May 2025 arXiv papers — page 82

Elisa G. Vergamini, Leonardo F. Dos Santos, Cícero Zanette, Yecid Moreno

Controlling the physical interaction with the environment or objects, as humans do, is a shared requirement across different types of robots. To effectively control this interaction, it is necessary to control the power delivered to the load, that is, the interaction force and the interaction velocity. However, it is not possible to control these two quantit

Soumita Hait, Ping Li, Haipeng Luo, Mengxiao Zhang

In the classic expert problem, $\Phi$-regret measures the gap between the learner's total loss and that achieved by applying the best action transformation $\phi \in \Phi$. A recent work by Lu et al., [2025] introduces an adaptive algorithm whose regret against a comparator $\phi$ depends on a certain sparsity-based complexity measure of $\phi$, (almost) rec

Svala Sverrisdóttir

We develop an algebraic geometric framework for Fock space coupled cluster theory in second quantization. In quantum chemistry, many-electron states are represented as elements of the exterior algebra. The fermionic creation and annihilation operators generate the Fermi-Dirac algebra, which can be realized as a Clifford algebra acting on the exterior algebra

Hossein Maghsoumi, Yaser Fallah

The advancement of cooperative autonomous vehicle systems depends heavily on effective coordination between multiple agents, aiming to enhance traffic efficiency, fuel economy, and road safety. Despite these potential benefits, real-world testing of such systems remains a major challenge and is essential for validating control strategies, trajectory modeling

Yangyang Liu, Lei Xue, Sishan Wang, Xiapu Luo

The progress of automotive technologies has made cybersecurity a crucial focus, leading to various cyber attacks. These attacks primarily target the Controller Area Network (CAN) and specialized Electronic Control Units (ECUs). In order to mitigate these attacks and bolster the security of vehicular systems, numerous defense solutions have been proposed.Thes

A. Andrade, I. Saviane, L. Monaco, M. Gullieuszik

Studying metal-poor galaxies is crucial for understanding the physical mechanisms driving the evolution of galaxies. Most observational works in dwarf galaxies employ integral field unit data to investigate gas physics in the entire galaxy body. However, previous studies have not explored the detailed spatially resolved properties of individual extragalactic

Mingyu Yang, Mehdi Rezagholizadeh, Guihong Li, Vikram Appia

With the growing demand for deploying large language models (LLMs) across diverse applications, improving their inference efficiency is crucial for sustainable and democratized access. However, retraining LLMs to meet new user-specific requirements is prohibitively expensive and environmentally unsustainable. In this work, we propose a practical and scalable

David Sychrovský, Jakub Černý, Martin Loebl

Resource distribution is a fundamental problem in economic and policy design, particularly when demand and supply are not naturally aligned. Without regulation, wealthier individuals may monopolize this resource, leaving the needs of others unsatisfied. While centralized distribution can ensure fairer division, it can struggle to manage logistics efficiently

Tamas G. Molnar

Collision-free motion is a fundamental requirement for many autonomous systems. This paper develops a safety-critical control approach for the collision-free navigation of polytope-shaped agents in polytope-shaped environments. A systematic method is proposed to generate control barrier function candidates in closed form that lead to controllers with formal

Abdramane Annour Saad, Julien Arino, Patrick M Tchepmo Djomegni, Mahamat S Daoussa Haggar

We consider a metapopulation model of cholera describing explicitly the movement of individuals and contaminated water between locations as well as a simple vaccination mechanism. The global stability of the disease-free equilibrium point when the location-specific reproduction numbers are all less than unity is established. We then conduct some numerical in

Senol Gulgonul

This paper presents SOSTIAE (Second-Order System Target IAE), a novel PID tuning method that combines IAE minimization with explicit transient response shaping for practical control applications. The algorithm generates optimal PID parameters by matching the closed-loop response to a target second-order system with user-defined settling time (Ts) and percent

Odysseas S. Chlapanis, Dimitrios Galanis, Nikolaos Aletras, Ion Androutsopoulos

We introduce GreekBarBench, a benchmark that evaluates LLMs on legal questions across five different legal areas from the Greek Bar exams, requiring citations to statutory articles and case facts. To tackle the challenges of free-text evaluation, we propose a three-dimensional scoring system combined with an LLM-as-a-judge approach. We also develop a meta-ev

Cehao Yang, Xueyuan Lin, Xiaojun Wu, Chengjin Xu

A practical approach to activate long chain-of-thoughts reasoning ability in pre-trained large language models is to perform supervised fine-tuning on instruction datasets synthesized by strong Large Reasoning Models such as DeepSeek-R1, offering a cost-effective alternative to reinforcement learning. However, large-scale instruction sets with more than 100k

