May 2025 arXiv papers — page 38

Charlotte Peale, Vinod Raman, Omer Reingold

We introduce "representative generation," extending the theoretical framework for generation proposed by Kleinberg et al. (2024) and formalized by Li et al. (2024), to additionally address diversity and bias concerns in generative models. Our notion requires outputs of a generative model to proportionally represent groups of interest from the training data.

Can Chen, Yunping Huang, Hongwei Zhang, Shimin Wang

Leveraging the concept of the macroscopic fundamental diagram (MFD), perimeter control can alleviate network-level congestion by identifying critical intersections and regulating them effectively. Considering the time-varying nature of travel demand and the equilibrium of accumulation state, we extend the conventional set-point perimeter control (SPC) proble

Xiaomeng Yang, Lei Lu, Qihui Fan, Changdi Yang

Diffusion models have demonstrated exceptional capabilities in generating high-fidelity images. However, their iterative denoising process results in significant computational overhead during inference, limiting their practical deployment in resource-constrained environments. Existing acceleration methods often adopt uniform strategies that fail to capture t

Amr Keleg, Sharon Goldwater, Walid Magdy

Arabic has diverse dialects, where one dialect can be substantially different from the others. In the NLP literature, some assumptions about these dialects are widely adopted (e.g., ``Arabic dialects can be grouped into distinguishable regional dialects") and are manifested in different computational tasks such as Arabic Dialect Identification (ADI). However

Mohammad Attar, Andrew Carse, Yeming Chen, Thomas Green

Particle Builder Online is a web-based education game designed for high school physics students. Students can play against an AI opponent or peers to familiarise themselves with the Standard Model of Particle Physics. The game is aimed at a high school level and tailored to the International Baccalaureate and the Australian Curriculum. Students from four sch

Yunyi Zhang, Ruozhen Yang, Siqi Jiao, SeongKu Kang

Scientific paper retrieval is essential for supporting literature discovery and research. While dense retrieval methods demonstrate effectiveness in general-purpose tasks, they often fail to capture fine-grained scientific concepts that are essential for accurate understanding of scientific queries. Recent studies also use large language models (LLMs) for qu

Alan Moore, Lynna Chu, Zhengyuan Zhu

We present a non-parametric change-point detection approach to detect potentially sparse changes in a time series of high-dimensional observations or non-Euclidean data objects. We target a change in distribution that occurs in a small, unknown subset of dimensions, where these dimensions may be correlated. Our work is motivated by a remote sensing applicati

Madi Matymov, Ba-Hien Tran, Michael Kampffmeyer, Markus Heinonen

Data Augmentation (DA) has become an essential tool to improve robustness and generalization of modern machine learning. However, when deciding on DA strategies it is critical to choose parameters carefully, and this can be a daunting task which is traditionally left to trial-and-error or expensive optimization based on validation performance. In this paper,

Clemens Korn, Joerg Robert

Radio Frequency Identification (RFID) is a widely used technology for identifying and locating objects equipped with low-cost RFID transponders (tags). UHF (Ultra High Frequency) RFID operates in frequency bands around 900 MHz and supports communication distances of up to 15 m between the reader and the tag. Reliable motion detection is therefore a highly re

Aakash Garg, Libing Zeng, Andrii Tsarov, Nima Khademi Kalantari

In this paper, we propose a novel diffusion-based approach to generate stereo images given a text prompt. Since stereo image datasets with large baselines are scarce, training a diffusion model from scratch is not feasible. Therefore, we propose leveraging the strong priors learned by Stable Diffusion and fine-tuning it on stereo image datasets to adapt it t

Clark Mingxuan Ju, Leonardo Neves, Bhuvesh Kumar, Liam Collins

Sequential recommendation is a popular paradigm in modern recommender systems. In particular, one challenging problem in this space is cross-domain sequential recommendation (CDSR), which aims to predict future behaviors given user interactions across multiple domains. Existing CDSR frameworks are mostly built on the self-attention transformer and seek to im

Nathaniel E. Putzig, Gareth A. Morgan, Matthew R. Perry, Bruce A. Campbell

Throughout its mission, the Mars Reconnaissance Orbiter (MRO) has often rolled about its along-track axis by up to 28{\deg} to partially compensate for the suboptimal location of the Shallow Radar (SHARAD) antenna along an edge of the spacecraft that is opposite the imaging payload deck, thereby enhancing the signal-to-noise ratio (S/N) of echoes returned fr

