Search: “safety”

Open-Ended Scenario Reasoning for Specialist Model Adaptation

arXiv:2607.06625v1 Announce Type: cross Abstract: Process industries have accumulated validated specialist models, yet sensor drift, feedstock variation, and regime switching cause these models to deg

ORAN-DEFEND: Subspace Detection and Sanitization of Backdoor DRL xApps in Open RAN

DGX agent

arXiv:2607.06647v1 Announce Type: cross Abstract: Open Radio Access Networks (O-RAN) increasingly delegate near-real-time control to deep reinforcement learning (DRL) xApps obtained from third-party v

safetyarxiv-cs-lg

9 Jul 2026

ORCAID: Oblique Rule-Based Continuous-Action Interpretation for Deep RL Policies

arXiv:2607.07235v1 Announce Type: cross Abstract: Explainability remains a key issue in reinforcement learning (RL). Distilling an interpretable policy from an agent trained in a complex environment i

Pelican-VLA 0.5: Attending Before Acting Benefits Generalization

arXiv:2607.06655v1 Announce Type: cross Abstract: In this report, we present Pelican-VLA 0.5, a unified VLA model that integrates vision-language understanding, future-frame generation, and action pre

Provably Efficient Off-Policy Adversarial Imitation Learning with Convergence Guarantees

arXiv:2405.16668v2 Announce Type: replace Abstract: Adversarial Imitation Learning (AIL) faces challenges with sample inefficiency because of its reliance on sufficient on-policy data to evaluate the

Rapidly Learning Soft Robot Control via Implicit Time-Stepping

arXiv:2511.06667v2 Announce Type: replace-cross Abstract: With the explosive growth of rigid-body simulators, policy learning in simulation has become the de facto standard for most rigid morphologies

RoboSnap: One-Shot Real-to-Sim Scene Generation for Generalizable Robot Learning and Evaluation

arXiv:2607.06699v1 Announce Type: new Abstract: Recovering real-world scenes as interactive simulation environments can enable generalizable robot learning and reproducible policy evaluation. However,

safetyarxiv-cs-ro

Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence

arXiv:2607.07675v1 Announce Type: new Abstract: Despite the recent promise in robot control, video generative models suffer from a domain mismatch due to their primary focus on content creation. For e

Selective Timestep Weighting and Advantage-Based Replay for Sample-Efficient Diffusion RLHF

arXiv:2607.07693v1 Announce Type: cross Abstract: Reinforcement learning from human feedback (RLHF) has emerged as a powerful paradigm for aligning generative models with human preferences. However, a

SHTA: Semantic Hard Token Correction and Center Alignment for Semi-Supervised Medical Image Segmentation

arXiv:2607.07019v1 Announce Type: new Abstract: Recent advances in semi-supervised medical image segmentation have achieved remarkable performance through prediction consistency, pseudo-label supervis

Single-Rollout Asynchronous Optimization for Agentic Reinforcement Learning

arXiv:2607.07508v1 Announce Type: cross Abstract: Reinforcement learning (RL) is becoming increasingly important for post-training large language models (LLMs). Previous RL pipelines for LLMs were mos

SOMtime the World Ain't Fair: Violating Fairness Using Self-Organizing Maps

arXiv:2602.18201v2 Announce Type: replace Abstract: Unsupervised representations are widely assumed to be neutral with respect to sensitive attributes when those attributes are withheld from training.

T2T-VICL: Cross-Task Visual In-Context Learning via Implicit Text-Driven VLMs

arXiv:2511.16107v3 Announce Type: replace-cross Abstract: Visual in-context learning (VICL) solves visual tasks by conditioning on a few input-output demonstrations without any model training. Recent

TACoS: Weakly Supervised Learning of Two-Dimensional Materials from Scribble Annotations to Precise Segmentation

arXiv:2607.07169v1 Announce Type: new Abstract: The precise pixel-level localization of 2D material flakes is crucial for high-throughput screening. However, traditional fully supervised methods rely

The New York Times and several news outlets are accusing OpenAI of hiding its ability to search its AI training data and output logs for mor…

The New York Times and several news outlets are accusing OpenAI of hiding its ability to search its AI training data and output logs for more than two years during an ongoing copyright lawsuit: • The

safetygary-marcus--x

Thinking Seeds: Leveraging Historical Diversity for Position-Aware RL in LLMs

arXiv:2601.21476v2 Announce Type: replace Abstract: On-policy reinforcement learning (RL) for language model post-training suffers from a fundamental tension: as training progresses, policy entropy co

safetyarxiv-cs-cl

This is a very important observation. AI is not adopting the language and shaping it to fit its known ideas. It has no known ideas. It is re…