Xiao Yu Cindy Zhang, Carlos R. Ferreira, Francis Rossignol, Raymond T. Ng

Rare diseases, including Inborn Errors of Metabolism (IEM), pose significant diagnostic challenges. Case reports serve as key but computationally underutilized resources to inform diagnosis. Clinical dense information extraction refers to organizing medical information into structured predefined categories. Large Language Models (LLMs) may enable scalable in

Giuseppe De Giacomo, Yong Li, Sven Schewe, Christoph Weinhuber

The temporal logics LTLf+ and PPLTL+ have recently been proposed to express objectives over infinite traces. These logics are appealing because they match the expressive power of LTL on infinite traces while enabling efficient DFA-based techniques, which have been crucial to the scalability of reactive synthesis and adversarial planning in LTLf and PPLTL ove

Camillo Brena

We construct two sequences of closed $4$-dimensional manifolds with non-negative Ricci curvature, diameter bounded from above by $1$, and volume bounded from below by $v>0$, with different fundamental groups but with the same Gromov-Hausdorff limit. This provides a negative answer to the question posed in [J. Pan. Ricci Curvature and Fundamental Groups of Ef

L. R. Colaço, R. F. L. Holanda, Z. C. Santana, R. Silva

We present a cosmological model-independent determination of the Hubble constant, $H_0$, by combining time-delay measurements from seven TDCOSMO systems, Einstein radius measurements, and Type Ia Supernovae data sourced from the Pantheon+ sample. For each lens of time-delay system, we calculate the angular diameter distance $D_{A_l}$ using the product $D^{\t

Arghya Ghosh, Subrata Jana, Manoar Hossain, Dimple Rani

Predicting accurate band gaps and optical properties of lower-dimensional materials, including two-dimensional van der Waals (vdW) materials and their heterostructures, remains a challenge within density functional theory (DFT) due to their unique screening compared to their bulk counterparts. Additionally, accurate treatment of the dielectric response is cr

Yihuai Hong, Yiran Zhao, Wei Tang, Yang Deng

Over time, a growing wave of large language models from various series has been introduced to the community. Researchers are striving to maximize the performance of language models with constrained parameter sizes. However, from a microscopic perspective, there has been limited research on how to better store knowledge in model parameters, particularly withi

Henrique Perez G. Silva

The determination of musical key is a fundamental aspect of music theory and perception, providing a harmonic context for melodies and chord progressions. Automating this process, known as automatic key detection, is a significant task in the field of Music Information Retrieval (MIR). This article outlines an algorithmic methodology for estimating the music

Pablo Barros, Roger Behling, Vincent Guigues, Luiz-Rafael Santos

This paper introduces the Parallelized Circumcentered Reflection Method (P-CRM), a circumcentric approach that parallelizes the Circumcentered Reflection Method (CRM) for solving Convex Feasibility Problems in affine settings. Beyond feasibility, P-CRM solves the best approximation problem for any finite collection of affine subspaces; that is, it not only f

Qihao Duan, Bingding Huang, Zhenqiao Song, Irina Lehmann

Large language models (LLMs) have revolutionized natural language processing and are increasingly applied to other sequential data types, including genetic sequences. However, adapting LLMs to genomics presents significant challenges. Capturing complex genomic interactions requires modeling long-range dependencies within DNA sequences, where interactions oft

Liang Shi, Yun Fu

Recent advances in diffusion models have significantly improved text-to-face generation, but achieving fine-grained control over facial features remains a challenge. Existing methods often require training additional modules to handle specific controls such as identity, attributes, or age, making them inflexible and resource-intensive. We propose ExpertGen,

Guilherme S. Costa, Marcel Novaes, Ricardo Fariello, Marcus A. M. de Aguiar

We study the dynamics of the Kuramoto model on the sphere under higher-order interactions and an external periodic force. For identical oscillators, we introduce a novel way to incorporate three- and four-body interactions into the dynamics of the order parameter, allowing for a full dimensional reduction of this system. We discuss how such reduction can be

Alexey Boldyrev, Fedor Ratnikov, Andrey Shevelev

The rapid development of machine learning (ML) and artificial intelligence (AI) applications requires the training of large numbers of models. This growing demand highlights the importance of training models without human supervision, while ensuring that their predictions are reliable. In response to this need, we propose a novel approach for determining mod

Zakia Tamanna Tisha, Ujjwal Guin

The modern semiconductor industry requires memory solutions that can keep pace with the high-speed demands of high-performance computing. Embedded non-volatile memories (eNVMs) address these requirements by offering faster access to stored data at an improved computational throughput and efficiency. Furthermore, these technologies offer numerous appealing fe