Sun Ye, Zuo Cuiming, Zhang Rui, Shi Bin

Myoelectric interfaces enable intuitive and natural control by decoding residual muscle activity, providing an effective pathway for motor restoration in individuals with preserved musculature. However, in patients with severe muscular atrophy or high-level spinal cord injury, the absence of reliable muscle activity renders myoelectric control infeasible. In

Jaya Narain, Vasudha Kowtha, Colin Lea, Lauren Tooley

Perceptual voice quality dimensions describe key characteristics of atypical speech and other speech modulations. Here we develop and evaluate voice quality models for seven voice and speech dimensions (intelligibility, imprecise consonants, harsh voice, naturalness, monoloudness, monopitch, and breathiness). Probes were trained on the public Speech Accessib

Jam Kraprayoon, Zoe Williams, Rida Fayyaz

This report serves as an accessible guide to the emerging field of AI agent governance. Agents - AI systems that can autonomously achieve goals in the world, with little to no explicit human instruction about how to do so - are a major focus of leading tech companies, AI start-ups, and investors. If these development efforts are successful, some industry lea

Tommy Xu, Zhitian Zhang, Xiangyu Sun, Lauren Kelly Zung

Predictive modeling on tabular data is the cornerstone of many real-world applications. Although gradient boosting machines and some recent deep models achieve strong performance on tabular data, they often lack interpretability. On the other hand, large language models (LLMs) have demonstrated powerful capabilities to generate human-like reasoning and expla

Zhenghai You, Zhenyu Zhou, Lantian Li, Dong Wang

Target confusion, defined as occasional switching to non-target speakers, poses a key challenge for end-to-end speaker extraction (E2E-SE) systems. We argue that this problem is largely caused by the lack of generalizability and discrimination of the speaker embeddings, and introduce a simple yet effective speaker augmentation strategy to tackle the problem.

Manisha Dhillon, Kuldeep Kumar Kataria

In this paper, we introduce and study a time-changed variant of the Erlang queue with multiple arrivals where the time-changing component used is the first hitting time of a tempered stable subordinator. The system of fractional difference-differential equations that governs its state probabilities is derived which is solved to obtain their explicit expressi

Naomi Andrew, Irakli Patchkoria

Using L\"uck's Chern character isomorphism we obtain a general formula in terms of centralisers for the $p$-adic Farrell--Tate $K$-theory of any discrete group $G$ with a finite classifying space for proper actions. We apply this formula to $\text{Out}(F_n)$. The case $n=p+1$ turns out to be especially interesting for the following reason: Up to conjugacy th

Noah Walker

In a recent paper by Harada, Seceleanu, and \c{S}ega, the Hilbert function, betti table, and graded minimal free resolution of a general principal symmetric ideal are determined when the number of variables in the polynomial ring is sufficiently large. In this paper, we strengthen that result by giving a effective bound on the number of variables needed for

Masoud Safilian, Amin Beheshti, Stephen Elbourn

Automated answer grading is a critical challenge in educational technology, with the potential to streamline assessment processes, ensure grading consistency, and provide timely feedback to students. However, existing approaches are often constrained to specific exam formats, lack interpretability in score assignment, and struggle with real-world applicabili

Josefa Lia Stoisser, Marc Boubnovski Martell, Kaspar Märtens, Lawrence Phillips

Electronic health records (EHRs) contain richly structured, longitudinal data essential for predictive modeling, yet stringent privacy regulations (e.g., HIPAA, GDPR) often restrict access to individual-level records. We introduce \textbf{Query, Don't Train} (QDT): a \textbf{structured-data foundation-model interface} enabling \textbf{tabular inference} via

Stanley Yu, Vaidehi Bulusu, Oscar Yasunaga, Clayton Lau

Large Language Models (LLMs) exhibit strong conversational abilities but often generate falsehoods. Prior work suggests that the truthfulness of simple propositions can be represented as a single linear direction in a model's internal activations, but this may not fully capture its underlying geometry. In this work, we extend the concept cone framework, rece

Tim Tsz-Kit Lau, Qi Long, Weijie Su

The ever-growing scale of deep learning models and training data underscores the critical importance of efficient optimization methods. While preconditioned gradient methods such as Adam and AdamW are the de facto optimizers for training neural networks and large language models, structure-aware preconditioned optimizers like Shampoo and Muon, which utilize

Birch Bryant

A well-known result of Walsh states that if $\mathcal T^*$ is an ideal triangulation of an atoroidal, acylindrical, irreducible, compact 3-manifold with torus boundary components, then every properly embedded, two-sided, incompressible surface $S$ is isotopic to a spun-normal surface unless $S$ is isotopic to a fiber or virtual fiber. Previously it was unkno

