safety

Pretty profound for an ipo filing: “For the entirety of its existence, human civilization has lived on a single celestial body: Earth. The c…

Pretty profound for an ipo filing: “For the entirety of its existence, human civilization has lived on a single celestial body: Earth. The current paradigm, in which human civilization is confined to

safetyelon-musk--x

Safety

Principled RL for Flow Matching Emerges from the Chunk-level Policy Optimization

DGX agent

arXiv:2510.21583v2 Announce Type: replace Abstract: Recent Progress in post-training flow matching for text-to-image (T2I) generation with Group Relative Policy Optimization (GRPO) has demonstrated st

safetyarxiv-cs-cv

21 May 2026

Proximal State Nudging: Reducing Skill Atrophy from AI Assistance

arXiv:2605.20355v1 Announce Type: cross Abstract: Skill atrophy, the gradual decline of human capability under AI assistance, poses a safety risk in shared-control of semi-autonomous systems, where op

Q-SpiRL: Quantum Spiking Reinforcement Learning for Adaptive Robot Navigation

arXiv:2605.20801v1 Announce Type: new Abstract: Adaptive robot navigation in dynamic environments requires policies that can reach the target reliably while producing efficient and stable trajectories

Quadratic Characterizations for Reachability Analysis of Neural Networks

arXiv:2605.20482v1 Announce Type: new Abstract: Quadratic constraints (QCs) are widely used to characterize nonlinearities and uncertainties, but generic analytical characterizations can be conservati

Quantum End-to-End Learning for Contextual Combinatorial Optimization

arXiv:2605.20222v1 Announce Type: cross Abstract: Contextual combinatorial optimization (CCO) plays a critical role in decision-making under uncertainty, yet remains a significant challenge. We presen

REFLECTOR: Internalizing Step-wise Reflection against Indirect Jailbreak

arXiv:2605.20654v1 Announce Type: new Abstract: While Large Language Models (LLMs) demonstrate remarkable capabilities, they remain susceptible to sophisticated, multi-step jailbreak attacks that circ

Regulating Anatomy-Aware Rewards via Trajectory-Integral Feedback for Volumetric Computed Tomography Analysis

arXiv:2605.20277v1 Announce Type: new Abstract: Medical vision-language models (VLMs) have rapidly advanced as general-purpose multimodal assistants, yet their deployment in 3D Computed Tomography (CT

Reinforcement Learning for Risk Adaptation via Differentiable CVaR Barrier Functions

arXiv:2605.21257v1 Announce Type: new Abstract: Planning through crowded environments under uncertain obstacle motions remains difficult, as stochastic interactions often induce overly conservative be

Reinforcement Learning with Discrete Diffusion Policies for Combinatorial Action Spaces

arXiv:2509.22963v3 Announce Type: replace Abstract: Reinforcement learning (RL) struggles to scale to large, combinatorial action spaces common in many real-world problems. This paper introduces a nov

rePIRL: Learn PRM with Inverse RL for LLM Reasoning

arXiv:2602.07832v2 Announce Type: replace Abstract: Process rewards have been widely used in deep reinforcement learning to improve training efficiency, reduce variance, and prevent reward hacking. In

Rethinking Cross-Layer Information Routing in Diffusion Transformers

arXiv:2605.20708v1 Announce Type: new Abstract: Diffusion Transformers (DiTs) have become a de facto backbone of modern visual generation, and nearly every major axis of their design -- tokenization,

Robust Recommendation from Noisy Implicit Feedback: A GMM-Weighted Bayes-label Transition Matrix Framework

arXiv:2605.20721v1 Announce Type: new Abstract: Learning from implicit feedback in recommender systems is fundamentally challenged by pervasive label noise. While conventional denoising approaches oft

SAM-Sode: Towards Faithful Explanations for Tiny Bacteria Detection

arXiv:2605.21186v1 Announce Type: new Abstract: Interpretability in object detection provides crucial confidence support for clinical auxiliary diagnosis. However, in tiny bacteria detection, traditio

SCRIBE: Diagnostic Evaluation and Rich Transcription Models for Indic ASR

arXiv:2605.20712v1 Announce Type: new Abstract: Automatic speech recognition replaces typing only when correction costs less than manual entry, a threshold determined by error types, not counts: fixin

Secure, Verifiable, and Scalable Multi-Client Data Sharing via Consensus-Based Privacy-Preserving Data Distribution

arXiv:2601.00418v2 Announce Type: replace-cross Abstract: We propose the Consensus-Based Privacy-Preserving Data Distribution (CPPDD) framework, a lightweight and post-setup autonomous protocol for se