This is a very important observation. AI is not adopting the language and shaping it to fit its known ideas. It has no known ideas. It is repeating patterns found in the language. super interesting -

safetygary-marcus--x

TIR-Agent: Training an Explorative and Efficient Agent for Image Restoration

arXiv:2603.27742v2 Announce Type: replace Abstract: Vision-language agents that orchestrate specialized tools for image restoration (IR) have emerged as a promising method, yet most existing framework

Towards a Pseudo-Labeling Workflow for Celltype-Classification from Explanted Brain Slice Recordings

arXiv:2607.06569v1 Announce Type: cross Abstract: This paper proposes an unsupervised workflow to pseudo-label extracellular spikes from human brain slice MEA recordings into two putative cell types:

Trajectory Inference of Human Aging from Cross-Sectional DNA Methylation Data

arXiv:2607.06583v1 Announce Type: cross Abstract: DNA methylation (DNAm) serves as one of the most robust molecular biomarkers of biological aging. While conventional epigenetic clocks accurately pred

Unmasking On-Policy Distillation: Where It Helps, Where It Hurts, and Why

On-policy distillation offers dense, per-token supervision for training reasoning models; however, it remains unclear under which conditions this signal is beneficial and under which it is detrimental

safetyapple-ml-research

UP: Unbounded Positive Asymmetric Optimization for Breaking the Exploration-Stability Dilemma

arXiv:2607.06987v1 Announce Type: new Abstract: Reinforcement learning (RL) has become the standard paradigm for enhancing the complex reasoning capabilities of large language models (LLMs). To achiev

VFM-Loc: Training-Free Cross-View Geo-Localization via Aligning Discriminative Visual Hierarchies

arXiv:2603.13855v2 Announce Type: replace Abstract: Cross-View Geo-Localization (CVGL) in remote sensing aims to locate a drone-view query by matching it to geo-tagged satellite images. Although super

Vision Foundation Models in Radiology: A Scoping Review of Data, Methodology, Evaluation and Clinical Translation

arXiv:2607.07219v1 Announce Type: cross Abstract: Vision foundation models (VFMs) are increasingly being developed for radiological imaging, yet their definition, development and evaluation remain het

When Agents Remember Too Much: Memory Poisoning Attacks on Large Language Model Agents

arXiv:2607.06595v1 Announce Type: cross Abstract: Personal AI agents powered by large language models can reason and act using available tools to access emails, manage calendars, and push code to remo

WHERE to Generate Matters: Budget-Aware Synthetic Augmentation for Label Skewed Federated Learning

arXiv:2607.06616v1 Announce Type: cross Abstract: Label skew in federated learning (FL) causes client drift and degrades global accuracy. Synthetic data augmentation can reduce this imbalance; however

Zoom In Disparities in Healthcare LLM Q&A

arXiv:2510.17476v2 Announce Type: replace Abstract: Equitable access to reliable health information is vital when integrating AI into healthcare. Yet, information quality varies across languages, rais

safetyarxiv-cs-cl

A Coin Flip Per Token: Bernoulli Sparse Steering of Large Language Models

arXiv:2607.05615v1 Announce Type: new Abstract: Activation steering via sparse autoencoders (SAEs) enables behavioral control of large language models without task-specific fine-tuning, but standard m

A Function-Space Dichotomy for Compositional Learning: Exponential Sub-Optimality of the Neural Tangent Kernel

arXiv:2607.06382v1 Announce Type: cross Abstract: A persistent empirical observation is that trained neural networks outperform their neural tangent kernel (NTK) limit on tasks with compositional stru

A toy framework for single and multi-agent human-AI curiosity ecosystems

arXiv:2607.06214v1 Announce Type: new Abstract: This paper offers a toy framework for considering curiosity as an ecosystem. First, it suggests that a single agent's inquiry policy (how, when, and why

Model Releases

A VLM-Enhanced Framework for Comprehensive Traffic Sign Condition Assessment Integrating Daytime Visual Performance and Nighttime Retroreflectivity Evaluation

model-releasesarxiv-cs-cv

arXiv:2607.06478v1 Announce Type: new Abstract: Traffic signs are crucial components of road safety, serving as visual tools under all lighting conditions. The Manual on Uniform Traffic Control Device

ARMS: Anchor-Relational Motion Streaming for Seamless Solo-Social Motion Transitions

arXiv:2607.05733v1 Announce Type: new Abstract: Generating temporally continuous and socially coherent human motion from text remains a fundamental challenge, particularly in realistic streams where p