Jiale Tan, Marisa C. Eisenberg

To be fully useful for public health practice, models for epidemic response must be able to do more than predict -- it is also important to incorporate the mechanisms underlying transmission dynamics to enable policymakers and practitioners to be able to evaluate what-if scenarios and intervention options. However, most mechanistic models suffer from uncerta

Purnanand Elango, Abraham P. Vinod, Kenji Kitamura, Behçet Açıkmeşe

We present an optimization-based approach for fuel-efficient spacecraft rendezvous to the Gateway, a space station that will be deployed on a near rectilinear halo orbit (NRHO) around the Moon. The approach: i) ensures passive safety and satisfies path constraints at all times, ii) meets the specifications for critical decision points along the trajectory, i

Razvan-Gabriel Dumitru, Darius Peteleaza, Vikas Yadav, Liangming Pan

Large language models excel at complex tasks by breaking down problems into structured reasoning steps. However, reasoning traces often extend beyond reaching a correct answer, causing wasted computation, reduced readability, and hallucinations. To address this, we introduce a novel hyperparameter-free conciseness score used as a reward signal within a reinf

Yuran Sun, Susu Xu, Chenguang Wang, Xilei Zhao

Big trajectory data hold great promise for human mobility analysis, but their utility is often constrained by the absence of critical traveler attributes, particularly sociodemographic information. While prior studies have explored predicting such attributes from mobility patterns, they often overlooked underlying cognitive mechanisms and exhibited low predi

Chace Ashcraft, Ted Staley, Josh Carney, Cameron Hickert

Backdoor attacks, or trojans, pose a security risk by concealing undesirable behavior in deep neural network models. Open-source neural networks are downloaded from the internet daily, possibly containing backdoors, and third-party model developers are common. To advance research on backdoor attack mitigation, we develop several trojans for deep reinforcemen

Shayan Dadman, Bernt Arild Bremdal, Børre Bang, Rune Dalmo

This paper introduces an unsupervised framework for detecting audio patterns in musical samples (loops) through anomaly detection techniques, addressing challenges in music information retrieval (MIR). Existing methods are often constrained by reliance on handcrafted features, domain-specific limitations, or dependence on iterative user interaction. We addre

Tim Morrison

We study the question of how best to stratify units into matched pairs in online experiments, so that units within a pair receive opposite treatment. Past work by Bai, Romano, and Shaikh (2022) has demonstrated the asymptotic variance improvement that comes from pairing units with similar covariates in this way. However, their method requires knowing the cov

Justine C. Obidowski, Jeremy J. Webb, Simon Portegies Zwart, Maxwell X. Cai

It is unknown if an Oort cloud reaches its maximum mass within its star's birth cluster or millions of years later. Complicating the Oort cloud evolution process is the fact that comets can be stripped from orbit due to perturbations from passing stars. We explore how a star's cluster escape time (t$_{ \rm esc}$) and the time its Oort cloud reaches maximum m

Ryota Yagi

Dataset pruning -- selecting a small yet informative subset of training data -- has emerged as a promising strategy for efficient machine learning, offering significant reductions in computational cost and storage compared to alternatives like dataset distillation. While pruning methods have shown strong performance in image classification, their extension t

Changyi Li, Jiayi Wang, Xudong Pan, Geng Hong

Large Reasoning Models (LRMs) leverage transparent reasoning traces, known as Chain-of-Thoughts (CoTs), to break down complex problems into intermediate steps and derive final answers. However, these reasoning traces introduce unique safety challenges: harmful content can be embedded in intermediate steps even when final answers appear benign. Existing moder

Yakov Berchenko-Kogan, Evan S. Gawlik

The tensor product of two differential forms of degree $p$ and $q$ is a multilinear form that is alternating in its first $p$ arguments and alternating in its last $q$ arguments. These forms, which are known as double forms or $(p,q)$-forms, play a central role in certain differential complexes that arise when studying partial differential equations. We cons

Ramón Fernandez Astudillo, Md Arafat Sultan, Aashka Trivedi, Yousef El-Kurdi

Inference scaling helps LLMs solve complex reasoning problems through extended runtime computation. On top of long chain-of-thought (long-CoT) models, purely inference-time techniques such as best-of-N (BoN) sampling, majority voting, or more generally, minimum Bayes risk decoding (MBRD), can further improve LLM accuracy by generating multiple candidate solu

Niklas Holzner, Sebastian Maier, Stefan Feuerriegel

Generative artificial intelligence (GenAI) is increasingly used to support a wide range of human tasks, yet empirical evidence on its effect on creativity remains scattered. Can GenAI generate ideas that are creative? To what extent can it support humans in generating ideas that are both creative and diverse? In this study, we conduct a meta-analysis to eval

