safety

WASD: Wasserstein-based Knowledge Distillation for Large Language Models

arXiv:2610.07706v1 Announce Type: cross Abstract: Autoregressive large language models (LLMs) have rapidly advanced in capability, but their increasing scale comes with substantial computational and m

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

We Need Explanation Cards to Connect Explanation Algorithms to the Real World

DGX agent

arXiv:2606.16786v2 Announce Type: replace Abstract: Algorithmic explanations are intended to help stakeholders understand opaque algorithmic decisions, but in practice, they often fall short. First, t

safetyarxiv-cs-lg

7 Oct 2026

What Does Frechet Distance Measure? A Directional Decomposition

arXiv:2610.05518v1 Announce Type: cross Abstract: The Frechet distance is a de facto standard for evaluating generative models across domains, appearing as FID for images and FVD for videos. It summar

What Must Replay Preserve? Separating Correctable Bias from Class Correspondence

arXiv:2610.07077v1 Announce Type: new Abstract: Class-incremental learning must recognize all classes seen so far without task labels. Logit replay methods such as DER and DER++ mitigate forgetting by

What pass@k Cannot Measure: Evaluating Diversity and Capability Retention after Post-Training

arXiv:2610.07405v1 Announce Type: new Abstract: pass@k, the fraction of problems a model solves within k sampled attempts, is the field's default protocol for deciding whether reinforcement-learning (

When better traffic forecasts fail to improve signal control: a layered diagnostic study of forecast-to-decision value

arXiv:2610.06992v1 Announce Type: new Abstract: Improved traffic forecasts do not necessarily yield better signal-control decisions. We investigate this gap through a layered diagnostic study using 29

When Does AI Supervision Help? A Role-Aware Study of Network Fraud Decision Management with Blockchain Auditability

arXiv:2610.07434v1 Announce Type: new Abstract: When does a second artificial intelligence (AI) component improve a primary network-fraud decision rather than add operational burden? We study this que

When Does External Guidance Help LLM Reasoning? A Bias-Variance Theory of Guidance-Augmented GRPO

arXiv:2610.06861v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has become the dominant paradigm for eliciting multi-step reasoning in large language models, an

safetyarxiv-cs-cl

When Plans Change Answers: Formalizing Cost-Accuracy Optimization for Semantic Queries

arXiv:2610.08089v1 Announce Type: cross Abstract: In semantic query engines, predicates are evaluated by machine-learned models, and the choice of a query plan affects not only the cost of a query but

When Tools Lie: Reliability of Mathematical Agents Under Corrupted Tool Feedback

arXiv:2610.08097v1 Announce Type: cross Abstract: Mathematical problem solving often requires deterministic computational steps that agents delegate to tools and implicitly trust. Yet tools can fail s

Where Does the Audio Jailbreak Live? A Controlled Frequency-Depth Audit of AdvWave-P on Qwen2-Audio

arXiv:2610.07005v1 Announce Type: cross Abstract: We audit frequency and decoder-depth claims for AdvWave-P, an additive audio jailbreak, on Qwen2-Audio. The protocol masks frequency components of the

Which Image Property Carries the Jailbreak? A Controlled Dissection of Image-to-Text Jailbreaks

arXiv:2610.07009v1 Announce Type: cross Abstract: Image-to-text jailbreaks place harmful intent in text, image content, or the relationship between them. We examine image-side factors across four publ

Whose Face Is It Anyway? A Multi-Model Audit of Facial Affect Recognition on Children, and Why the Gap Is the Head, Not the Features

arXiv:2610.08279v1 Announce Type: new Abstract: Facial affect models are trained almost entirely on adults, yet are increasingly applied to children in education, health, and developmental research. W

WorldSonus: Bringing Sound to Worlds

arXiv:2610.08760v1 Announce Type: cross Abstract: Recent advances in world models have enabled increasingly realistic visual synthesis. However, these generated environments remain largely silent. Bri

yep, this, too, like a bunch of other similar tweets, is typical of the cluelessness of people who literarily don’t understand the differenc…

yep, this, too, like a bunch of other similar tweets, is typical of the cluelessness of people who literarily don’t understand the difference between a. a neural network solving a problem by itself an

safetygary-marcus--x

A new design of a fall detection system integrating landmark identification and deep learning techniques

arXiv:2610.05749v1 Announce Type: new Abstract: This article introduces an innovative system that integrates landmark identification with deep learning to enhance fall detection accuracy and reliabili

