safety

Mask, Sample, Revise: A Revisable CTMC Inference Stack for Guided Discrete Flow Matching Text-to-Speech

arXiv:2606.13989v2 Announce Type: replace-cross Abstract: Recent alignment-free non-autoregressive (NAR) text-to-speech (TTS) models formulate synthesis as a conditional infilling task, bypassing expl

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Math for AI safety: an invitation for mathematicians

DGX agent

arXiv:2609.15289v1 Announce Type: cross Abstract: Artificial intelligence threatens to outrun human understanding and control. New mathematics is needed to design AI that is legible, steerable, and co

safetyarxiv-cs-ai

15 Sept 2026

NeuroProlog: Multi-Task Fine-Tuning for Neurosymbolic Mathematical Reasoning via the Cocktail Effect

arXiv:2603.02504v3 Announce Type: replace Abstract: Large Language Models (LLMs) achieve strong performance on natural language tasks but remain unreliable in mathematical reasoning, frequently genera

New Conditions for Philosophers to Catch the Wave of Citizen Deliberation in the Age of Artificial Intelligence in advance

arXiv:2609.15707v1 Announce Type: new Abstract: Powerful technologies labeled ``AI''-without sufficient epistemic caution-are already reshaping political and private life, bringing both new dangers an

Non-Asymptotic Global Convergence of PPO-Clip

arXiv:2512.16565v2 Announce Type: replace-cross Abstract: Reinforcement learning has gained attention for modern Large Language Model post-training. The actor-only variants of Proximal Policy Optimiza

Nonparametric Variance-Penalized Actor-Critic: Statistical Inference for Risk-Sensitive Reinforcement Learning

arXiv:2609.14327v1 Announce Type: new Abstract: Variance penalization is a principled approach to risk-sensitive reinforcement learning (RL) that explicitly trades expected return for policy stability

Not All Prompts Are Equal: Exploration-Guided Prompt Scaffolding for Multimodal Reinforcement Post-Training

arXiv:2609.15051v1 Announce Type: cross Abstract: Training prompts in online reinforcement learning (RL) differ substantially in how informative they are for the current policy: some are already satur

One Feedback System Does Not Fit All: Localising Data-to-Text Driver Coaching for the United Kingdom and Nigeria

arXiv:2609.14687v1 Announce Type: new Abstract: Data-to-text driver coaching is often presented as a generic pipeline from telematics events to advice. This paper argues that its content requires loca

safetyarxiv-cs-cl

Online Bayesian Node Classification on Inductive Graphs under Distribution Shift

arXiv:2609.13655v1 Announce Type: new Abstract: On evolving graphs, node classifiers must satisfy two key requirements: inductive generalization to newly arriving nodes under distribution shift and ca

OpenAI backs a provision of the FRONTIER Act, a bipartisan House proposal, that would require AI companies to embed outside evaluators to ensure model safety (Brendan Bordelon/Politico)

Brendan Bordelon / Politico: OpenAI backs a provision of the FRONTIER Act, a bipartisan House proposal, that would require AI companies to embed outside evaluators to ensure model safety — Artificial

safetytechmeme

OpenAI says it has been working with Anthropic and Google for several weeks on AI safety, and it does not need an antitrust waiver to coordinate on safety (Maggie Eastland/Bloomberg Law)

Maggie Eastland / Bloomberg Law: OpenAI says it has been working with Anthropic and Google for several weeks on AI safety, and it does not need an antitrust waiver to coordinate on safety — OpenAI is

safetytechmeme

Operational Range Bounding in Spectroscopy: A Safety Cage Framework for Machine Learning Models

arXiv:2609.13514v1 Announce Type: new Abstract: Ensuring the reliability of black-box machine learning models in safety-critical space missions remains a significant challenge, particularly when groun

Optimizing Sparse Outcomes Through Dense Behavioral Signals via Value-Guided Preference Distillation

arXiv:2609.14648v1 Announce Type: new Abstract: Aligning multi-turn dialogue agents is usually framed as matching turn-level human preferences, yet direct optimization of long-term outcomes is often i

safetyarxiv-cs-cl

OptoAgent: A Trustworthy Multi-Agent Framework for Opportunistic Vision Micro-Screening in Classroom Environments

arXiv:2609.14514v1 Announce Type: new Abstract: A child with reduced distance vision often does not know that anything is wrong. Children adapt, move closer, and rarely report the difficulty, so the p

Orientation Control of Soft Robots via Adiabatic Spectral Submanifolds

arXiv:2609.14376v1 Announce Type: new Abstract: Soft robots are commonly sought for safety-critical interactions in delicate environments, where accurate position and orientation control is imperative