V. Vilasini, Lin-Qing Chen, Liuhang Ye, Renato Renner

The notions of events and their localisation fundamentally differ between quantum theory and general relativity, reconciling them becomes even more important and challenging in the context of quantum gravity where a classical spacetime background can no longer be assumed. We therefore propose an operational approach drawing from quantum information, to defin

Sajad Khodadadian, Martin Zubeldia

Polyak-Ruppert averaging is a widely used technique to achieve the optimal asymptotic variance of stochastic approximation (SA) algorithms, yet its high-probability performance guarantees remain underexplored in general settings. In this paper, we present a general framework for establishing non-asymptotic concentration bounds for the error of averaged SA it

Saman Khamesian, Asiful Arefeen, Bithika M. Thompson, Maria Adela Grando

High quality real world datasets are essential for advancing data driven approaches in type 1 diabetes (T1D) management, including personalized therapy design, digital twin systems, and glucose prediction models. However, progress in this area has been limited by the scarcity of publicly available datasets that offer detailed and comprehensive patient data.

Claudia Cuttano, Gabriele Trivigno, Giuseppe Averta, Carlo Masone

Few-shot segmentation aims to segment unseen object categories from just a handful of annotated examples. This requires mechanisms that can both identify semantically related objects across images and accurately produce segmentation masks. We note that Segment Anything 2 (SAM2), with its prompt-and-propagate mechanism, offers both strong segmentation capabil

Sean G. Sellers, Juie Shetye, Damian J. Christian, David B. Jess

The Sunspot Solar Observatory Data Archive (SSODA) stores data acquired with the suite of instruments at the Richard B. Dunn Solar Telescope (DST) from February 2018 to the present. The instrumentation at the DST continues to provide high cadence imaging, spectroscopy, and polarimetry of the solar photosphere and chromosphere across a wavelength range from 3

Mohammad Mahdi Naderi, Megan Harris, Ehsanoddin Ghorbanichemazkati, John C. Little

Understanding the complexity and interdependence of systems in the Anthropocene is essential for making informed decisions about societal challenges spanning geophysical, biophysical, sociocultural, and sociotechnical domains. This paper explores the potential of Hetero-functional Graph Theory (HFGT) as a quantification tool for converting Model-based System

Yuanzhe Peng, Jieming Bian, Lei Wang, Yin Huang

Multimodal Federated Learning (MFL) lies at the intersection of two pivotal research areas: leveraging complementary information from multiple modalities to improve downstream inference performance and enabling distributed training to enhance efficiency and preserve privacy. Despite the growing interest in MFL, there is currently no comprehensive taxonomy th

Julia Nakhleh, Robert D. Nowak

Overparameterized neural networks can interpolate a given dataset in many different ways, prompting the fundamental question: which among these solutions should we prefer, and what explicit regularization strategies will provably yield these solutions? This paper addresses the challenge of finding the sparsest interpolating ReLU network--i.e., the network wi

Hilal Asi, Vinod Raman, Kunal Talwar

We design new differentially private algorithms for the problems of adversarial bandits and bandits with expert advice. For adversarial bandits, we give a simple and efficient conversion of any non-private bandit algorithm to a private bandit algorithm. Instantiating our conversion with existing non-private bandit algorithms gives a regret upper bound of $O\

Gabriel Conant, Aycin Iplikci Arodirik, Tora Ozawa, David Zeng

We analyze generalized progressions in some nonabelian groups using a measure of complexity called VC-dimension, which was originally introduced in statistical learning theory by Vapnik and Chervonenkis. Here by a "generalized progression" in a group $G$, we mean a finite subset of $G$ built from a fixed set of generators in analogy to a (multidimensional) a

Eleanor Weckwerth, Andrew J. Howard, Chuan Cheng, Ian Gabalski

We have investigated strong-field-induced electronic coherences in argon and molecular nitrogen ions created by high-intensity, few-cycle infrared laser pulses. This is a step toward the long-sought goal of strong-field coherent control in molecular chemistry. We employed high-intensity, few-cycle infrared laser pulses in a pump-probe setup to investigate a

Yang Xiao, Hisashi Kurata, Ting Wang

Dual-channel closed-loop supply chains (DCCLSCs) play a vital role in attaining both sustainability and profitability. This paper introduces a game-theoretic model to analyze optimal pricing strategies for primary and replacement customers within three distinct recycling frameworks: manufacturer-led, retailer-led, and collaborative recycling. The model ident