A Safe Action Is Not Enough: Feasible-Future Decoding for Vision-Language-Action Policies

arXiv:2610.05166v1 Announce Type: cross Abstract: A safe action is not necessarily a viable one. Under a frozen vision-language-action (VLA) policy, an action can be likely and locally admissible yet

A Statistical Inference Framework for PMI Estimation and SGNS Word Embeddings

arXiv:2610.05058v1 Announce Type: cross Abstract: Pointwise Mutual Information (PMI) is a core measure of testing word association, and Skip-gram with Negative Sampling (SGNS) is essentially a method

A Unified Dynamics Framework for Reinforcement Learning and Classical Control of a Six-DOF Pipeline-Tracking ROV in NVIDIA Isaac Sim

arXiv:2610.04949v1 Announce Type: new Abstract: Reinforcement learning controllers for underwater vehicles are usually trained against one physics representation and deployed against another, so repor

Adaptive Bregman Alternating Projections for Feasible Gromov-Wasserstein Learning

arXiv:2610.04264v1 Announce Type: new Abstract: The Gromov-Wasserstein (GW) problem compares structured distributions without requiring a shared feature space or known correspondences, but its nonconv

Agentic Resource Allocation for Batch Multi-Objective Bayesian Optimization in Autonomous Materials Discovery

arXiv:2610.04134v1 Announce Type: cross Abstract: The discovery and development of advanced materials is a challenging process constrained by the high time and monetary costs of synthesis, processing,

Agentic-ZTA: A Multi-Agent Architecture for Autonomous Zero Trust Enforcement

arXiv:2610.05782v1 Announce Type: cross Abstract: Agentic AI is emerging as a promising paradigm for automating complex cybersecurity decisions, yet its use in enforcing zero trust introduces signific

AID: A Framework for AI Infrastructure Dynamics

arXiv:2610.04801v1 Announce Type: cross Abstract: A useful model of AI inference infrastructure must specify the system state, the information available to an observer, and the decisions the model is

Aligning Multimodal Patient Evidence with Biomedical Knowledge Graphs for Clinical LLMs

arXiv:2610.06685v1 Announce Type: new Abstract: Clinical questions often depend on linking a patient's multimodal evidence to external biomedical knowledge, yet existing predictive systems rarely repr

Amortized Score-Hamiltonian Policy Iteration: A Grid-Free Scheme for Relaxed Stochastic Control Problems

arXiv:2610.04285v1 Announce Type: cross Abstract: We develop an amortized, grid-free implementation of continuous Langevin dynamics based policy-value iteration for entropy-regularized, infinite-horiz

An evolutionary origin of collective decision making in humans and machines

arXiv:2610.05676v1 Announce Type: cross Abstract: Groups of individuals can solve collective problems more accurately than any single member, by aggregating their opinions. Recent theoretical work has

An LLM-in-the-loop RL Framework for Bioinformatics Feature Selection

arXiv:2610.05600v1 Announce Type: new Abstract: High-dimensional bioinformatics data, characterized by a large number of features relative to the number of samples, pose major challenges such as the `

Anatomy-aware Fine-grained Multimodal Fusion for Laryngopharyngeal Cancer T-Staging Prediction Using CT and Radiology Report

arXiv:2610.06837v1 Announce Type: new Abstract: Accurate T-staging is crucial for guiding personalized treatment strategies for laryngopharyngeal cancer. However, current clinical practice relies on i

Articulatory Entrainment and Coordination Complexity in Spontaneous Autistic and Non-autistic Dialogue

arXiv:2610.04071v1 Announce Type: new Abstract: Articulatory entrainment, the adaptation of vocal tract coordination to facilitate interaction remains underexplored in spontaneous dialogue, particular

ASCENT: Online Test-Time Training of Long-Horizon Agents via Self-Distillation of Verified Experience

arXiv:2610.05303v1 Announce Type: new Abstract: A large language model (LLM) agent solves long-horizon tasks through many reasoning-action turns, with one verification signal at termination. Deployed

Asking the Crowd the Right Question: Bias-Cancelling Weights for Federated Learning

arXiv:2610.04671v1 Announce Type: cross Abstract: A federated objective is a weighted sum of client risks, and the weights are almost always fixed in advance. We treat them instead as the only instrum

At an Australian hearing, OpenAI Chief Strategy Officer Jason Kwon pledged faster safety incident disclosures; Anthropic representatives made similar pledges (ABC)