Carlos Areces, Valentin Cassano, Danae Dutto, Raul Fervari

This document serves as a companion to the paper of the same title, wherein we introduce a Gentzen-style sequent calculus for HXPathD. It provides full technical details and proofs from the main paper. As such, it is intended as a reference for readers seeking a deeper understanding of the formal results, including soundness, completeness, invertibility, and

Hongjian Zhou, Haoyu Yang, Ziang Ying, Nicholas Gangi

Driven by innovations in photonic computing and interconnects, photonic integrated circuit (PIC) designs advance and grow in complexity. Traditional manual physical design processes have become increasingly cumbersome. Available PIC layout tools are mostly schematic-driven, which has not alleviated the burden of manual waveguide planning and layout drawing.

Clayton Cohn, Surya Rayala, Caitlin Snyder, Joyce Fonteles

Collaborative dialogue offers rich insights into students' learning and critical thinking, which is essential for personalizing pedagogical agent interactions in STEM+C settings. While large language models (LLMs) facilitate dynamic pedagogical interactions, hallucinations undermine confidence, trust, and instructional value. Retrieval-augmented generation (

Ezequiel A. Galpern, Federico Caamaño, Diego U. Ferreiro

Natural protein sequences somehow encode the structural forms that these molecules adopt. Recent developments in structure-prediction are agnostic to the mechanisms by which proteins fold and represent them as static objects. However, the amino acid sequences also encode information about how the folding process can happen, and how variations in the sequence

Md Shariful Islam, M. Sohel Rahman

Log management is crucial for ensuring the security, integrity, and compliance of modern information systems. Traditional log management solutions face challenges in achieving tamper-proofing, scalability, and real-time processing in distributed environments. This paper presents a blockchain-based log management framework that addresses these limitations by

Omar Moured, Yufan Chen, Ruiping Liu, Simon Reiß

Charts play a critical role in data analysis and visualization, yet real-world applications often present charts with challenging or noisy features. However, "outlier charts" pose a substantial challenge even for Multimodal Large Language Models (MLLMs), which can struggle to interpret perturbed charts. In this work, we introduce CHAOS (CHart Analysis with O

Aryan Sharma, Jaden Li, Christina Chu, Anna Sisk

Congressional Research Service (CRS) reports provide detailed analyses of major policy issues to members of the US Congress. We extract and analyze data from 2,010 CRS reports written between 1996 and 2024 to quantify inter-country relationships, representing 172 countries as nodes and 4,137 shared interests as edges within a weighted, bidirectional network.

Abhinav Raina, Maurizio Sanfilippo, Chang-Ki Moon, Manuel Neubauer

We present a graphene photodetector coupled to a layer of aggregated organic semiconductor. A graphene phototransistor is covered with a thin film of merocyanine molecules. The aggregation of the molecular layer can be controlled by the deposition parameters and post-deposition annealing to obtain films ranging from amorphous to a highly aggregated state. Th

Jipeng Zhang, Haolin Yang, Kehao Miao, Ruiyuan Zhang

Recent text-to-SQL models have achieved strong performance, but their effectiveness remains largely confined to SQLite due to dataset limitations. However, real-world applications require SQL generation across multiple dialects with varying syntax and specialized features, which remains a challenge for current models. The main obstacle in building a dialect-

Arjun Dey, Sergey E. Koposov, Joan R. Najita, Andrew P. Cooper

The Milky Way Backup Program (MWBP), a survey currently underway with the Dark Energy Spectroscopic Instrument (DESI) on the Nicholas U. Mayall 4-m Telescope, works at the margins of the DESI Main surveys to obtain spectra of millions of additional stars from the Gaia catalog. Efficiently utilizing twilight times (<18 deg) and poor weather conditions, the MW

S. A. Syzganbayeva, A. V. Filinov, Jesus Ara, A. B. Ashikbayeva

This paper presents a detailed study of the polarizational stopping power of a homogeneous electron gas in moderate and strong coupling regimes using the self-consistent version of the method of moments as the key theoretical approach capable of expressing the dynamic characteristics of the system in terms of the static ones, which are the moments. We develo

Arash Afkanpour, Omkar Dige, Fatemeh Tavakoli, Negin Baghbanzadeh

Current evaluation frameworks for foundation models rely heavily on static, manually curated benchmarks, limiting their ability to capture the full breadth of model capabilities. This paper introduces Active learning for Capability Evaluation (ACE), a novel framework for scalable, automated, and fine-grained evaluation of foundation models. ACE leverages the