Dimitrios Kafetzis, Nikos Fotiou, Savvas Argyropoulos, Jad Nasreddine

The delivery of high-quality, low-latency video streams is critical for remote autonomous vehicle control, where operators must intervene in real time. However, reliable video delivery over Fourth/Fifth-Generation (4G/5G) mobile networks is challenging due to signal variability, mobility-induced handovers, and transient congestion. In this paper, we present

Dasha Metropolitansky, Jonathan Larson

Even when instructed to adhere to source material, language models often generate unsubstantiated content - a phenomenon known as "closed-domain hallucination." This risk is amplified in processes with multiple generative steps (MGS), compared to processes with a single generative step (SGS). However, due to the greater complexity of MGS processes, we argue

Mayank Jobanputra, Yana Veitsman, Yash Sarrof, Aleksandra Bakalova

Transformers have theoretical limitations in modeling certain sequence-to-sequence tasks, yet it remains largely unclear if these limitations play a role in large-scale pretrained LLMs, or whether LLMs might effectively overcome these constraints in practice due to the scale of both the models themselves and their pretraining data. We explore how these archi

Badhan Chandra Das, M. Hadi Amini, Yanzhao Wu

The system prompt in Large Language Models (LLMs) plays a pivotal role in guiding model behavior and response generation. Often containing private configuration details, user roles, and operational instructions, the system prompt has become an emerging attack target. Recent studies have shown that LLM system prompts are highly susceptible to extraction attac

Tharindu Kumarage, Ninareh Mehrabi, Anil Ramakrishna, Xinyan Zhao

Safety reasoning is a recent paradigm where LLMs reason over safety policies before generating responses, thereby mitigating limitations in existing safety measures such as over-refusal and jailbreak vulnerabilities. However, implementing this paradigm is challenging due to the resource-intensive process of creating high-quality policy-embedded chain-of-thou

Hyunsik Yun

Over-smoothing remains a major challenge in Graph Neural Networks (GNNs), where repeated message passing causes node representations to converge and lose discriminative power. To address this, we propose a novel node selection strategy based on Poisson processes, introducing stochastic but structure-aware updates. Specifically, we equip each node with an ind

Trenton Chang, Tobias Schnabel, Adith Swaminathan, Jenna Wiens

Despite advances in large language models (LLMs) on reasoning and instruction-following tasks, it is unclear whether they can reliably produce outputs aligned with a variety of user goals, a concept called steerability. Two gaps in current LLM evaluation impede steerability evaluation: (1) many benchmarks are built with past LLM chats and Internet-scraped te

Thomas Fischer, Yury Person

A conjecture of Talagrand (2010) states that the so-called expectation and fractional expectation thresholds are always within at most some constant factor from each other. In this note we generalize a method of DeMarco and Kahn and settle a few more special cases.

Chutong Meng, Antonios Anastasopoulos

This paper describes the GMU systems for the IWSLT 2025 low-resource speech translation shared task. We trained systems for all language pairs, except for Levantine Arabic. We fine-tuned SeamlessM4T-v2 for automatic speech recognition (ASR), machine translation (MT), and end-to-end speech translation (E2E ST). The ASR and MT models are also used to form casc

Yanbo Wang, Justin Dauwels, Yilun Du

Generative models have demonstrated remarkable abilities in generating high-fidelity visual content. In this work, we explore how generative models can further be used not only to synthesize visual content but also to understand the properties of a scene given a natural image. We formulate scene understanding as an inverse generative modeling problem, where

Daniel Hauck, Stefan Rex, Markus Garst

Identifying realistic platforms capable of controlled operations with Majorana bound states is a key challenge in the study of topological superconductivity. Among the most promising proposals are magnet-superconductor hybrid devices, which employ magnetic textures to engineer regions of non-trivial topology. Here, we consider the remarkably simple case of $

Miguel Ballesteros, Ramsés H. Mena, Arno Siri-Jégousse, Gabor Toth

The Curie-Weiss model is used to study phase transitions in statistical mechanics and has been the object of rigorous analysis in mathematical physics. We analyse the problem of reconstructing the probability measure of a multi-group Curie-Weiss model from a sample of data by employing the maximum likelihood estimator for the coupling parameters of the model

Bao Pham, Gabriel Raya, Matteo Negri, Mohammed J. Zaki

Dense Associative Memories (DenseAMs) are generalizations of Hopfield networks, which have superior information storage capacity and can store training data points (memories) at local minima of the energy landscape. When the amount of training data exceeds the critical memory storage capacity of these models, new local minima, which are different from the tr

