The paper investigates the implicit bias of Riemannian gradient flow in hyperbolic multiclass classification with fixed class prototypes. It introduces a decomposition of distances into radial and direction-dependent terms via the Busemann function, leading to two key results: a radial dichotomy that predicts whether trajectories move toward the boundary or return inward, and convergence of boundary directions to critical points of the Busemann risk. These findings offer a rigorous asymptotic explanation for boundary saturation and near-boundary clustering observed in hyperbolic representation learning.
By Xingrun Li, Sho Kuno, Yusuke Mukuta, Xin Yang, Tatsuya Harada
arXiv:2610. 07529v1 Announce Type: new Abstract: Wave-based processors promise fast, energy-efficient Fourier layers for neural operators.
By Samrendra Roy, Jason Yoo, Souvik Chakraborty, Syed Bahauddin Alam
The paper introduces Contrastive‑SDXL, a day‑to‑night augmentation framework that uses SDXL‑Turbo and Low‑Rank Adaptation to generate realistic night‑time images for pedestrian detection. It employs a patch‑wise semantic contrastive loss guided by a pretrained DINOv2 encoder and an object consistency loss to preserve pedestrian and local semantic structure during translation, achieving a FID of 22.5. Detectors trained on these synthetic images show a 6‑7% reduction in miss rate compared to a daytime‑only baseline, approaching the performance of models trained on real night‑time data.
By Franky George, Muhammad Khalid, Adil Khan, Koorosh Aslansefat
The paper examines four leading motion‑based AI‑generated video detectors and finds that three of them suffer from preprocessing and sampling biases that inflate their reported performance. These detectors rely heavily on motion patterns—specifically, the lower inter‑frame movement typical of synthetic videos—so their accuracy drops to near random when tested on datasets lacking this bias or after simple spatial augmentations. In contrast, a frequency‑based detector remains robust across all datasets, indicating that frequency‑domain methods may generalize better for detecting AI‑generated videos.
By Joren Michels, Lode Jorissen, Nick Michiels
The paper introduces the Residual Patch Adapter (RPA), a lightweight module that aligns EEG and MEG signals with all patch tokens from an intermediate layer of a Vision Transformer, rather than a single global embedding. Ablation studies show that retaining the full set of patch tokens is crucial for effective alignment, while the CLS token adds little value. Using RPA, the authors achieve state‑of‑the‑art Top‑1 accuracies on THINGS‑EEG2 and THINGS‑MEG datasets, and demonstrate that the approach generalizes to other brain encoders and can be swapped with various neural network backbones.
By Yuhui Jin, Yonghao Song, Bingchuan Liu
SymNetPro is a localization framework that extends SymNet by adding a line‑of‑sight aware attention bias and a transmitter‑drop augmentation technique. These components help the model learn obstruction‑aware spatial relations and handle varying numbers of transmitters. Experiments on ray‑traced urban environments demonstrate lower OSPA errors compared to baseline methods, especially under sparse sampling, noise, and higher transmitter counts.
By Lyuzhou Ye, Heng Fan, Yan Huang
The paper addresses the problem of setting optimal non‑discriminatory prices in a network of strategic agents with private information and potential misreporting. It shows that the optimal prices can be decomposed according to each agent’s squared centrality in a preference‑weighted network, leading to the Poll algorithm. Poll samples one agent per round, briefly explores its neighborhood, and updates prices locally, achieving significant computational, statistical, and economic efficiency gains over exact or other distributed methods.
By Chenyu Zhang, Rohit Parasnis, Saurabh Amin
The paper investigates how to recover low‑dimensional latent geometry from a random geometric graph built from noisy, high‑dimensional data. It analyzes a spectral embedding algorithm on the Signal+Noise Graph Model, where vertices are perturbed points and edges are added when their inner product exceeds a threshold. In the regime where both the number of points and ambient dimension grow, the authors prove that, under a spectral gap condition, the leading eigenvectors and eigenvalues of the adjacency matrix can approximately recover the original point cloud up to an orthogonal transformation, and demonstrate the method on nested spheres and high‑dimensional sinusoid curves.
By Tatiana Brailovskaya, Nicholas A. Cook, Sofia Poinelli
SIGMA is a self‑improving alignment framework that uses a model’s own reasoning to generate diverse alignment dilemma scenarios from a simple "Model Spec" and then fine‑tunes the model with self‑judged rewards. By performing spec‑guided task synthesis and rubric‑based reinforcement learning, SIGMA improves safety alignment in multi‑turn agentic settings, reducing harmfulness and misalignment scores while preserving general capability. The approach demonstrates that a model can autonomously strengthen its own safety reasoning without external supervision.
