safety

Rationalize: Shared Semantic Reasoning for Human-AI Alignment

arXiv:2605.30632v1 Announce Type: cross Abstract: We introduce Rationalize, a role-pair framework for shared semantic reasoning between humans and AI models in data-driven sensemaking. Building on ide

RDGen: Demonstration Generation for High-Quality Robot Learning via Reinforcement Learning

DGX agent

arXiv:2605.30957v1 Announce Type: new Abstract: Vision-Language-Action (VLA) models have emerged as a promising paradigm for general-purpose robot control. However, their performance remains fundament

safetyarxiv-cs-ro

1 Jun 2026

Reading Between the Citations: A Typed Claim Network for Scientific Literature

arXiv:2605.30966v1 Announce Type: cross Abstract: Knowledge graphs over corpora of inter-referencing documents - scholarly papers, legal opinions, policy briefs - encode the topology of reference but

REAL: Regression-Aware Reinforcement Learning for LLM-as-a-Judge

arXiv:2603.17145v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly deployed as automated evaluators that assign numeric scores to model outputs, a paradigm known a

Reassessing Extractive QA Datasets at Scale: LLM-as-a-Judge and In-Depth Analyses

arXiv:2504.11972v3 Announce Type: replace Abstract: Extractive QA tasks are commonly evaluated using Exact Match (EM) and F1-score, but these metrics often fail to reflect true model performance. Rece

Reinforced sequential Monte Carlo for amortised sampling

arXiv:2510.11711v2 Announce Type: replace Abstract: This paper proposes a synergy of amortised and particle-based methods for sampling from distributions defined by unnormalised density functions. We

Reinforcement Learning Amplifies Emergent Misalignment from Harmless Rewards

arXiv:2605.31328v1 Announce Type: new Abstract: Emergent misalignment (EM) is the surprising tendency of language models to become broadly misaligned after fine-tuning on narrowly misaligned examples.

Reinterpreting Safety Thresholds as Neuron Spiking Thresholds

arXiv:2605.30368v1 Announce Type: cross Abstract: Surrogate Safety Measures (SSMs) are extensively utilised in the evaluation of traffic risk in automated driving contexts. However, the majority of SS

Representation Collapse in Sequential Post-Training of Large Language Models

arXiv:2605.30524v1 Announce Type: new Abstract: Large language models are now adapted through chains of post-training stages rather than through a single instruction-tuning pass. This paper studies wh

Rethinking Multimodal Few-Shot 3D Point Cloud Segmentation: From Fused Refinement to Decoupled Arbitration

arXiv:2601.01456v2 Announce Type: replace-cross Abstract: In this paper, we revisit multimodal few-shot 3D point cloud semantic segmentation (FS-PCS), identifying a conflict in 'Fuse-then-Refine' para

// Reusable Context Engineering // Context bloat quietly kills long-horizon runs, but you can fix it from the outside without fine-tuning th…

// Reusable Context Engineering // Context bloat quietly kills long-horizon runs, but you can fix it from the outside without fine-tuning the underlying agent. (bookmark this) Context management is us

safetydair-ai--x

Revisiting Zeroth-Order Hessian Approximation: A Single-Step Policy Optimization Lens

arXiv:2605.30960v1 Announce Type: new Abstract: Accurate Zeroth-Order (ZO) Hessian estimation is a cornerstone of derivative-free methods, essential for tasks such as bilevel optimization, Bayesian in

Routing on the Stiefel Manifold: When Does Adaptive Subspace Selection Help for Cross-Domain EEG Decoding?

arXiv:2605.31043v1 Announce Type: cross Abstract: Cross-domain EEG decoding remains challenging despite advances in Riemannian deep learning: covariance matrices from different subjects occupy systema

Safeguarding Text-to-Image Generation via Inference-Time Prompt-Noise Optimization

arXiv:2412.03876v2 Announce Type: replace Abstract: Text-to-Image (T2I) diffusion models are widely recognized for their ability to generate high-quality and diverse images based on text prompts. Howe

Scalable Constrained Multi-Agent Reinforcement Learning via State Augmentation and Consensus for Separable Dynamics

arXiv:2605.30461v1 Announce Type: cross Abstract: We present a distributed approach for constrained Multi-Agent Reinforcement Learning (MARL) that combines state-augmented policy learning with distrib

Scaling Multi-Agent Environment Co-Design with Diffusion Models

arXiv:2511.03100v2 Announce Type: replace-cross Abstract: The agent-environment co-design paradigm jointly optimises agent policies and environment configurations in search of improved system performa