Pengcheng Wan, Zhong-Heng Tan, S. T. Chui, Tiexiang Li

We recently found that the electromagnetic scattering problem can be very fast in an approach expressing the fields in terms of orthonormal basis functions. In this paper we apply computational conformal geometry with the conformal energy minimization (CEM) algorithm to make possible fast solution of finite-frequency electromagnetic problems involving arbitr

Kun Yang, Neena Imam

Federated Learning (FL) enables collaborative machine learning across decentralized data sources without sharing raw data. It offers a promising approach to privacy-preserving AI. However, FL remains vulnerable to adversarial threats from malicious participants, referred to as Byzantine clients, who can send misleading updates to corrupt the global model. Tr

Doohyuk Jang, Yoonjeon Kim, Chanjae Park, Hyun Ryu

Large language models have demonstrated remarkable proficiency in long and complex reasoning tasks. However, they frequently exhibit a problematic reliance on familiar reasoning patterns, a phenomenon we term \textit{reasoning rigidity}. Despite explicit instructions from users, these models often override clearly stated conditions and default to habitual re

Naima Sultana Ayesha, Mehrin Anannya, Md Biplob Hosen, Rashed Mazumder

In this cloud-dependent era, various security techniques, such as encryption, steganography, and hybrid approaches, have been utilized in cloud computing to enhance security, maintain enormous storage capacity, and provide ease of access. However, the absence of data type-specific encryption and decryption procedures renders multimedia data vulnerable. To ad

Siyang Song, Micol Spitale, Xiangyu Kong, Hengde Zhu

In dyadic interactions, a broad spectrum of human facial reactions might be appropriate for responding to each human speaker behaviour. Following the successful organisation of the REACT 2023 and REACT 2024 challenges, we are proposing the REACT 2025 challenge encouraging the development and benchmarking of Machine Learning (ML) models that can be used to ge

Georgios Chochlakis, Peter Wu, Arjun Bedi, Marcus Ma

Modeling complex subjective tasks in Natural Language Processing, such as recognizing emotion and morality, is considerably challenging due to significant variation in human annotations. This variation often reflects reasonable differences in semantic interpretations rather than mere noise, necessitating methods to distinguish between legitimate subjectivity

Korel Gundem, Juncheng Dong, Dennis Zhang, Vahid Tarokh

In-Context Learning (ICL) allows Large Language Models (LLMs) to adapt to new tasks with just a few examples, but their predictions often suffer from systematic biases, leading to unstable performance in classification. While calibration techniques are proposed to mitigate these biases, we show that, in the logit space, many of these methods are equivalent t

V. Anagnostou, E. Papadimitriou, V. Karakostas, T. Back

In March 2024, a swarm like seismic activity occurred north of Kefalonia Island, in the central Ionian Islands area. Following a machine-learning aided workflow, we compiled an enhanced seismic catalog of 2495 low to moderate magnitude earthquakes throughout a 2 month period. Spatiotemporal analysis reveals a narrow epicentral distribution of nearly E-W alig

Thiago T. Bergamaschi

The aim of this work is to present the black hole information problem and discuss the assumptions and hypotheses necessary for its formulation. As the problem arises in the framework of semiclassical gravity, we first review the necessary notions to describe Lorentzian manifolds equipped with physical properties, as well as the physical concepts of the theor

Karoly J. Boroczky, Shibing Chen, Weiru Liu, Christos Saroglou

We prove the $C^0$ estimate for the $L_p$ $q$th dual Minkowski problem on $S^2$ under fairly general conditions; namely, when $p$ lies in [0,1) and $q>2+p$, and the $L_p$ $q$th dual curvarture is bounded and bounded away from zero. We note that it is known that the analogous $C^0$ estimate does not hold if $p<-1$ and $q=3$. As a corollary of our $C^0$ estima

Lianghuan Huang, Shuo Li, Sagnik Anupam, Insup Lee

Reinforcement learning (RL) has emerged as a promising strategy for improving the reasoning capabilities of language models (LMs) in domains such as mathematics and coding. However, most modern RL algorithms were designed to target robotics applications, which differ significantly from LM reasoning. We analyze RL algorithm design decisions for LM reasoning,

Kangda Wei, Hasnat Md Abdullah, Ruihong Huang

Large Language Models (LLMs) often exhibit gender bias, resulting in unequal treatment of male and female subjects across different contexts. To address this issue, we propose a novel data generation framework that fosters exploratory thinking in LLMs. Our approach prompts models to generate story pairs featuring male and female protagonists in structurally

Lior Alon, Gregory Berkolaiko, Mark Goresky

Motivated by the nodal distribution universality conjecture for discrete operators on graphs and by the spectral analysis of their maximal abelian covers, we consider a family of Hermitian matrices $h_{\alpha}$ obtained by varying the complex phases of individual matrix elements. This family is parametrized by a $\beta$-dimensional torus, where $\beta$ is th