Self-Refining Video Sampling

arXiv:2601.18577v2 Announce Type: replace Abstract: Modern video generators still struggle with complex physical dynamics, often falling short of physical realism. Existing approaches address this usi

Shiny Stories, Hidden Struggles: Investigating the Representation of Disability Through the Lens of LLMs

arXiv:2605.20191v1 Announce Type: new Abstract: Modern Large Language Models (LLMs) have recently attracted much attention for their ability to simulate human behavior and generate text that reflects

Shipping features to production just got easier with new feature flags in AppLifecycle Manager

Many development teams are familiar with the hesitation that comes right before pushing a new feature live. As AI helps developers write code faster, the gap between rapid code generation and safe pro

safetygoogle-cloud-ai

SMA-DP: Spectral Memory-Aware Differential Privacy for Deep Learning

arXiv:2605.20450v1 Announce Type: new Abstract: Differentially private stochastic gradient descent (DP-SGD) enables private deep learning through per-example clipping and calibrated Gaussian noise, bu

Spacetime Optimal-Transport Attention for Visuo-Haptic Imitation Learning of Contact-Rich Manipulation

arXiv:2605.20433v1 Announce Type: new Abstract: Contact-rich manipulation tasks such as tight-clearance insertion, connector mating, polishing, and surface-conforming wiping remain difficult for data-

SpaceX claims a 28.5 TRILLION market let us look at Musk's pitch deck to banks for the twitter takeover in 2022, he said he'd create a 10 …

SpaceX claims a 28.5 TRILLION market let us look at Musk's pitch deck to banks for the twitter takeover in 2022, he said he'd create a 10 billion dollar subs biz by 2028 today, ad revs down 75% & subs

safetygary-marcus--x

Spatial Gram Alignment for Ultra-High-Resolution Image Synthesis

arXiv:2605.20808v1 Announce Type: new Abstract: Modern ultra-high-resolution image synthesis relies heavily on the robust generative capacity of large-scale pre-trained Latent Diffusion Models (LDMs).

Spectral Souping: A Unified Framework for Online Preference Alignment

arXiv:2605.20408v1 Announce Type: new Abstract: Reinforcement Learning from Human Feedback (RLHF) effectively aligns Large Language Models (LLMs) with aggregate human preferences but often fails to ad

Stage-Audit: Auditable Source-Frontier Discovery for Cross-Wiki Tables

arXiv:2605.20478v1 Announce Type: new Abstract: LLM-curated tables can appear source-grounded while containing unsupported rows: the curator may recall entries from parametric memory and retroactively

Statistical Guarantees in the Search for Less Discriminatory Algorithms

arXiv:2512.23943v2 Announce Type: replace-cross Abstract: U.S. discrimination law can impose liability on firms that fail to adopt a less discriminatory alternative (LDA): a decision policy that achie

STEAM: A Training-Free Congestion-Aware Enhancement Framework for Decentralized Multi-Agent Path Finding

arXiv:2605.20929v1 Announce Type: new Abstract: We propose STEAM (Spatial, Temporal, and Emergent congestion Awareness for MAPF), a training-free test-time enhancement framework for learning-based dec

STiTch: Semantic Transition and Transportation in Collaboration for Training-Free Zero-Shot Composed Image Retrieval

arXiv:2605.21261v1 Announce Type: new Abstract: Training-free zero-shot composed image retrieval models are recently gaining increasing research interest due to their generalizability and flexibility

Strict Subgoal Execution: Reliable Long-Horizon Planning in Hierarchical Reinforcement Learning

arXiv:2506.21039v3 Announce Type: replace Abstract: Long-horizon goal-conditioned tasks pose fundamental challenges for reinforcement learning (RL), particularly when goals are distant and rewards are

Subword boundaries are the second meaningful effect. Adding end-of-subword markers as input embeddings produces a large gain throughout trai…

Subword boundaries are the second meaningful effect. Adding end-of-subword markers as input embeddings produces a large gain throughout training (H3): end-boundaries leak future bytes (whitespace alwa

safetynous-research--x

SUGAR: A Scalable Human-Video-Driven Generalizable Humanoid Loco-Manipulation Learning Framework

arXiv:2605.20373v1 Announce Type: cross Abstract: Building humanoid robots capable of generalizable whole-body loco-manipulation in the real world remains a fundamental challenge. Existing methods eit