Tom Gustafsson, Rolf Stenberg

We consider two methods for treating elastic contact problems with the finite element method; the penalty method and Nitsche's method. For the penalty method we discuss how the penalty parameter should be chosen. Both the theoretical analysis and numerical examples show that an optimal convergence rate cannot be achieved. The method is contrasted to that of

Michael Klamkin, Arnaud Deza, Sikai Cheng, Haoruo Zhao

Consider the following task taught in introductory optimization courses which addresses challenges articulated by the community at the intersection of (generative) AI and OR: generate the dual of a linear program. LLMs, being trained at web-scale, have the conversion process and many instances of Primal to Dual Conversion (P2DC) at their disposal. Students m

Mohammadreza Iranpour, Mohammad Rasoul Narimani, Xudong Jia

The growing adoption of electric vehicles (EVs) is expected to significantly increase demand on electric power distribution systems, many of which are already nearing capacity. To address this, the paper presents a comprehensive framework for analyzing the impact of large-scale EV integration on distribution networks. Using the open-source simulator OpenDSS,

Reza Khanmohammadi, Erfan Miahi, Mehrsa Mardikoraem, Simerjot Kaur

Miscalibration in Large Language Models (LLMs) undermines their reliability, highlighting the need for accurate confidence estimation. We introduce CCPS (Calibrating LLM Confidence by Probing Perturbed Representation Stability), a novel method analyzing internal representational stability in LLMs. CCPS applies targeted adversarial perturbations to final hidd

Vincent Guan, Joseph Janssen, Nicolas Lanzetti, Antonio Terpin

Identifying the drift and diffusion of an SDE from its population dynamics is a notoriously challenging task. Researchers in machine learning and single-cell biology have only been able to prove a partial identifiability result: for potential-driven SDEs, the gradient-flow drift can be identified from temporal marginals if the Brownian diffusivity is already

Jashanjot Singh Sidhu, Abdelhak Bentaleb

The QUIC transport protocol represents a significant evolution in web transport technologies, offering improved performance and reduced latency compared to traditional protocols like TCP. Given the growing number of QUIC implementations, understanding their performance, particularly in video streaming contexts, is essential. This paper presents a comprehensi

Lucia De Luca, Michael Goldman, Marcello Ponsiglione

This paper deals with the dynamics - driven by the gradient flow of negative fractional seminorms - of empirical measures towards equi-spaced ground states. Specifically, we consider periodic empirical measures $\mu$ on the real line that are screened by the Lebesgue measure, i.e., with $\mu-d x$ having zero average. To each of these measures $\mu$ we associ

Xiaoyan Li, Shixin Xu, Faisal Habib, Arvind Gupta

Reconstructing ECG from PPG is a promising yet challenging task. While recent advancements in generative models have significantly improved ECG reconstruction, accurately capturing fine-grained waveform features remains a key challenge. To address this, we propose a novel PPG-to-ECG reconstruction method that leverages a Vision Transformer (ViT) as the core

Zehua Liu, Xiaolou Li, Li Guo, Lantian Li

Visual Speech Recognition (VSR) transcribes speech by analyzing lip movements. Recently, Large Language Models (LLMs) have been integrated into VSR systems, leading to notable performance improvements. However, the potential of LLMs has not been extensively studied, and how to effectively utilize LLMs in VSR tasks remains unexplored. This paper systematicall

Sohyun An, Ruochen Wang, Tianyi Zhou, Cho-Jui Hsieh

While recent success of large reasoning models (LRMs) significantly advanced LLMs' reasoning capability by optimizing the final answer accuracy using reinforcement learning, they may also drastically increase the output length due to overthinking, characterized by unnecessarily complex reasoning paths that waste computation and potentially degrade the perfor

Albin Petersson

In the paper, we analyze the Lebesgue exponents $p_\Phi$ and $q_\Phi$, and show that for any $p_\Phi< p < \infty$ and $1< q<q_\Phi$, there exists an equivalent Young function $\Psi$ with $p < p_\Psi < \infty$ and $1<q_\Psi < q$. This type of construction is used to improve upon the inclusions $L^{p_\Phi}\cap L^{q_\Phi}\subseteq L^\Phi \subseteq L^{p_\Phi} +

Harrison Gaebler, Wesley R Perkins

When studying the stability of $T$-periodic solutions to partial differential equations, it is common to encounter subharmonic perturbations, i.e. perturbations which have a period that is an integer multiple (say $n$) of the background wave, and localized perturbations, i.e. perturbations that are integrable on the line. Formally, we expect solutions subjec