Xiaochen Wang, Yuan Zhong, Lingwei Zhang, Lisong Dai

Medical deep learning models depend heavily on domain-specific knowledge to perform well on knowledge-intensive clinical tasks. Prior work has primarily leveraged unimodal knowledge graphs, such as the Unified Medical Language System (UMLS), to enhance model performance. However, integrating multimodal medical knowledge graphs remains largely underexplored,

Ramanakumar Sankar, Michael H Wong, Csaba Palotai, Shawn Brueshaber

Recent observations of convection in the jovian atmosphere have demonstrated that convection is strongly concentrated at specific locations on planet. For instance, observations of lightning show that the cyclonic features (e.g,. belts and folded filamentary regions - FFRs) show increased convective activity compared to anti-cyclonic regions. Meanwhile, the

J. W. González, T. Brumme, E. Suárez Morell, A. M. León

Altermagnetic materials combine compensated magnetic order with momentum-dependent spin splitting, offering a fundamentally new route for spintronic functionality beyond conventional ferromagnets and antiferromagnets. While most studies have focused on three-dimensional compounds, the emergence of altermagnetism in few-layer two-dimensional materials remains

Nandini Das, Juhi Dutta, Dilip Kumar Ghosh, Santosh Kumar Rai

We investigate the possibility of having a vector boson dark matter in a $U(1)_X$ extended two-Higgs-doublet model (2HDM) setup. The gauge boson gains mass when a SM singlet complex scalar, which is charged under the dark $U(1)_X$ symmetry, acquires vacuum expectation value (\textit{vev}). This scalar acts as the connection between the SM sector and DM via t

Martin Villagrana, Francisco Lopez-Tiro, Clement Larose, Gilberto Ochoa-Ruiz

The segmentation of kidney stones is regarded as a critical preliminary step to enable the identification of urinary stone types through machine- or deep-learning-based approaches. In urology, manual segmentation is considered tedious and impractical due to the typically large scale of image databases and the continuous generation of new data. In this study,

Huaiyuan Yao, Pengfei Li, Bu Jin, Yupeng Zheng

Recent advances in autonomous driving research towards motion planners that are robust, safe, and adaptive. However, existing rule-based and data-driven planners lack adaptability to long-tail scenarios, while knowledge-driven methods offer strong reasoning but face challenges in representation, control, and real-world evaluation. To address these challenges

Suifang Zhou, Kexue Fu, Huanmin Yi, Ray Lc

Rapid changes in social networks have transformed the way people express themselves, turning past neologisms, values, and mindsets embedded in these expressions into online heritage. How can we preserve these expressions as cultural heritage? Instead of traditional archiving methods for static material, we designed an interactive and experiential form of arc

Adeep Hande, Kishorekumar Sundararajan, Sardar Hamidian, Ferhan Ture

Millions of people rely on search functionality to find and explore content on entertainment platforms. Modern search systems use a combination of candidate generation and ranking approaches, with advanced methods leveraging deep learning and LLM-based techniques to retrieve, generate, and categorize search results. Despite these advancements, search algorit

Kushal Chawla, Alfy Samuel, Anoop Kumar, Daben Liu

Traditional Retrieval-Augmented Generation (RAG) struggles with complex queries that lack strong signals to retrieve the most relevant context, forcing a trade-off between choosing a small context that misses key information and a large context that confuses the LLM. To address this, we propose Forward-Backward RAG (FB-RAG), a new training-free framework bas

Agripino Sousa-Neto, Maria Aldinez Dantas

In this work, we reconstruct cosmic history via supervised learning through three methods: Classification and Regression Trees (CART), Multi-layer Perceptron Regressor (MLPR), and Support Vector Regression (SVR). For this purpose, we use ages of simulated galaxies based on 32 massive, early-time, passively evolving galaxies in the range $0.12 < z < 1.85$, wi

Minh-Duong Nguyen, Le-Tuan Nguyen, Quoc-Viet Pham

Federated Continual Learning (FCL) has recently emerged as a crucial research area, as data from distributed clients typically arrives as a stream, requiring sequential learning. This paper explores a more practical and challenging FCL setting, where clients may have unrelated or even conflicting data and tasks. In this scenario, statistical heterogeneity an

Jayshawn Cooper, Pilhwa Lee

The sliced Wasserstein flow (SWF), a nonparametric and implicit generative gradient flow, is transformed into a Liouville partial differential equation (PDE)-based formalism. First, the stochastic diffusive term from the Fokker-Planck equation-based Monte Carlo is reformulated as a Liouville PDE-based transport without the diffusive term, essentially reflect