Supervised Latent Restructuring for Small-Data Quantum Learning in Plant Phenomics

arXiv:2605.20413v1 Announce Type: new Abstract: High-dimensional biological data often exhibit a severe mismatch between feature dimensionality and sample size, making reliable classification difficul

SURF: Steering the Scalarization Weight to Uniformly Traverse the Pareto Front

arXiv:2605.20619v1 Announce Type: new Abstract: Scalarization is widely used in multi-objective optimization owing to its simplicity and scalability. In many applications, the goal is to generate solu

SurgOnAir: Hierarchy-Aware Real-Time Surgical Video Commentary

arXiv:2605.21132v1 Announce Type: new Abstract: Understanding surgical workflow in real time is fundamental for intelligent surgical embodiment, where AI systems continuously perceive and respond as s

SynCB: A Synergy Concept-Based Model with Dynamic Routing Between Concepts and Complementary Neural Branches

arXiv:2605.20908v1 Announce Type: new Abstract: Concept-based (CB) models provide interpretability and support test-time human intervention, while standard neural networks (NN) offer strong task perfo

Synchronization and Turn-Taking in Full-Duplex Speech Dialogue Models

arXiv:2605.20356v1 Announce Type: new Abstract: Full-duplex spoken dialogue models (SDMs) can listen and speak simultaneously, enabling interaction dynamics closer to human conversation than turn-base

Task-conditioned probing of instruction-tuned multimodal LLMs: Region-specific brain alignment patterns under naturalistic stimuli

arXiv:2506.08277v3 Announce Type: replace-cross Abstract: Recent voxel-wise multimodal brain encoding studies have shown that multimodal large language models (MLLMs) exhibit a higher degree of brain

Tech researchers are suing the Trump administration over the future of online safety

Since its earliest days back in office, the Trump administration has been going after researchers who study and try to counter hate speech, harassment, propaganda, and disinformation online. Now, some

safetymit-tech-review

TelePhysics: Physics-Grounded Multi-Object Scene Generation from a Single Image with Real-Time Interaction

arXiv:2605.20290v1 Announce Type: cross Abstract: Recent generative video models achieve impressive visual quality but remain constrained by limited physical consistency and controllability. Existing

The Download: online safety’s future and climate tech’s big pivot

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Tech researchers are suing the Trump administration over the f

safetymit-tech-review

The Economics of AI Inference: Inflation Dynamics, Welfare Costs, and Optimal Monetary Policy under the Inference-Cost Phillips Curve

arXiv:2605.20281v1 Announce Type: cross Abstract: We develop a unified microeconomic and monetary theory of artificial intelligence inference costs and their pass-through to inflation, welfare, and op

The Illusion of Intervention: Your LLM-Simulated Experiment is an Observational Study

arXiv:2605.20767v1 Announce Type: new Abstract: Large language models (LLMs) show potential as simulators of human behavior, offering a scalable way to study responses to interventions. However, becau

Time-Prompt: Integrated Heterogeneous Prompts for Unlocking LLMs in Time Series Forecasting

arXiv:2506.17631v4 Announce Type: replace Abstract: Time series forecasting aims to model temporal dependencies among variables for future state inference, holding significant importance and widesprea

Time-To-Reach Separation and Safety Filtering for Safe, Fair, and Efficient Multi-Agent Coordination

arXiv:2605.20625v1 Announce Type: cross Abstract: Advanced Air Mobility (AAM) operations are expected to significantly increase aerial traffic in urban airspace, requiring autonomous traffic managemen

Towards Context-Invariant Safety Alignment for Large Language Models

arXiv:2605.20994v1 Announce Type: new Abstract: Preference-based post-training aligns LLMs with human intent, yet safety behavior often remains brittle. A model may refuse a harmful request in a stand

Trusted Weights, Treacherous Optimizations? Optimization-Triggered Backdoor Attacks on LLMs

arXiv:2605.20641v1 Announce Type: cross Abstract: Inference optimization is a vital technique for deploying LLMs at scale. Compilation is the most widely adopted optimization technique for LLMs. While

Tutor-Student Reinforcement Learning: A Dynamic Curriculum for Robust Deepfake Detection

arXiv:2603.24139v2 Announce Type: replace Abstract: Standard supervised training for deepfake detection treats all samples with uniform importance, which can be suboptimal for learning robust and gene

Uh-oh, the International Space Station is leaking again

The International Space Station has an ongoing air leak in its Russian PrK module that has been known since 2019 , and this leak has delayed the launch of Axiom Space's private astronaut mission . NAS

safetyars-technica