Kong, Aran Nayebi

Tactile sensing remains far less understood in neuroscience and less effective in artificial systems compared to more mature modalities such as vision and language. We bridge these gaps by introducing a novel Encoder-Attender-Decoder (EAD) framework to systematically explore the space of task-optimized temporal neural networks trained on realistic tactile in

Muhammad Iqbal Rochman, Joshua Roy Palathinkal, Vanlin Sathya, Mehmet Yavuz

The 3.55 - 3.7 GHz Citizens Broadband Radio Service (CBRS) band in the U.S., shared with incumbent Navy radars, is witnessing increasing deployments both indoors and outdoors using a shared, licensed model. Among the many use-cases of such private networks is the indoor neutral-host, where cellular customers of Mobile Network Operators (MNOs) can be seamless

Seda Dogan-Tusha, Armed Tusha, Muhammad Iqbal Rochman, Hossein Nasiri

Standard Power (SP) Wi-Fi 6E in the U.S. is just beginning to be deployed outdoors in the shared but unlicensed 6 GHz band under the control of an Automated Frequency Coordination (AFC) system to protect incumbents, while low-power-indoor (LPI) usage has been steadily increasing over the past 2 years. In this paper, we present the first comprehensive measure

Jianyang Gu, Haonan Wang, Ruoxi Jia, Saeed Vahidian

Dataset distillation (DD) has witnessed significant progress in creating small datasets that encapsulate rich information from large original ones. Particularly, methods based on generative priors show promising performance, while maintaining computational efficiency and cross-architecture generalization. However, the generation process lacks explicit contro

Walid A. Hanafy, Li Wu, David Irwin, Prashant Shenoy

Accelerating computing demand, largely from AI applications, has led to concerns about its carbon footprint. Fortunately, a significant fraction of computing demand comes from batch jobs that are often delay-tolerant and elastic, which enables schedulers to reduce carbon by suspending/resuming jobs and scaling their resources down/up when carbon is high/low.

Lucas Bandarkar, Nanyun Peng

Large language models (LLMs) still struggle across tasks outside of high-resource languages. In this work, we investigate cross-lingual transfer to lower-resource languages where task-specific post-training data is scarce. Building on prior work, we first validate that the subsets of model parameters that matter most for mathematical reasoning and multilingu

Yiming Sun, Shuo Chen, Shengyu Chen, Chonghao Qiu

Methane (CH$_4$) is the second most powerful greenhouse gas after carbon dioxide and plays a crucial role in climate change due to its high global warming potential. Accurately modeling CH$_4$ fluxes across the globe and at fine temporal scales is essential for understanding its spatial and temporal variability and developing effective mitigation strategies.

H. Erington, G. Bollen, G. Dykstra, A. Hamaker

Although it is known that the $N=28$ spherical shell closure erodes, the strength of the closure with decreasing proton number $Z<20$ is an open question in nuclear structure. In this region of interest, direct high-precision mass measurements of neutron-rich $^{43-45}$Cl isotopes were performed at the Low Energy Beam and Ion Trap (LEBIT) when coupled to the

Ramin Babaee, Shahab Oveis Gharan, Martin Bouchard

We propose a novel digital-to-analog converter (DAC) weighting architecture that statistically minimizes the distortion caused by random current mismatches. Unlike binary, thermometer-coded, and segmented DACs, the current weights of the proposed architecture are not an integer power of 2 or any other integer number. We present a heuristic algorithm for a st

Karen Fonseca, Leon Suarez-Rodriguez, Andres Jerez, Felipe Gutierrez-Barragan

Phase retrieval (PR) reconstructs phase information from magnitude measurements, known as coded diffraction patterns (CDPs), whose quality depends on the number of snapshots captured using coded phase masks. High-quality phase estimation requires multiple snapshots, which is not desired for efficient PR systems. End-to-end frameworks enable joint optimizatio

Sola Kim, Dongjune Chang, Jieshu Wang

Despite advances in designing personas for Large Language Models (LLM), challenges remain in aligning them with human cognitive processes and representing diverse stakeholder perspectives. We introduce a Social Cognitive Theory (SCT) agent design framework for designing, evaluating, and implementing psychologically grounded LLMs with consistent behavior. Our

Waleed Reda, Abhinav Jangda, Krishna Chintalapudi

As Large Language Models (LLMs) are increasingly deployed for narrow tasks in resource-constrained settings, a central question arises: how much of an LLM is truly necessary for a given task? We present LLM-Sieve, a framework that prunes LLMs down to the minimal parameter subset needed to preserve task performance. Our approach introduces two innovations: (i

Cesar Ayala, S. V. Mikhailov, A. V. Pimikov

We obtain the electromagnetic pion form factor (emFF) $F_\pi$ for spacelike mid-range of momentum transfer in QCD. We use renormalization group (RG) summation within the light cone sum rules (LCSRs) to obtain the QCD radiative corrections to the $F_\pi$ and involve contributions of the leading twist 2 and, twists 4, 6. The additional conditions to apply here

Thalles Silva, Helio Pedrini, Adín Ramírez Rivera

We present Self-Organizing Visual Prototypes (SOP), a new training technique for unsupervised visual feature learning. Unlike existing prototypical self-supervised learning (SSL) methods that rely on a single prototype to encode all relevant features of a hidden cluster in the data, we propose the SOP strategy. In this strategy, a prototype is represented by

Marwa Banna, Nicolas Gilliers, Pei-Lun Tseng

We provide the first quantitative estimates for the rate of convergence in the free multiplicative central limit theorem (CLT), in terms of the Kolmogorov and $r$-Wasserstein distances for $r \geq 1$. While the free additive CLT has been thoroughly studied, including convergence rates, the multiplicative setting remained open in this regard. We consider prod

Mohamed A. Mohamed, Kateryna Nekhomiazh, Vedant Vyas, Marcos M. Jose

Continual reinforcement learning (RL) concerns agents that are expected to learn continually, rather than converge to a policy that is then fixed for evaluation. This setting is well-suited to environments that the agent perceives as changing over time, rendering any static policy ineffective. In continual RL, researchers often simulate such changes either b

Behrad Moniri, Hamed Hassani

Weak-to-strong generalization, where a student model trained on imperfect labels generated by a weaker teacher nonetheless surpasses that teacher, has been widely observed but the mechanisms that enable it have remained poorly understood. In this paper, through a theoretical analysis of simple models, we uncover three core mechanisms that can drive this phen