Yi Zhang, Elynn Chen, Yujun Yan

We study contextual dynamic pricing when a target market can leverage K auxiliary markets -- offline logs or concurrent streams -- whose mean utilities differ by a structured preference shift. We propose Cross-Market Transfer Dynamic Pricing (CM-TDP), the first algorithm that provably handles such model-shift transfer and delivers minimax-optimal regret for

Arnav Verma, Kushin Mukherjee, Christopher Potts, Elisa Kreiss

Data visualizations are powerful tools for communicating patterns in quantitative data. Yet understanding any data visualization is no small feat -- succeeding requires jointly making sense of visual, numerical, and linguistic inputs arranged in a conventionalized format one has previously learned to parse. Recently developed vision-language models are, in p

Chaim Chai Elchik, Fatemeh Karimi Nejadasl, Seyed Sahand Mohammadi Ziabari, Ali Mohammed Mansoor Alsahag

Multi-object tracking (MOT) in computer vision has made significant advancements, yet tracking small fish in underwater environments presents unique challenges due to complex 3D motions and data noise. Traditional single-view MOT models often fall short in these settings. This thesis addresses these challenges by adapting state-of-the-art single-view MOT mod

Hyunjin Kim, Gautam Rai, Lorenzo Crippa, Dumitru Călugăru

Magic-angle twisted multilayer graphene stands out as a highly tunable class of moir\'e materials that exhibit strong electronic correlations and robust superconductivity. However, understanding the relations between the low-temperature superconducting phase and the preceding correlated phases established at higher temperatures remains a challenge. Here, we

Arya Kimiaghalam

The Arctic sea ice cover has significantly declined over the recent decades. The debate on whether this decline is caused by anthropogenic activity or internal cycles is still ongoing. However, despite this uncertainty, some physical factors reinforce this declining trend, one of which is sea ice thickness. The thinning of Arctic sea ice facilitates the melt

Shuang Wu, Meijie Wang, Lun Yu

Peptide compounds demonstrate considerable potential as therapeutic agents due to their high target affinity and low toxicity, yet their drug development is constrained by their low membrane permeability. Molecular weight and peptide length have significant effects on the logD of peptides, which in turn influences their ability to cross biological membranes.

Keighley E. Rockcliffe, Elisabeth R. Newton, Allison Youngblood, Girish M. Duvvuri

Atmospheric escape's potential to shape the exoplanet population motivates detailed observations of systems actively undergoing escape. AU Mic is a young and active M dwarf hosting two close-in transiting sub- to Neptune-sized planets. Atmospheric escape was previously detected on the inner planet b, with radially-blown neutral hydrogen producing ~30% blue-s

ShengYun Peng, Pin-Yu Chen, Jianfeng Chi, Seongmin Lee

Finetuning large language models (LLMs) enables user-specific customization but introduces critical safety risks: even a few harmful examples can compromise safety alignment. A common mitigation strategy is to update the model more strongly on examples deemed safe, while downweighting or excluding those flagged as unsafe. However, because safety context can

Leah R. Weiss, Grant T. Smith, Ryan A. Murphy, Bahman Golesorkhi

Optically addressable electronic spins in polyatomic molecules are a promising platform for quantum information science with the potential to enable scalable qubit design and integration through atomistic tunability and nanoscale localization. However, optical state- and site-selection are an open challenge. Here we introduce an organo-erbium spin qubit in w

Abhishek Muhuri, Ayan Patra, Rivu Gupta, Tamoghna Das

We analyze the security of two-way quantum key distribution using arbitrary finite-dimensional systems, considering both individual and collective eavesdropping attacks, without the effective use of entangled states, by incorporating two mutually unbiased bases and Heisenberg-Weyl operators in higher dimensions. For individual attacks, we consider cloning op

Christoph Brause, Rafał Kalinowski, Monika Pilśniak, Ingo Schiemeyer

The distinguishing chromatic number of a graph $G$, denoted $\chi_D(G)$, is the minimum number of colours in a proper vertex colouring of $G$ that is preserved by the identity automorphism only. Collins and Trenk proved that $\chi_D(G)\le 2\Delta(G)$ for any connected graph $G$, and the equality holds for complete balanced bipartite graphs $K_{p,p}$ and for

Zongmeng Yang, Xingyue Yang, Jianhua Wang, Qiang Li

In conventional magnetic tunnel junctions (MTJs), the tunnel magnetoresistance (TMR) typically increases with barrier thickness as electron transmission in the antiparallel configuration decays faster than that of the parallel configuration. In this work, we reveal an anomalous scaling effect in altermagnetic tunnel junctions (AMTJs), where the TMR decreases