Tomppa Pakarinen

This work explores the behaviour of demagnetizing tensors for general ellipsoids under arbitrary rotations in homogeneous magnetic fields. The work is motivated by the concerns in magnetic resonance imaging safety and their practical evaluation in clinical environments. Whereas demagnetizing tensor is a well-defined concept in the principal axes, its transfo

Pranta Rahman Sarkar, Outi Tammisola, Ranajay Ghosh

We investigate the nonlinear viscoelastic behavior of a biomimetic scale-covered beam in which shear-dependent complex fluids are trapped between overlapping scales under bending loads. These fluids mimic biological mucus and slime layers commonly enveloping the skins found in nature. An energy-based analytical model is developed to quantify the interplay be

Shabnam Nikbakhsh, Eija I. Tanskanen, Thomas Hackman

Aims. This study investigates the magnetic evolution of solar active regions (ARs), with a particular focus on understanding how the magnetic morphology of simple and complex ARs changes throughout their lifetime. Methods. To analyse the magnetic evolution of ARs, we developed a Magnetic Evolution Method (MEM) that segments each region's lifetime into three

Uğur Önal, Sanem Sariel, Metin Sezgin, Ergun Akleman

This article presents a case study comparing the capabilities of humans and artificial intelligence (AI) for visual storytelling. We developed detailed instructions to recreate a three-panel Nancy cartoon strip by Ernie Bushmiller and provided them to both humans and AI systems. The human participants were 20-something students with basic artistic training b

Maria Patrou, Thomas Wang, Wael Elwasif, Markus Eisenbach

With high-performance computing systems now running at exascale, optimizing power-scaling management and resource utilization has become more critical than ever. This paper explores runtime power-capping optimizations that leverage integrated CPU-GPU power management on architectures like the NVIDIA GH200 superchip. We evaluate energy-performance metrics tha

Yubin Kim, Zhiyuan Hu, Hyewon Jeong, Eugene Park

Large Language Models (LLMs) as clinical agents require careful behavioral adaptation. While adept at reactive tasks (e.g., diagnosis reasoning), LLMs often struggle with proactive engagement, like unprompted identification of critical missing information or risks. We introduce BehaviorBench, a comprehensive dataset to evaluate agent behaviors across a clini

Bernardo A. Huberman, Jing Wang

The use of quantm mechanisms in the service of voting security suffers from the problem that in order to generate keys for voters and verifiers a point to point connection has to be physically established for each pair, rendering this impractical. We thus propose using Point-to-Multipoint quantum key distribution (QKD) via time division multiplexing (TDM) an

Chengyue Huang, Brisa Maneechotesuwan, Shivang Chopra, Zsolt Kira

Visual question answering (VQA) systems face significant challenges when adapting to real-world data shifts, especially in multi-modal contexts. While robust fine-tuning strategies are essential for maintaining performance across in-distribution (ID) and out-of-distribution (OOD) scenarios, current evaluation settings are primarily unimodal or particular to

Martin Büchner, Liza Dahiya, Simon Dorer, Vipul Ramtekkar

Visual loop closure detection traditionally relies on place recognition methods to retrieve candidate loops that are validated using computationally expensive RANSAC-based geometric verification. As false positive loop closures significantly degrade downstream pose graph estimates, verifying a large number of candidates in online simultaneous localization an

Roberto Ulloa, Eve M. Zucker, Daniel Bultmann, David J. Simon

The proliferation of large language models (LLMs) can influence how historical narratives are disseminated and perceived. This study explores the implications of LLMs' responses on the representation of mass atrocity memory, examining whether generative AI systems contribute to prosthetic memory, i.e., mediated experiences of historical events, or to what we

Mengchen Dong, Levin Brinkmann, Omar Sherif, Shihan Wang

Experimental evidence on worker responses to AI management remains mixed, partly due to limitations in experimental fidelity. We address these limitations with a customized workplace in the Minecraft platform, enabling high-resolution behavioral tracking of autonomous task execution, and ensuring that participants approach the task with well-formed expectati

Radoslaw Klimek

Ambient-awareness in conjunction with pervasive computing is a significant challenge for system designers. It follows the necessity of gathering raw, massive and heterogeneous environmental data \newrrr{which we} obtained, while middleware processes must merge context modelling and reasoning seamlessly. We proposed a system supporting mountain rescuers which

Vivienne Huiling Wang, Tinghuai Wang, Joni Pajarinen

Hierarchical reinforcement learning (HRL) learns to make decisions on multiple levels of temporal abstraction. A key challenge in HRL is that the low-level policy changes over time, making it difficult for the high-level policy to generate effective subgoals. To address this issue, the high-level policy must capture a complex subgoal distribution while also