By Jingyu Zhang, Shruti Palaskar, Daniel Khashabi, Benjamin Van Durme, Leon A. Gatys, Joseph Yitan Cheng
BluffJAX is an open‑source suite of adversarial imperfect‑information games implemented in JAX, designed for high‑throughput simulation and GPU parallelization. It includes classic benchmarks like Texas Hold’Em and Kuhn Poker, as well as newer games such as Bluff, Stud Poker, and Kemps, aiming to broaden research in game‑theoretic reinforcement learning. The authors benchmarked throughput, memory usage, and algorithmic performance across single and multi‑GPU setups, showing scalability to hundreds of millions of samples per second and providing baseline results for future work.
By Aryaman Reddi, Jan Peters, Carlo D'Eramo
POLAR is a guardrail framework designed for small tool‑calling language‑model agents that evaluates the reversibility of each action using a two‑layer ontology. By assigning graded reversibility scores and pruning calls that fall below a threshold, POLAR aims to prevent operational risks before they occur. In experiments on the τ²‑bench across six agent models, POLAR increased mean task reward by 0.11 to 0.18 points on airline tasks for four of six agents, though improvements were limited to eight of eighteen model‑domain cells, with some domains showing regression.
By Yunju Kang, Seonghyeon Cho, Irene Li, Yeo-Chan Yoon, Chanjun Park
The paper "Illusory Pattern Perception Drives Spurious Inference in Large Language Models" reports that large language models (LLMs) exhibit a cognitive bias similar to humans’ tendency to see patterns where none exist. By adapting classic psychological experiments to three tasks, the authors find that LLMs often over‑associate frequent positive attributes with majority groups or large organizations and are prone to constructing causal narratives from ambiguous events. Using a Sparse Autoencoder‑based interpretability framework, the study links these illusory perceptions to holistic frequency perception and analytic cognitive orientation within the models.
By Peihua Mai, Zhuoyan Shao, Xinbao Qiao, Meng Zhang, Xinyue Zhou, Yan Pang
The paper introduces STEPGATE, an uncertainty‑aware handoff framework that evaluates each step of a small language model (SLM) agent and selectively escalates difficult steps to a stronger model. On a 52‑task single‑step benchmark, the Qwen2.5‑1.5B/7B pair achieved 82.7% task success with only 30.8% escalation, outperforming local‑only and random escalation baselines. In multi‑turn tests, STEPGATE reached 69.0% trajectory success and 84.0% action success while using only 30.0% cloud actions, demonstrating that step‑level escalation can close much of the performance gap to a stronger backend with fewer remote tokens.
By Abolfazl Younesi
The paper introduces SSRFT (Supervised Safe-Role Fine‑Tuning), a new framework that reframes safety alignment for large language models as internalizing a predefined safe role rather than relying on explicit refusal patterns. SSRFT builds a Safe‑Role Question‑Answer dataset from psychometric questions, limited jailbreak prompts, and a safe‑role description, then synthesizes and validates role‑consistent responses to expose models to diverse safety scenarios. Experiments on multiple base and instruct models demonstrate that SSRFT yields more robust and generalizable safety alignment, better resilience to prefilling attacks, improved generalization to unseen jailbreak domains, and reduced over‑refusal on benign queries while preserving general capabilities.
By Jinghao Pang, Jitai Hao, Qiang Huang, Zhaochun Ren, Jun Yu
arXiv:2610.07514v1 Announce Type: new
Abstract: A vision-language model may need to combine an image with a prompt to recognize a safety risk that neither reveals alone. Where does this joint safety...
By Faezeh Dehghan Tarzjani, Mevan Wijewardena, Alexander Romanus, Sampad Mohanty
arXiv:2610.03829v1 Announce Type: cross
Abstract: Real-world outpatient oncology notes contain specialised terminology, tumour staging expressions, treatment names, toxicity descriptions, and institu...
By Wuraola Oyewusi, Eliana Vasquez Osorio, Goran Nenadic, Gareth Price
arXiv:2610.06885v1 Announce Type: cross
Abstract: Moving target defense (MTD) against advanced persistent threats (APTs) in industrial control systems (ICS) has well-established game-theoretic formul...
By Tian Zijian, Zhang He, Chen XinJie, Liu Xinggao
arXiv:2610.07125v1 Announce Type: cross
Abstract: While open-weight models have enjoyed steady progress in capabilities and wide adoption across multiple domains, their safety remains an important co...
By Abhinav Sudhakar Dubey (University of California Santa Cruz), Scott Sirri (University of California Santa Cruz), Vaggos Chatziafratis (University of California Santa Cruz), C. Seshadhri (University of California Santa Cruz)
arXiv:2610.07204v1 Announce Type: cross
Abstract: Recent AI alignment work often frames alignment as a pre-deployment optimization problem: collect human feedback, learn preferences or principles, fi...
By Tao Long, Lydia B. Chilton
arXiv:2610.07205v1 Announce Type: cross
Abstract: Institutional Analytics (IA) dashboards inform decision-making in higher education, yet data limitations, constraints in analytical techniques, and m...
By Francielle Marques, Ariel Ortiz-Beltr\'an, Ishari Amarasinghe, Davinia Hern\'andez-Leo