ABC: At an Australian hearing, OpenAI Chief Strategy Officer Jason Kwon pledged faster safety incident disclosures; Anthropic representatives made similar pledges — By national AI reporter Cam Wilson

safetytechmeme

Backdoor Attacks on Discrete Graph Diffusion Models

arXiv:2503.06340v2 Announce Type: replace-cross Abstract: Diffusion models have demonstrated remarkable generative capabilities in continuous data domains such as images and videos. Recently, discrete

Barrier-Shaped Recurrent Reinforcement Learning for Autonomous Landing on a Heaving Ship Deck

arXiv:2610.03924v1 Announce Type: cross Abstract: This paper addresses autonomous landing of unmanned aerial vehicle (UAV) rotorcrafts on a heaving ship deck using a recurrent policy trained via asymm

Base Models Can Reason By Taking a Cue From Training Data

arXiv:2610.06851v1 Announce Type: new Abstract: In this paper, we study how training data creates associations between the tokens at the start of a base model's response and the reasoning behavior tha

Benchmarking CLIP for Zero-Shot Face and Periocular Gender Estimation

arXiv:2610.06102v1 Announce Type: new Abstract: We investigate CLIP for zero-shot gender estimation from full-face and periocular images. Three CLIP backbones are evaluated on 11,299 frontal images fr

Best-of-N Guidance for Test-time Diffusion Alignment

arXiv:2610.05108v1 Announce Type: new Abstract: Diffusion models achieve strong generative performance but often struggle to align generated samples with human preferences measured by a reward model.

Better Call Reward: Reward Hacking as Strategic Abstention in Legal Reasoning Models

arXiv:2610.06439v1 Announce Type: new Abstract: What happens when a legal AI model learns to look like a lawyer instead of reasoning like one? We fine tune Qwen3-8B with Group Relative Policy Optimisa

Beyond In-Distribution Preservation: Recovering Generalization in Quantized VLAs via Vulnerability-Oriented Tuning

arXiv:2610.05745v1 Announce Type: new Abstract: Post-training quantization has been shown to preserve VLA performance under standard evaluation conditions, but whether it preserves the full-precision

Beyond Token Accuracy: Prioritizing What Matters for Visual Reconstruction

arXiv:2610.03822v1 Announce Type: new Abstract: Tokens have become a unified interface for multimodal foundation models, making visual-token communication a natural paradigm for efficient image delive

Bi-manual Stabilization of the Cervical Spine for Safe Physical Human-Robot Interaction in Rolling Maneuvers

arXiv:2610.04220v1 Announce Type: new Abstract: The neck, specifically the cervical spine, is a highly fragile and mobile part of the body with a high risk of catastrophic injury. Whether someone is i

Bilinear Flow Policy: Distributional Extrapolation for Goal-Conditioned Visuomotor Imitation

arXiv:2610.05765v1 Announce Type: new Abstract: Goal-conditioned imitation learning (GCIL) with flow matching is a promising framework that can represent multimodal behaviors while adapting to diverse

Boosting Transferable Adversarial Attacks against Deep Reinforcement Learning

arXiv:2610.06083v1 Announce Type: new Abstract: Most adversarial attacks on deep reinforcement learning (DRL) assume white-box access to the victim policy, which rarely holds in practice. This paper s

BridgeCast: Bridging Ocean Wave Forecasts to Reanalysis via Flow Matching with Exogenous Variables

arXiv:2610.03759v1 Announce Type: cross Abstract: Ocean wave forecasting is essential for maritime safety, offshore operations, and coastal resilience, yet remains challenging due to systematic biases

Build a voice travel concierge with Amazon Bedrock AgentCore, Managed Knowledge Base and Nova Sonic

Add a voice travel concierge to an airline app with Amazon Bedrock AgentCore, Amazon Nova Sonic for real-time speech, and Amazon Bedrock Knowledge Bases for policy answers. Travelers change seats, che

safetyaws-ml-blog

CACFG: Curvature-Aware Classifier-Free Guidance and Optimal Control

arXiv:2610.05845v1 Announce Type: new Abstract: Diffusion models generate samples by learning to reverse a fixed corruption process, and classifier-free guidance (CFG) is the standard mechanism for co

Causally Fair Generation with Large Language Models

arXiv:2610.04444v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used to generate, complete, and transform information in settings where their outputs can shape conseque

CEENs: Causality-enforced evolutional networks for solving time-dependent partial differential equations

arXiv:2610.04405v1 Announce Type: new Abstract: Despite the growing popularity of physics-informed neural networks (PINNs), their applicability in the long-time integration of partial differential equ