M. Reza Ebrahimi, Roland Memisevic

The role of hidden units in recurrent neural networks is typically seen as modeling memory, with research focusing on enhancing information retention through gating mechanisms. A less explored perspective views hidden units as active participants in the computation performed by the network, rather than passive memory stores. In this work, we revisit bilinear

John Hood, Caterina De Bacco, Aaron Schein

Complex systems are often driven by higher-order interactions among multiple units, naturally represented as hypergraphs. Understanding dependency structures within these hypergraphs is crucial for understanding and predicting the behavior of complex systems but is made challenging by their combinatorial complexity and computational demands. In this paper, w

K. L. Luhman

I present infrared spectroscopy of 37 brown dwarf candidates in the Upper Sco association, 35 of which are classified as young and cool, making them likely members. This sample includes many of the faintest spectroscopically confirmed members ($K=16$-17 mag), which should have masses down to $\sim0.007$-0.01 $M_\odot$ for the range of ages in Upper Sco (7-14

Arif Masrur, Peder A. Olsen, Paul R. Adler, Carlan Jackson

Unmanned Aircraft Systems (UAS) and satellites are key data sources for precision agriculture, yet each presents trade-offs. Satellite data offer broad spatial, temporal, and spectral coverage but lack the resolution needed for many precision farming applications, while UAS provide high spatial detail but are limited by coverage and cost, especially for hype

Jules Vliem, Najac Chloe, Beatrice Lena, Andrew Webb

Purpose: One of the intrinsic limitations of low-field MRI is low signal-to-noise ratio (SNR), which can be further reduced by electromagnetic interference (EMI) due to the lack of a Faraday shielded room. To address this issue, we propose a novel RF coil design that is inherently less sensitive to EMI while maintaining high receive sensitivity. Methods: The

Simone Severini

This piece plays with the idea of the Computocene: an era defined not merely by the ubiquity of computers, but by their deepening role in how we observe, interpret, and make sense of the world. Rather than emphasizing automation, speed, scale, or intelligence, computation is reframed as a mode of attention: filtering information, guiding inquiry, reframing q

Zihao Li, Xinyuan Cao, Xiangbo Gao, Kexin Tian

Traffic safety science has long been hindered by a fundamental data paradox: the crashes we most wish to prevent are precisely those events we rarely observe. Existing crash-frequency models and surrogate safety metrics rely heavily on sparse, noisy, and under-reported records, while even sophisticated, high-fidelity simulations undersample the long-tailed s

Minjae Lee, Minhyuk Seo, Tingyu Qu, Tinne Tuytelaars

In continual instruction tuning (CIT) scenarios, where new instruction tuning data continuously arrive in an online streaming manner, training delays from large-scale data significantly hinder real-time adaptation. Data selection can mitigate this overhead, but existing strategies often rely on pretrained reference models, which are impractical in CIT setups

Briglia Maria Rosaria, Mujtaba Hussain Mirza, Giuseppe Lisanti, Iacopo Masi

We answer the question in the title, showing that adversarial training (AT) for diffusion models (DMs) fundamentally differs from classifiers: while AT in classifiers enforces output invariance, AT in DMs requires equivariance to keep the diffusion process aligned with the data distribution. AT is a way to enforce smoothness in the diffusion flow, improving

Dongjune Chang, Sola Kim, Young Soo Park

Nuclear waste management requires rigorous regulatory compliance assessment, demanding advanced decision-support systems capable of addressing complex legal, environmental, and safety considerations. This paper presents a multi-agent Retrieval-Augmented Generation (RAG) system that integrates large language models (LLMs) with document retrieval mechanisms to

Marvin Limpijankit, Yanda Chen, Melanie Subbiah, Nicholas Deas

LLMs can be unpredictable, as even slight alterations to the prompt can cause the output to change in unexpected ways. Thus, the ability of models to accurately explain their behavior is critical, especially in high-stakes settings. One approach for evaluating explanations is counterfactual simulatability, how well an explanation allows users to infer the mo

Zosia Krusberg

This paper presents a photo elicitation project that invites students to explore their identities in STEM and surface the cultural wealth they bring into the physics classroom. Grounded in critical race theory, educational theory, and affective neuroscience, the project asks students to take original photographs representing aspects of their lived experience

Bárbara Muniz

We give an alternative proof of Naito--Sagaki's conjecture, which states that the restriction of $gl_{2n}(\mathbb{C})$-representations to $sp_{2n}(\mathbb{C})$ can be described in terms of crystals. Using the tableau model for crystals, we construct an explicit and self-contained bijection between their highest weight elements and Sundaram's branching model.