Personalizing Personal Health Interfaces: Co-Design with Generative AI

arXiv:2609.15046v1 Announce Type: cross Abstract: Personal health interfaces present wellbeing data through standardized dashboards that rarely fit how people interpret or act on it. Personalizing the

Physical Kernel: Structured Visual Latents for Dark Manipulation

arXiv:2609.13244v1 Announce Type: cross Abstract: We study dark manipulation: after a brief lit Write encodes z0 = Enc(rgb), a policy pi(z) and open-loop dynamics f(z,a) complete contact-rich skills w

PolicyMem: Geometric Policy Memory for LLM Governance

arXiv:2609.13734v1 Announce Type: cross Abstract: As large language models (LLMs) are increasingly deployed in real-world high-stakes applications, effective governance has become essential. Existing

Predictive audio representations for early detection and tracking of hidden dynamic objects

arXiv:2609.13595v1 Announce Type: cross Abstract: Predicting potential dangers is core to safety. Forecasting the presence of other traffic agents is core to danger prediction. Occluded traffic agents

PredTac: Learning Contact-Rich Manipulation with Predicted Touch

arXiv:2609.15198v1 Announce Type: new Abstract: Contact-rich manipulation benefits from tactile feedback, yet physical tactile sensors introduce hardware, calibration, synchronization, and maintenance

Prelimit Coupling and Steady-State Convergence of Constant-stepsize Nonsmooth Contractive SA

arXiv:2404.06023v3 Announce Type: replace-cross Abstract: Motivated by Q-learning, we study nonsmooth contractive stochastic approximation (SA) with constant stepsize. We focus on two important classe

PRI-Net: A Lightweight Multimodal Framework for 3D UAV Localization

arXiv:2609.14469v1 Announce Type: new Abstract: Accurate 3D localization of unmanned aerial vehicles (UAVs) remains challenging for existing multimodal approaches due to sparse LiDAR geometry, modalit

Primitive-Informed Sampling-Based MPC for Multi-Fingered Dexterous Manipulation

arXiv:2609.14868v1 Announce Type: new Abstract: We present a primitive-informed sampling-based model predictive control (MPC) framework for multi-fingered dexterous manipulation. Sampling-based MPC av

Principal-timestep Restricted Init via Sparse Matrix-decomposition in Flow-matching

arXiv:2609.15643v1 Announce Type: new Abstract: Flow-matching diffusion models have recently emerged as a strong paradigm for high-fidelity visual generation. However, their prohibitively high fine-tu

Privileged observations enable rapid and reliable policy discovery directly in the physical world

arXiv:2512.08463v2 Announce Type: replace Abstract: We study how privileged information about a physical system affects the discovery of high-performing policies when training a reinforcement learning

Real-Time Synthesis of Robust Controlled Invariant Sets for Monotone Systems

arXiv:2609.14115v1 Announce Type: cross Abstract: Safety-critical control of autonomous systems requires formal safety certificates, such as controlled invariant sets, that must be computed online as

Real-World Reinforcement Learning with MPC Scaffolding for Dexterous Manipulation

arXiv:2609.14878v1 Announce Type: new Abstract: Real-world reinforcement learning (RL) offers a promising route to dexterous manipulation policies that can adapt directly from physical interaction, bu

Reason What Matters: Retrieval-Grounded Reasoning for Universal Multimodal Embeddings

arXiv:2609.15296v1 Announce Type: new Abstract: Universal multimodal embedding (UME) learns unified representations across modalities, enabling a single model to support diverse retrieval tasks. Recen

ReCAST: Reward Credit Assignment across Timesteps for Online Diffusion Reinforcement

arXiv:2609.13425v1 Announce Type: cross Abstract: Training diffusion models with multiple rewards requires distinguishing user preference from reward informativeness. User preference determines how mu

Recoverability as a System Primitive for Long-Horizon AI Agents

arXiv:2609.13672v1 Announce Type: new Abstract: AI agents can be interrupted while editing files, calling tools, or carrying out multi-step tasks. Restarting repeats completed work, but continuing fro

Recurrent GraphNeural NetworkswithSet-BasedAggregation

arXiv:2609.15932v1 Announce Type: new Abstract: Recurrent GNNs iterate message passing to convergence, and their logical characterizations to date rely on multi-set aggregation, graded (counting) logi

Reducing the Output-Mode Gap in Speech Language Models via Joint-Output On-Policy Distillation

arXiv:2609.15313v1 Announce Type: cross Abstract: Autoregressive generation of interleaved text and acoustic tokens is a common approach to spoken-response generation in speech large language models.

safetyarxiv-cs-cl

Refinement-based Flow Policy Optimization

arXiv:2609.15123v1 Announce Type: cross Abstract: Flow-based policies offer an expressive representation for online reinforcement learning, but conventional flow matching requires samples drawn from t

ResSafe: Learning Safety Filtering with Residual Reinforcement Learning for Humanoids

arXiv:2609.15988v1 Announce Type: new Abstract: Safe control of humanoid robots remains challenging due to their high-dimensional dynamics, contact-rich interactions, and sensitivity to disturbances.