SCOPE: Selective Conformal Optimized Pairwise LLM Judging

arXiv:2602.13110v3 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used as scalable judges in pairwise evaluation, but they remain prone to miscalibration and bias

SDM-Q: Cost-Aware Staged Decision-Making for Multi-Omics Classification with Deep Q-Learning

arXiv:2605.31014v1 Announce Type: new Abstract: Multi-omics data provide complementary molecular characterizations of disease phenotypes and play an important role in disease diagnosis and subtype cla

Secure AI agents with Policy and Lambda interceptors in Amazon Bedrock AgentCore gateway

In this post, we use a lakehouse data agent to demonstrate how you can use Policy for deterministic access control and Lambda interceptors for dynamic validation. We then show how to combine Lambda in

safetyaws-ml-blog

Seeing Before Agreeing: Aligning Multi-Agent Consensus with Visual Evidence

arXiv:2605.30698v1 Announce Type: cross Abstract: Vision-language models (VLMs) have achieved strong performance on visual question answering (VQA). To mitigate individual hallucinations and blind spo

Semantic Motion Anchors: Bridging Motion and Meaning in Co-Speech Gestures

arXiv:2605.30608v1 Announce Type: new Abstract: Learning a shared representation between spoken text and gesture is central to co-speech gesture retrieval, synthesis, and understanding, but remains ch

SemStruct: Contextualizing Semantic Embeddings with Structural Information for Schema Matching

arXiv:2605.30729v1 Announce Type: new Abstract: Schema matching is a fundamental step in integrating heterogeneous data sources. While Pre-trained Language Models (PLMs) have revolutionized this task

Simulation of collision avoidance behavior in crowd movement by data-driven approach

arXiv:2605.31210v1 Announce Type: cross Abstract: Crowd movement simulation is essential for pedestrian safety management and facility layout optimization. Data-driven models enhance trajectory predic

Skill is Not One-Size-Fits-All: Model-Aware Skill Alignment for LLM Agents

arXiv:2605.30723v1 Announce Type: new Abstract: LLM agents increasingly retrieve externally curated skills-procedural instructions retrieved at decision time-to improve performance on long-horizon int

Smaller Models are Natural Explorers for Policy-Level Diversity in GRPO

arXiv:2605.30789v1 Announce Type: cross Abstract: We identify a new dimension for enhancing rollout diversity in Group Relative Policy Optimization (GRPO) for LLMs. While GRPO relies on diverse rollou

Softly Constrained Denoisers for Diffusion Models Applied to Partial Differential Equations

arXiv:2512.14980v4 Announce Type: replace Abstract: Diffusion models have become a powerful generative prior for solutions of partial differential equations (PDEs). Existing approaches enforce physica

SpongeBob: Sync-Aware Harmonious Audio-Visual Generative Editing

arXiv:2605.25193v2 Announce Type: replace Abstract: Visual and acoustic events in the physical world are inherently coupled, yet existing video editing methods typically adopt decoupled pipelines, lac

Stateful Online Monitoring Catches Distributed Agent Attacks

arXiv:2605.31593v1 Announce Type: cross Abstract: Language models can find thousands of severe software vulnerabilities, and agents are increasingly being misused for cyberattacks. To avoid detection,

Structural Bias Beyond Homophily: A Study of Fairness in Link Prediction

arXiv:2602.11802v2 Announce Type: replace Abstract: Graph link prediction (LP) plays a critical role in socially impactful applications such as job recommendation and friendship formation, making fair

Structure-Induced Information for Rerooting Levin Tree Search

arXiv:2605.30664v1 Announce Type: new Abstract: Subgoal-based policy tree search, which uses a policy to guide search, is effective for complex single-agent deterministic problems but often relies on

Supervised Learning as Lossy Compression: Characterizing Generalization and Sample Complexity via Finite Blocklength Analysis

arXiv:2602.04107v2 Announce Type: replace Abstract: This paper presents a novel information-theoretic perspective on generalization in machine learning by framing the learning problem within the conte

Supervised Training Rapidly Degrades Early Visual Cortex Alignment Across Biologically Plausible Learning Rules

arXiv:2605.30556v1 Announce Type: new Abstract: Random, untrained neural networks consistently match or exceed trained networks in representational similarity to early visual cortex. This puzzling fin

Surface Constraint Policy for Learning Surface-Constrained and Dynamically Feasible Robot Skills

arXiv:2605.31321v1 Announce Type: new Abstract: Diffusion-based imitation learning methods have driven rapid progress in robot dexterous manipulation tasks. However, they have limitations when applied

safetyarxiv-cs-ro

Synthetic Stimuli, Real Gains: Rethinking VLM Fine-Tuning Through Fully Controlled Data Generation

arXiv:2511.11440v3 Announce Type: replace-cross Abstract: Performance gains of Vision Language Models (VLMs) obtained by fine-tuning are generally based on ad hoc data collection and annotation of rea