William A. Sands, Jing-Mei Qiu, Daniel Hayes, Nanyi Zheng

In this paper, we propose a novel adaptive-rank method for simulating multi-scale BGK equations, based on a greedy sampling strategy. The method adaptively selects important rows and columns of the solution matrix and updates them using a local semi-Lagrangian solver. An adaptive cross approximation then reconstructs the full solution matrix. This extends ou

Baran Hashemi, Kurt Pasque, Chris Teska, Ruriko Yoshida

Can algebraic geometry enhance the sharpness, robustness, and interpretability of modern neural reasoning models by equipping them with a mathematically grounded inductive bias? To answer this, we introduce Tropical Attention, an attention mechanism grounded in tropical geometry that lifts the attention kernel into tropical projective space, where reasoning

Meysam Motaharfar, Parampreet Singh

We provide further evidence that information is preserved during black hole evaporation and may be recoverable, provided quantum gravitational effects resolve the singularity. We demonstrate that due to quantum gravity effects, black holes acquire quantum hair, manifested by non-zero tidal Love numbers, revealing a distinct internal structure similar to neut

Soumak Maitra, Girish Kulkarni, Shikhar Asthana, James S. Bolton

Ly$\alpha$ emitters (LAEs) have now been discovered out to redshift $z=13$, and are valuable probes of the reionization history at redshifts beyond the reach of other currently available tracers. Most inferences of the neutral hydrogen fraction from LAE observations rely on one-point and two-point statistics like the luminosity function and the power spectru

Ruizhe Shen, Tianqi Chen, Ching Hua Lee

Developing methods to accurately characterize and mitigate the impact of noise is crucial for enhancing the fidelity of quantum simulations on Noisy Intermediate-Scale Quantum (NISQ) devices. In this work, we present a circuit structure-preserving error mitigation framework for parameterized quantum circuits. A key advantage of our approach lies in its abili

Sebastian H. Völkel

Understanding the physical significance and spectral stability of black hole quasinormal modes is fundamental to high-precision spectroscopy with future gravitational wave detectors. Inspired by Mashhoon's idea of relating quasinormal modes of black holes with their equivalent bound states in an inverted potential, we investigate, for the first time, energy

Chiara Capecci, Gopal Chandra Santra, Alberto Bottarelli, Emanuele Tirrito

Quantum optimization has emerged as a promising approach for tackling complicated classical optimization problems using quantum devices. However, the extent to which such algorithms harness genuine quantum resources and the role of these resources in their success remain open questions. In this work, we investigate the resource requirements of the Quantum Ap

David Osten

A novel decoupling limit of the membrane is proposed, leading to the $(1+2)$-dimensional classically integrable model originally introduced by Manakov, Zakharov, and Ward. This limit is the large-wrapping regime of a membrane propagating toy background of the form $\mathbb{R}_t \times T^2 \times G$ subject to scaling limit, where $G$ is a Lie group and the g

Alexander Rawlings, Peter H. Johansson, Thorsten Naab, Antti Rantala

We study the detectability of supermassive black holes (SMBHs) with masses of $M_{\bullet}\gtrsim 10^{9}\,\mathrm{M}_\odot$ displaced by gravitational wave recoil kicks $(v_{\rm kick}=0\mathrm{-}2000\,\mathrm{km\,s}^{-1})$ in simulations of merging massive $(M_{\star}>10^{11}\,\mathrm{M}_\odot)$ early-type galaxies. The used KETJU code combines the GADGET-4

Nabamita Banerjee, Vedant Bhutra, Ranveer Kumar Singh

We study the twisted partition function of quarter BPS states in CHL models and show that for a large class of single-centered black holes, the degeneracy of microstates is given by the Fourier coefficients of mock Jacobi forms. Our analysis is a continuation of the programme initiated by Dabholkar, Murthy and Zagier (DMZ) for $1/4$-BPS dyons in $\mathcal{N}

Tanmay Bhore, Lei Su, Ivar Martin, Aashish A. Clerk

The Mpemba effect, where a system initially farther from equilibrium relaxes faster than one closer to equilibrium, has been extensively studied in classical systems and recently explored in quantum settings. While previous studies of the quantum Mpemba effect (QME) have largely focused on isolated systems with global symmetries, we argue that the QME is ubi

Yan-Kun Qu, Zhong-Xiao Man, Yu-Peng Yang, Shuang-Xi Yi

Gamma-ray bursts (GRBs) are generally categorized into long and short bursts based on their duration ($T_{90}$). Recently, it has been proposed that GRBs can also be classified into type I (merger) and type II (collapsar) bursts based on the different origin. From a sample of \textit{Swift} long GRBs~(LGRBs) with a redshift completeness of 60\% and $P \geq 2