The paper introduces a source‑grounded integrity gate for AI‑assisted personal health records, ensuring that data generated by large language models remains provisional until a deterministic monitor verifies it against the source document. The monitor only accepts candidates that contain a unique supporting quotation, appear within the same laboratory row, and preserve provenance, preventing the model from approving its own output. In Medical DataCloud, the system passed all 22 conformance and mutation tests and, in a replay of nine historical lab reports, admitted 72 of 97 numeric candidates while retaining 25 for human review.
By Nora Girda, Adrian Groza
This study evaluates machine learning and deep‑learning models for classifying balanced versus imbalanced postural states in immersive virtual reality using a multimodal dataset of kinematic, EMG, and EDA signals. The Mamba‑inspired CNN (MI‑CNN) achieved the highest accuracy (96.76%) and, through SHapley Additive exPlanations (SHAP), identified kinematic features as the most influential for detecting imbalance. Even after reducing input dimensionality by 33% based on SHAP importance, the model maintained near‑optimal performance (0.957 accuracy and F1‑score).
By Nipa Anjum, Md Irfan Pavel, Robert Gonzalez Jr, Kevin Desai, Alberto Cordova, M. Rasel Mahmud, John Quarles
RAGDiffusion++ advances garment generation by addressing the high‑frequency texture gap that previous retrieval‑augmented models left unresolved. The approach introduces a dual‑image FLUX architecture trained on a large, complex garment dataset, coupled with a new attribute‑aware reward model that guides reinforcement learning to favor realistic high‑frequency patterns. An adversarial‑regularized RL strategy (AR‑GRPO) further prevents artifact exploitation, ensuring the model samples authentic, detailed garment textures.
By Yuhan Li, Xianfeng Tan, Fangao Zeng, Wenxiang Shang, Pipei Huang, Hao Zhou, Zhiyu Jin, Wenjun Zhang, Bingbing Ni
The paper introduces a fully distributed continuous‑time algorithm for solving Generalized Nash Equilibrium Problems (GNEPs) with shared linear equality constraints. Unlike existing methods that require exchanging Lagrange multipliers, this approach converges to any GNE without multiplier communication, thereby reducing communication overhead and enhancing privacy. Discrete‑time variants are also presented and the method is demonstrated on a multi‑robot placement task.
By Shao-An Yin, Mingyi Hong, Nicola Elia
The paper introduces a rubric-based benchmark to evaluate Saudi Arabic dialect and cultural competence in large language models. It comprises 31 expert-authored prompts covering idiomatic, pragmatic, lexical, and culturally embedded aspects, each paired with an expert-established ground truth. Four state-of-the-art models were scored, revealing that none exceeded 55% accuracy and that ambiguous framing was the most common error type.
By Ghassan Al-Sumaidaee, Sajjad Abdoli, Ahmed Rashad, Maxim Legg
BCPPO is a new variant of Proximal Policy Optimization that uses Bachelier-inspired cost‑prediction networks to generate a smooth penalty based on disagreement among critics. The method keeps temporal‑difference learning unchanged, applies a saturation‑aware controller to manage cost penalties, and deploys only the policy network. Across extensive experiments, BCPPO outperforms comparators in achieving higher mean returns while maintaining lower or comparable CVaR in all tested tasks.
By Dongsheng Hou, Yanqiao Chen, Yuhan Rui
This survey reviews tensor methods applied to large language models, framing them through a seven‑stage lifecycle (tokenization, embeddings, pre‑training, adaptation, compression, inference, interpretability) and a component view (embeddings, attention, feed‑forward networks). It offers unified notation, theoretical foundations, and comparative analyses of tensorization strategies for Transformer components, while highlighting evaluation protocol differences and model scale effects. The paper also introduces a new metric, ρ_gap, to quantify the gap between theoretical memory savings and actual system‑level speedup, and connects tensor techniques to related efficiency and probabilistic methods.
By Matvei Tarasov, Salman Ahmadi-Asl, Andre L. F. de Almeida, Andrzej Cichocki
The paper introduces a German-English benchmark dataset to evaluate anti‑LGBTQ biases in language models, combining community‑sourced stereotypes from German‑speaking queer individuals with a German translation of WinoQueer. Eight language models of varying sizes and architectures were assessed, revealing that they reproduce anti‑queer stereotypes with differences across identities and models. Fine‑tuning on community and progressive media content reduced bias on average, though the effect was not consistent across all models and identities.
By Melina Morch, Daniel Braun
The paper introduces a curiosity‑driven quantized Mixture‑of‑Experts framework that routes inputs based on Bayesian epistemic uncertainty across heterogeneous experts (BitNet ternary, 1‑16 bit BitLinear, post‑training quantization). On audio classification benchmarks, 4‑bit quantization preserves 99.9 % of full‑precision F1 while achieving 4× compression and 31 % energy savings, and curiosity‑driven routing further improves accuracy and reduces cross‑fold variance by up to 85 %. The routing is self‑organizing, allocating the most uncertain samples to the high‑precision expert, and the method demonstrates statistical parity with full precision across datasets.
By Sebasti\'an Andr\'es Cajas Ord\'o\~nez, Luis Fernando Torres Torres, Mackenzie J. Meni, Carlos Andr\'es Duran Paredes, Eric Arazo, Cristian Bosch, Ricardo Simon Carbajo, Yuan Lai, Leo Anthony Celi
The paper investigates the nature of agreement among repeated samples of large language models (LLMs), showing that strong agreement can arise even for incorrect answers. It introduces a pluralistic agreement index, Gamma, which is decomposed into a mechanical component driven solely by per‑case answer preferences and a residual component that captures preference‑unexplained agreement. Experiments on GPT‑4.1 and several open‑weight models demonstrate that mechanical agreement dominates in many settings, while the residual varies with benchmark type and sampling protocol.
By Lizhuo Zhang, Mengmeng Tang, Chenfeng Long, Xiaoyong Tang, Xiang Luo
The paper introduces an unsupervised framework that merges manifold learning with rank‑based interpretable graph embeddings to address the Geometric and Interpretability Gaps in visual representation learning. By first analyzing contextual information on the dataset manifold and then producing sparse, self‑explainable embeddings, the method achieves dimensionality reduction while preserving or improving performance in image retrieval and semi‑supervised Graph Convolutional Network classification. Experiments across varied datasets confirm that these context‑aware representations maintain high downstream effectiveness.
By Thiago C\'esar Castilho Almeida, Gustavo Rosseto Let\'icio, Vinicius Atsushi Sato Kawai, Daniel Carlos Guimar\~aes Pedronette
arXiv:2608. 29503v1 Announce Type: new Abstract: Worst-case online classification is governed by sequential complexity, such as Littlestone dimension, and can be impossible even for statistically simple classes, such as thresholds of VC dimension one.
By Roi Livni, Sahil Singla
The paper investigates how making responsible‑AI evaluations more efficient—through batching, quantization, and benchmark reduction—affects the stability of conclusions drawn about model behavior. By testing three dense and mixture‑of‑experts models on the BBQ and BBQ‑V datasets under seven different conditions, the authors compare accuracy, bias, reasoning quality, subgroup performance, subset‑membership stability, runtime, and GPU energy consumption against a full‑benchmark BF16 baseline. Findings show that larger batching preserves accuracy and reduces energy in most settings, INT8 largely maintains quality but can increase energy use, INT4 introduces larger, context‑dependent changes, and reduced benchmarks save resources but are highly sensitive to which items are retained, underscoring that efficient evaluation must be validated against the benchmark’s intended conclusions.
By Ahmed El Kady, Aravind Narayanan, Rehana Noorani, Yani Ioannou, Shaina Raza
The study challenges the prevailing belief that frequency-based features and compression-invariant learning are essential for robust deepfake detection. Using a controlled, pre‑registered protocol, a simple EfficientNet‑B0 trained on diverse multi‑quality data outperformed the more complex CAFRL model across all compression levels, with a 3.66 AUC point advantage at CRF 40. After identifying and correcting four experimental defects, the authors found that frequency features added no marginal benefit, while data diversity—particularly real constant‑rate‑factor variants—proved to be the key factor for robustness against H.264 re‑encoding.
By Abbas Aliyev, Samir Rustamov
ECGQuest is a new benchmark that evaluates language models on the contextual knowledge required for electrocardiogram interpretation, featuring 10,904 True/False questions derived from 23 ECG references and 2003‑2025 Computing in Cardiology proceedings. The study tested 23 commercial and open‑source models, finding that zero‑shot accuracy ranged from 49.5% to 74.4% and that fine‑tuning with Low‑Rank Adaptation improved all open‑source models by 6.5–14.1%, with the best fine‑tuned model achieving 76.3% accuracy and a five‑model ensemble reaching 78.5%. ECGQuest demonstrates that parameter‑efficient fine‑tuning can enable smaller models to compete with larger commercial ones on ECG‑specific tasks.
By Mohammadsina Hassannia, Matthew A. Reyna, Reza Sameni
The paper investigates multilingual confidence calibration in large language models, revealing that non‑English languages are systematically less well calibrated than English. By analyzing internal representations, the authors find that late‑intermediate layers provide a more reliable confidence signal than the final layer, which is biased by English‑centric training. They propose training‑free methods such as Language‑Aware Confidence Ensemble (LACE) to adaptively select optimal layers per language, aiming to improve global equity and trustworthiness of LLMs.
By Ej Zhou, Caiqi Zhang, Tiancheng Hu, Chengzu Li, Nigel Collier, Ivan Vuli\'c, Anna Korhonen
The paper introduces ClinMPO, an evidence‑guided reinforcement‑learning framework that enhances psychiatric reasoning in small language models (SLMs). ClinMPO leverages a reward model (ClinRM) trained on 18,569 question–answer pairs from 4,474 psychiatry articles and is guided by the psychiatrist‑defined Clinical Psychiatry Thinking Strategy (CPTS). Evaluations on four Qwen3 model sizes show that ClinMPO outperforms baseline, supervised fine‑tuning, and standard policy optimization, with the 8B model surpassing a human baseline of senior pre‑licensure medical students and achieving the highest rank among 31 models. Blinded clinician assessment confirms improved rationale quality across CPTS criteria, demonstrating that clinical evidence and specialist knowledge can be effectively incorporated into medical AI development.
By Xinxin Lin, Guangxin Dai, Yi Zhong, Xiang Li, Xue Xiao, Jian Liu, Yixin Zhang, Lingming Hu, Zhengdong Wu, Yongbo Zheng, Runchuan Zhu, Ming Zhao, Huizi Yu, Yi Zhang, Fangting Lu, Shuo Wu, Jun Zhao, Ping Yin, Joey W. Y. Chan, Ngan Yin Chan, Yumei Wang, Lejin Yang, Yanqiu Xing, Sijing Chen, Yun Kwok Wing, Lin Lu, Xin Ma, Lizhou Fan
FairReL is a fairness‑aware representation‑learning framework for deepfake detection that targets two subgroup‑sensitive components: multi‑scale spatial features and fine‑tuning‑induced residual features. It uses an SVD‑decomposed backbone to isolate residuals and introduces Group‑Conditional Wavelet Decorrelation (GCWD) and Subspace‑Localised Mean Alignment (SLMA) losses to suppress subgroup imbalance and align subgroup means. Experiments on FF++, Celeb‑DF, DFD, and DFDC show that FairReL improves unseen‑dataset AUC by 3.9% and reduces subgroup FPR disparity by 10.2% compared to the state‑of‑the‑art fairness‑aware detector.
By Xiaoman Lu, Jiaqi Li, Shuntian Zheng, Huiping Chen, Yu Guan
The paper investigates continual machine unlearning, where models must forget data over time. It identifies a fundamental issue called plasticity collapse, where successive unlearning requests cause geometric constraints that saturate parameter space, leading to two failure modes: forward failure (reduced forgetting quality) and backward failure (re‑memorization). Experiments across architectures and datasets confirm that plasticity collapse is a pervasive problem in continual unlearning.
By Yingdan Shi, Xiang Xu, Kaize Ding, Alfred O. Hero, Ren Wang
AgenticRag‑R1 is a reinforcement‑learning framework that integrates reasoning, retrieval, and memory through a stack and fine‑grained action space. It uses hierarchical action‑aware rewards and an information‑aware trajectory rejection strategy to support long‑horizon learning. Experiments on multi‑hop, open‑domain, and agentic reasoning benchmarks show that AgenticRag‑R1 outperforms strong baselines and produces robust, interpretable, memory‑aware reasoning behaviors.
By Xinke Jiang, Yue Fang, Zhibang Yang, Jiaran Gao, Zhixin Zhang, Tao Feng, Rihong Qiu, Wentao Zhang, Hongxin Ding, Ruizhe Zhang, Yongxin Xu, Yuheng Huang, Xu Chu, Junfeng Zhao, Yasha Wang