Nikolay Filonov, Michael Levitin, Iosif Polterovich, David A. Sher

We prove P\'olya's conjecture for the eigenvalues of the Dirichlet Laplacian on annular domains. Our approach builds upon and extends the methods we previously developed for disks and balls. It combines variational bounds, estimates of Bessel phase functions, refined lattice point counting techniques, and a rigorous computer-assisted analysis. As a by-produc

Marlis Ontivero-Ortega, Gorana Mijatovic, Luca Faes, Daniele Marinazzo

Factor analysis is a well-known statistical method to describe the variability of observed variables in terms of a smaller number of unobserved latent variables called factors. Even though latent factors are conceptually independent of each other, their influence on the observed variables is often joint and synergistic. We propose to quantify the synergy of

Zehua Liu, Xiaolou Li, Chen Chen, Lantian Li

This paper presents the second Chinese Continuous Visual Speech Recognition Challenge (CNVSRC 2024), which builds on CNVSRC 2023 to advance research in Chinese Large Vocabulary Continuous Visual Speech Recognition (LVC-VSR). The challenge evaluates two test scenarios: reading in recording studios and Internet speech. CNVSRC 2024 uses the same datasets as its

Michael A. McGuire, Matthew S. Cook, Brenden R. Ortiz, Jiaqiang Yan

We report results from our experimental investigation of the distorted-kagome compound FeCrAs. For this work, we developed a procedure using tin metal as a flux to produce needlelike crystals. The crystals were characterized by single crystal x-ray diffraction as well as measurements of magnetization, electrical transport, and heat capacity. The physical beh

Felix Jahncke, Johannes Betz

Developing robust, efficient navigation algorithms is challenging. Rule-based methods offer interpretability and modularity but struggle with learning from large datasets, while end-to-end neural networks excel in learning but lack transparency and modularity. In this paper, we present MIND-Stack, a modular software stack consisting of a localization network

Taein Kim, Karstan Bock, Claire Luo, Amanda Liswood

Online data scraping has taken on new dimensions in recent years, as traditional scrapers have been joined by new AI-specific bots. To counteract unwanted scraping, many sites use tools like the Robots Exclusion Protocol (REP), which places a robots$.$txt file at the site root to dictate scraper behavior. Yet, the efficacy of the REP is not well-understood.

Ruijie Zhang, Ziyue Liu, Zhengyang Wang, Zheng Zhang

Training foundation models such as ViTs and LLMs requires tremendous computing cost. Low-rank matrix or tensor factorization offers a parameter-efficient alternative, but often downgrades performance due to the restricted parameter space. In this work, we introduce {\textbf{Latent Crossing (LaX)}} -- a simple yet effective plug-and-play module that enhances

Quentin Delfosse, Jannis Blüml, Fabian Tatai, Théo Vincent

Deep reinforcement learning (RL) agents achieve impressive results in a wide variety of tasks, but they lack zero-shot adaptation capabilities. While most robustness evaluations focus on tasks complexifications, for which human also struggle to maintain performances, no evaluation has been performed on tasks simplifications. To tackle this issue, we introduc

Stéphane Aroca-Ouellette, Natalie Mackraz, Barry-John Theobald, Katherine Metcalf

Accommodating human preferences is essential for creating aligned LLM agents that deliver personalized and effective interactions. Recent work has shown the potential for LLMs acting as writing agents to infer a description of user preferences. Agent alignment then comes from conditioning on the inferred preference description. However, existing methods ofte

Priyam Das, Sarah Robinson, Christine B. Peterson

Motivation: Model selection is a ubiquitous challenge in statistics. For penalized models, model selection typically entails tuning hyperparameters to maximize a measure of fit or minimize out-of-sample prediction error. However, these criteria fail to reflect other desirable characteristics, such as model sparsity, interpretability, or smoothness. Results:

Amir Said, Xin Zhao, Marta Karczewicz, Hilmi E. Egilmez

For the last few decades, the application of signal-adaptive transform coding to video compression has been stymied by the large computational complexity of matrix-based solutions. In this paper, we propose a novel parametric approach to greatly reduce the complexity without degrading the compression performance. In our approach, instead of following the con

Aditya Sinha, Zilinghan Li, Tingkai Liu, Volodymyr Kindratenko

Federated learning (FL) is a distributed machine learning (ML) approach that allows multiple clients to collaboratively train ML models without exchanging original training data, offering a solution that is particularly valuable in sensitive domains such as biomedicine. However, training robust FL models often requires substantial computing resources from pa