Automated Recommendation of Programming Learning Content Using Pattern-based Knowledge Components

arXiv:2607.05409v1 Announce Type: cross Abstract: Introductory programming instruction relies on hands-on practice and short learning activities to support mastery of foundational concepts. Although m

Model Releases

Beyond Refusal: A Same-Lineage Study of Aligned and Abliterated LLMs for Vulnerability Analysis

model-releasesarxiv-cs-ai

arXiv:2607.05842v1 Announce Type: cross Abstract: Large language model (LLM)-assisted software security operates at a difficult boundary: the vulnerability-analysis terminology needed for legitimate c

Bridging Diffusion Pruning and Step Distillation with Teacher-Aligned Repair

arXiv:2607.06335v1 Announce Type: new Abstract: Diffusion models generate high-quality images, but their inference cost comes from two sources: large denoising networks and repeated denoising steps. E

Bridging Physical Reasoning and Task Generalization via Visual Action Outcome Reasoning Alignment

arXiv:2607.06522v1 Announce Type: new Abstract: Vision-language models (VLMs) struggle to generalize in interactive physical reasoning, particularly under unseen tasks and environments. Two key failur

Choose What to Observe: Task-Aware Semantic-Geometric Representations for Visuomotor Policy

arXiv:2603.07875v2 Announce Type: replace Abstract: Visuomotor policies learned from demonstrations often overfit to nuisance visual factors in raw RGB observations, resulting in brittle behavior unde

safetyarxiv-cs-ro

Controlling Tool Use with Heading-Specific Activation Steering

arXiv:2607.05790v1 Announce Type: new Abstract: Tool-augmented large language models extend their capabilities beyond parametric knowledge through external tools, but tend to invoke them unnecessarily

Cross-Contextual Vision-Language Adaptation with LoRA for Personalized Severe Adverse Event Detection in Clinical Wound Monitoring

arXiv:2607.05625v1 Announce Type: new Abstract: Wound monitoring is a critical yet underserved clinical challenge, where timely identification of severe adverse events (SAEs) such as infection, tissue

Deep Reinforcement Learning for Dynamic Origin-Destination Matrix Estimation in Microscopic Traffic Simulations Considering Credit Assignment

arXiv:2511.06229v3 Announce Type: replace Abstract: This paper focuses on dynamic origin-destination matrix estimation (DODE), a crucial calibration process necessary for the effective application of

DeSeG: Decoupling Semantic Intent and Geometric Constraints for Physically Plausible Human-Scene Interaction

arXiv:2607.05787v1 Announce Type: new Abstract: Synthesizing physically plausible human-scene interactions (HSI) remains a critical challenge in computer vision and the development of human avatars. A

Domain-Adaptive Climate Downscaling Under Temporal Distribution Shift

arXiv:2607.05645v1 Announce Type: new Abstract: Deep-learning-based climate downscaling aims to learn relationships from historical low-resolution (LR) and high-resolution (HR) climate data to generat

DreamPartGen: Semantically Grounded Part-Level 3D Generation via Collaborative Latent Denoising

arXiv:2603.19216v2 Announce Type: replace-cross Abstract: Understanding and generating 3D objects as compositions of meaningful parts is fundamental to human perception and reasoning. However, most te

DynaKRAG: A Unified Framework for Learnable Evidence Control in Multi-Hop Retrieval-Augmented Generation

arXiv:2607.06507v1 Announce Type: new Abstract: Multi-hop retrieval-augmented generation (RAG) acquires evidence sequentially, with each new document potentially revealing missing facts, bridge entiti

safetyarxiv-cs-cl

ELSA3D: Elastic Semantic Anchoring for Unified 3D Understanding and Generation

arXiv:2607.06565v1 Announce Type: cross Abstract: Unified 3D foundation models aspire to generate 3D assets and reason about them in language within a single backbone, but their text-3D interaction re

Embodied Human-Robot Interaction via Acoustics: A MARL Approach with AcoustoBots for Spatial Data Physicalization

arXiv:2607.06563v1 Announce Type: new Abstract: Traditional data physicalization is often static and disconnected from real environments, limiting its ability to convey embodied spatial dynamics and e

safetyarxiv-cs-ro

FADRA: Frequency-Aware Diffusion with Residual Adaptation for Video Face Restoration

arXiv:2607.06389v1 Announce Type: new Abstract: Video face restoration (VFR) aims to recover high-quality and temporally consistent facial details from severely degraded video sequences; however, exis

FreqDepthKV: Frequency-Guided Depth Sharing for Robust KV Cache Compression in Long-Context LLM Inference

arXiv:2607.06519v1 Announce Type: new Abstract: Long-context LLM inference is increasingly limited by the memory and bandwidth cost of KV caches, yet aggressive compression can remove the layer-specif