TALON: Token-Aligned Lightweight Adapters for 6-DoF Spacecraft Pose Estimation

arXiv:2605.31217v1 Announce Type: new Abstract: Monocular 6-DoF spacecraft pose estimation methods predominantly process individual frames, discarding the temporal information present in an image sequ

TARIC: Memory-Augmented Traversability-Aware Outdoor VLN under Interrupted Semantic Cues

arXiv:2605.31121v1 Announce Type: cross Abstract: Outdoor vision-language navigation (VLN) in long-range, open-world environments is frequently disrupted by semantic-cue interruptions, where informati

Task-Focused Memorization for Multimodal Agents

arXiv:2605.31075v1 Announce Type: new Abstract: Long-term memory is essential for multimodal agents to build coherent experience, accumulate world knowledge, and achieve continual learning. However, c

THE DEFINITIVE AI CONVERSATION OF THE YEAR. MUST LISTEN. @JG_Nuke @GaryMarcus @MacrostrategyP https://open.substack.com/pub/georgenoble/p/ai…

THE DEFINITIVE AI CONVERSATION OF THE YEAR. MUST LISTEN. @JG_Nuke @GaryMarcus @MacrostrategyP https://open.substack.com/pub/georgenoble/p/ai-the-biggest-capital-misallocation-d18?r=35saq&utm_medium=io

safetygary-marcus--x

The Flip Side of RLHF: On-Policy Feedback for Reward Model Self-Supervised Improvement

arXiv:2605.30888v1 Announce Type: new Abstract: Building strong reward models (RMs) for language model alignment is bottlenecked by the cost and difficulty of acquiring diverse and reliable preference

The Global Landscape of Environmental AI Regulation: From the Cost of Reasoning to a Right to Green AI

arXiv:2603.00068v2 Announce Type: replace-cross Abstract: Artificial intelligence (AI) systems impose substantial and growing environmental costs, yet transparency about these impacts has declined eve

The OpenAI Foundation is doing a lot of wonderful things. Helping society become resilient to AI is going to be incredibly important. Much m…

The OpenAI Foundation is doing a lot of wonderful things. Helping society become resilient to AI is going to be incredibly important. Much more to come here! AI is advancing quickly. Society’s ability

safetysam-altman--x

The Refutability Gap: Challenges in Validating Reasoning by Large Language Models

arXiv:2601.02380v4 Announce Type: replace-cross Abstract: Recent reports claim that Large Language Models (LLMs) have achieved the ability to derive new science and exhibit human-level general intelli

The Sword, Shield, and Achilles' Heel: Characterizing the Linguistic Inductive Bias of Large Language Models for Spatial Reasoning in Navigation Planning

arXiv:2605.31404v1 Announce Type: cross Abstract: Large Language Model (LLM)-based navigation systems commonly construct explicit spatial representations (e.g., topological graphs, semantic raster map

“The technology worked. The value didn’t arrive,” Bain concluded in the report. https://www.bloomberg.com/news/newsletters/2026-06-01/bain-s…

“The technology worked. The value didn’t arrive,” Bain concluded in the report. https://www.bloomberg.com/news/newsletters/2026-06-01/bain-survey-ai-delivers-less-cost-reduction-than-many-firms-predic

safetygary-marcus--x

This was right five years ago, and still is: “Large scale pretrained models are certainly likely to figure prominently in artificial intelli…

This was right five years ago, and still is: “Large scale pretrained models are certainly likely to figure prominently in artificial intelligence for the near future, and play an important role in com

safetygary-marcus--x

Traceable by Design: An LLM Pipeline and Dashboard for EU Regulatory Consultation Analysis

arXiv:2605.30995v1 Announce Type: cross Abstract: Public consultations generate large volumes of data in the form of stakeholder submissions that are practically unfeasible to analyse manually. We pre

Trust-Region Behavior Blending for On-Policy Distillation

arXiv:2605.31159v1 Announce Type: cross Abstract: On-policy distillation (OPD) trains a student on prefixes sampled from its own policy while matching a stronger teacher. This addresses the prefix mis

TunerDiT: Training-free Progressive Steering of Diffusion Transformer for Multi-Event Video Generation

arXiv:2605.31590v1 Announce Type: cross Abstract: Text-to-video (T2V) generation faces challenging questions when generating videos with long horizons containing multiple events. Inspired by the intri