arXiv AI June 15, 2026 By Abhinaw Priyadershi, Jelena Frtunikj

When and How Severely: Scenario-Specific Safety Envelopes for Driving VLAs

arXiv:2606. 14238v1 Announce Type: cross Abstract: Safety certification of Vision-Language-Action (VLA) driving planners under ISO 21448 (SOTIF) rests on an Operational Design Domain (ODD) specification that answers two complementary questions: when does the planner start to fail, and how severely does it fail once it does?

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

arXiv AI

Jun 4

Lost in Fog: Sensor Perturbations Expose Reasoning Fragility in Driving VLAs

arXiv:2605. 21446v2 Announce Type: replace-cross Abstract: Interpretable autonomous driving planners depend not only on generating explanations, but also on those explanations remaining reliable under real-world sensor degradation.

By Abhinaw Priyadershi, Jelena Frtunikj

agents robotics multimodal benchmarks safety

Hugging Face Trending Papers

Aug 26

Gating Before Commitment: Anticipating Intent Divergence to Prevent Post-Interaction Decision Failures in Autonomous Driving

Intent misinterpretation during vehicle interactions causes recurring planning failures. We study a decision layer in which a language-guided intent module reads structured descriptors, computes a smo...

agents robotics

arXiv Statistics ML

2d ago

CHOIR: heterogeneity-aware conformal prediction for crash injury severity across driver safety strata

arXiv:2609.11592v2 Announce Type: replace Abstract: Transportation agencies increasingly predict crash-injury severity with statistical and machine-learning models, but these models do not state how...

By Amir Rafe, Subasish Das

safety

arXiv Machine Learning

Jun 11

Seeing Before Colliding: Anticipatory Safe RL with Frozen Vision-Language Models

arXiv:2606. 11266v1 Announce Type: new Abstract: The cost signal that constrained-RL algorithms optimize against is almost always reactive: the simulator emits a non-zero cost only after a collision has begun, and the Lagrange multiplier of PPO-Lagrangian grows only after the episode budget has been exceeded.

By Samuel Tetteh, Cody Fleming

llms reinforcement-learning multimodal safety

arXiv AI

Jun 30

Budgeted Act-or-Defer Multi-Agent LLM Deliberation with Local Reliability Bounds

arXiv:2606. 29654v1 Announce Type: new Abstract: Multi-agent deliberation among LLMs can improve reasoning, but deployment requires deciding when the current answer is reliable enough to act on and when it should be escalated to human review.

By Mengdie Flora Wang, Haochen Xie, Guanghui Wang, Devin Zhang, Jae Oh Woo

llms agents benchmarks safety

arXiv Computer Vision

Sep 22

Beyond the Leaderboard: Counterfactual Diagnosis of End-to-End and VLA Driving Policies Under Domain Shift

arXiv:2609.22582v1 Announce Type: new Abstract: End-to-end and vision-language-action (VLA) driving policies are compared by leaderboard rank, but a rank reports an outcome, not the behaviour behind...

By Ruolin Yang, Zilin Huang, Buoyue Wang, Zhengyang Wan, Yuhao Luo, Zihao Sheng, Sikai Chen

multimodal benchmarks safety