Papers
Research papers from arXiv and related sources
Off-Policy Value-Based Reinforcement Learning for Large Language Models
Improving data utilization efficiency is critical for scaling reinforcement learning (RL) for long-horizon tasks where generating trajectories is expensive. However, the dominant RL methods for LLM...
Peng-Yuan Wang, Ziniu Li, Tian Xu, Bohan Yang, Tian-Shuo Liu, ChenYang Wang, Xiong-Hui Chen, Yi-C...
Design Space and Implementation of RAG-Based Avatars for Virtual Archaeology
Immersive technologies, such as virtual and augmented reality, are transforming digital heritage by enabling users to explore and interact with culturally significant sites. It is now possible to v...
Wilhelm Kerle-Malcharek, Giulio Biondi, Karsten Klein, Ulf Hailer, Steffen Diefenbach, Fabrizio G...
What a Mesh: Formal Security Analysis of WPA3 SAE Wireless Authentication
The latest Wi-Fi security standard, IEEE 802.11, includes a secure authentication protocol called SAE, whose use is mandatory for WPA3-Personal networks. The protocol is specified at two separate b...
Roberto Metere, Mario Lilli, Luca Arnaboldi, Elvinia Riccobene
Physics-Informed AI for Laser-Enhanced Contact Optimization in Silicon PV: Electrothermal Activation, Degradation Regimes, and Process Control
Laser-enhanced contact optimization (LECO) is increasingly used method to reduce contact resistance and recover fill factor in advanced crystalline silicon solar cells. However, the industrial tran...
Donald Intal, Abasifreke U. Ebong
RelayS2S: A Dual-Path Speculative Generation for Real-Time Dialogue
Real-time spoken dialogue systems face a fundamental tension between latency and response quality. End-to-end speech-to-speech (S2S) models respond immediately and naturally handle turn-taking, bac...
Long Mai
An Explainable AI-Driven Framework for Automated Brain Tumor Segmentation Using an Attention-Enhanced U-Net
Computer-aided segmentation of brain tumors from MRI data is of crucial significance to clinical decision-making in diagnosis, treatment planning, and follow-up disease monitoring. Gliomas, owing t...
MD Rashidul Islam, Bakary Gibba
Numerical Kernels on a Spatial Accelerator: A Study of Tenstorrent Wormhole
As AI accelerators gain prominence, their potential for traditional scientific computing workloads remains unclear. This paper explores Tenstorrent's Wormhole architecture, a spatial computing plat...
Maya Taylor, Carl Pearson, Luc Berger-Vergiat, Giovanni Long, Jan Ciesko
Data-calibrated point spread function prediction: General description of the method and demonstration on MUSE-NFM
Precise knowledge of the point spread function (PSF) underpins many data analysis steps in astronomy, from photometry and astrometry to source de-blending and deconvolution. In adaptive optics (AO)...
Arseniy Kuznetsov, Benoit Neichel, Sylvain Oberti, Thierry Fusco
Leveraging LLMs and Social Media to Understand User Perception of Smartphone-Based Earthquake Early Warnings
Android's Earthquake Alert (AEA) system provided timely early warnings to millions during the Mw 6.2 Marmara Ereglisi, Türkiye earthquake on April 23, 2025. This event, the largest in the region in...
Hanjing Wang, S. Mostafa Mousavi, Patrick Robertson, Richard M. Allen, Alexie Barski, Robert Bosc...
Unilateral Relationship Revision Power in Human-AI Companion Interaction
When providers update AI companions, users report grief, betrayal, and loss. A growing literature asks whether the norms governing personal relationships extend to these interactions. So what, if a...
Benjamin Lange
Time-Delay Systems with Discrete and Distributed delays: Discontinuous Initial Conditions and Reachability Sets
Time-invariant finite-dimensional systems, under reasonable continuity assumptions, exhibit the property that if solutions exist for all future times, the set of vectors reachable from a bounded se...
Hernan Haimovich, Jose L. Mancilla-Aguilar
Curriculum-Driven 3D CT Report Generation via Language-Free Visual Grafting and Zone-Constrained Compression
Automated radiology report generation from 3D computed tomography (CT) volumes is challenging due to extreme sequence lengths, severe class imbalance, and the tendency of large language models (LLM...
V. K. Cody Bumgardner, Mitchell A. Klusty, Mahmut S. Gokmen, Evan W. Damron
Security Barriers to Trustworthy AI-Driven Cyber Threat Intelligence in Finance: Evidence from Practitioners
Financial institutions face increasing cyber risk while operating under strict regulatory oversight. To manage this risk, they rely heavily on Cyber Threat Intelligence (CTI) to inform detection, r...
Emir Karaosman, Advije Rizvani, Irdin Pekaric
Steering LLMs for Culturally Localized Generation
LLMs are deployed globally, yet produce responses biased towards cultures with abundant training data. Existing cultural localization approaches such as prompting or post-training alignment are bla...
Simran Khanuja, Hongbin Liu, Shujian Zhang, John Lambert, Mingqing Chen, Rajiv Mathews, Lun Wang
Designing Agentic AI-Based Screening for Portfolio Investment
We introduce a new agentic artificial intelligence (AI) platform for portfolio management. Our architecture consists of three layers. First, two large language model (LLM) agents are assigned speci...
Mehmet Caner, Agostino Capponi, Nathan Sun, Jonathan Y. Tan
Pruning for efficient deterministic global optimization over trained ReLU neural networks
Neural networks are increasingly used as surrogates in optimization problems to replace computationally expensive models. However, embedding ReLU neural networks in mathematical programs introduces...
Giacomo Lastrucci, Tanuj Karia, Victor Schulte, Dominik Bongartz, Artur M. Schweidtmann
Granger Causality in Expectiles: an M-vine copula test
A model-free measure of Granger causality in expectiles is proposed, generalizing the traditional mean-based measure to arbitrary positions of the conditional distribution. Expectiles are the only ...
Roberto Fuentes-Martínez, Irene Crimaldi
LLM Olympiad: Why Model Evaluation Needs a Sealed Exam
Benchmarks and leaderboards are how NLP most often communicates progress, but in the LLM era they are increasingly easy to misread. Scores can reflect benchmark-chasing, hidden evaluation choices, ...
Jan Christian Blaise Cruz, Alham Fikri Aji
Unlocking AI's Potential in Agriculture: The Critical Role of Data
India generates substantial volumes of public agricultural data, yet artificial intelligence (AI) adoption in farming remains limited and largely confined to pilot initiatives. This paper examines ...
K. B. Vedamurthy, Manojkumar Patil, Vaishnavi, Priyanka V, Suman L, Ajayakumar, Sagar
Knot-10:A Tightness-Stratified Benchmark for Real-World Knot Classification with Topological Difficulty Analysis
Physical knot classification is a fine-grained visual classification (FGVC) scenario in which appearance cues are deliberately suppressed: different classes share the same rope material, color, and...
Shiheng Nie, Yunguang Yue