Restore What Matters: Lessons from Joint Restoration and Recognition

arXiv:2609.13791v1 Announce Type: new Abstract: Recognition pipelines typically adopt a restore-then-recognize workflow, yet decades of experience show that generating visually pleasing images seldom

Retrieval-Guided Fine-Tuning as Noisy Estimation: Risk bounds and Architectural Analysis

arXiv:2609.14485v1 Announce Type: new Abstract: Retrieval-Guided Fine-Tuning (RAG-FT) incorporates retrieved data directly into the training objective, but the statistical consequences of noisy retrie

REVOLVE: An Automated Closed-Loop Framework for Evolving Robot Manipulation with Minimal Human Intervention

arXiv:2609.14633v1 Announce Type: new Abstract: Recent advances in data-driven robot manipulation policies have substantially improved task execution and generalization. However, real-world deployment

ReWeight: Leveraging Human Data for VLA Post-Training via Demonstration Retrieval and Sample Weighting

arXiv:2609.13851v1 Announce Type: cross Abstract: Post-training vision-language-action (VLA) models for specific robots and tasks requires in-domain demonstrations, yet collecting diverse robot data i

Rotation-Based Subspace Tracking for Robust Kernel PCA on Streaming Data

arXiv:2609.15488v1 Announce Type: new Abstract: Machine learning models process large amounts of data, and Principal Component Analysis (PCA) is a widely used technique to reduce the dimensionality of

Runtime-Incremental Transformer for Reinforcement-Learning-Based Adaptive Control

arXiv:2609.13545v1 Announce Type: cross Abstract: Learning-based adaptive control of robotic manipulators with non-observable friction memory has been addressed by attention- based meta-controllers wh

Safe Meta-Reinforcement Learning via Information Space Reachability

arXiv:2609.15915v1 Announce Type: new Abstract: Meta-reinforcement learning (meta-RL) enables agents to adapt to unseen tasks with limited experience. Despite its promise, the application of meta-RL i

Safety as a Constraint: Fine-Tuning a LLM Recommender to Explain Itself

arXiv:2609.13657v1 Announce Type: new Abstract: Traditional recommender systems are typically trained to predict what item users will interact with next, but not why. However, offering personalized ev

Safety Signals to Verify NetOps Agents with Action-Level Granularity

arXiv:2609.14422v1 Announce Type: new Abstract: Agentic Network Operations (NetOps) are an emerging paradigm promising to enable workload-aware, self-adjustable, and reliable autonomous networks. Whil

SAM3D-Part: Interactive Part Selection and Generation from 3D Objects

arXiv:2609.15639v1 Announce Type: new Abstract: Part-level control is essential for modern 3D asset creation, where objects are frequently edited, reused, animated, or fabricated through their individ

SEA-Nav: Efficient Policy Learning for Safe and Agile Quadruped Navigation in Cluttered Environments

arXiv:2603.09460v2 Announce Type: replace Abstract: Efficiently learning safe and agile quadruped navigation in densely cluttered environments remains difficult: existing methods often lack safety and

Seeing What the Vehicle Sees: Video-Augmented Virtual Reality for Physical Autonomous Vehicles

arXiv:2609.13224v1 Announce Type: new Abstract: Autonomous vehicles are expected to improve road safety and efficiency, but passengers often remain uncertain about what the vehicle perceives and why i

Semantic Fibers and Cross-Gram Interference: A Calculus of Safety Drift in Overcomplete Representations

arXiv:2609.14861v1 Announce Type: cross Abstract: A deployed language model may refuse a harmful request in English yet comply with its faithful translation, revealing a cross-lingual safety failure t

SGWIB:Sliced Gromov-Wasserstein Information Bottleneck for Video Highlight Detection

arXiv:2609.13966v1 Announce Type: cross Abstract: Video highlight detection aims to identify temporally important segments that capture the most informative or engaging events in a video. Reliable pre