EvoDrive: Pareto Evolution for Safety-Critical Autonomous Driving via Self-Improving LLM Agents

arXiv:2606.03678 · cs.AI · Submitted 2026-06-02 · Read on arXiv

cs.AI

Submitted: 2026-06-02

Updated: 2026-10-09

Project page: https://tongnie.github.io/EvoDrive

Terminology

• Actor — 2× in this paper · explained in 3 episodes
• Actor-Critic Architecture — 1× in this paper · explained in 1 episode
• Benchmark — 1× in this paper · explained in 2 episodes
• Grounding — 1× in this paper · explained in 3 episodes

Sources

• Steerable Adversarial Scenario Generation through Test-Time Preference Alignment
• FREA: Feasibility-Guided Generation of Safety-Critical Scenarios with Reasonable Adversariality
• Enhancing Autonomous Vehicle Training with Language Model Integration and Critical Scenario Generation
• Voyager: An Open-Ended Embodied Agent with Large Language Models
• The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery
• AlphaEvolve: A coding agent for scientific and algorithmic discovery
• ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution
• CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery
• EvoX: Meta-Evolution for Automated Discovery
• ADV-0: Closed-Loop Min-Max Adversarial Training for Long-Tail Robustness in Autonomous Driving
• ReAct: Synergizing Reasoning and Acting in Language Models
• Meta-Harness: End-to-End Optimization of Model Harnesses

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems