Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Efficient Learning of Safe Driving Policy via Human-AI Copilot Optimization
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Train with HACO rather than SAC-RS | Driving policy · MetaDrive test environments | Test Success Rate: 0.82 { $\pm$0.18} -> { \textbf{0.83} { $\pm$ 0.04}} | No clear effect |
| Train with HACO rather than SAC-RS | Driving policy · MetaDrive training | Total Training Safety Violation: 2.76K { $\pm$ 0.95K } -> \textbf{30.14} { $\pm$ 11.36} | Helped |
| Train with HACO rather than PPO in town 1 | Driving policy · CARLA town 2 | Train Samples: 500,000 -> 8,000 | Helped |
| Use less frequent, longer human interventions instead of denser intervention signals (ablation) | HACO · MetaDrive test environments | Test Success Rate: 0.83 { $\pm$ 0.04} -> 0.397 | Hurt |
| Replace cosine similarity intervention cost with a constant cost (ablation) | HACO · MetaDrive test environments | Test Success Rate: { 0.83 { $\pm$ 0.04}} -> 0.00 | Hurt |
| Remove intervention minimization (ablation) | HACO · MetaDrive test environments | Test Success Rate: { 0.83 { $\pm$ 0.04}} -> 0.00 | Hurt |
Closest papers by meaning
- DriveDPO: Policy Learning via Safety DPO For End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.86)
- Predictive Preference Learning from Human Interventions (NeurIPS 2025 · similarity 0.85)
- Distill to Think, Foresee to Act: Cognitive-Physical Reinforcement Learning for Autonomous Driving (NeurIPS 2026 · similarity 0.85)
- Beyond Self-Play and Scale: A Behavior Benchmark for Generalization in Autonomous Driving (NeurIPS 2026 · similarity 0.84)
- Prioritizing Perception-Guided Self-Supervision: A New Paradigm for Causal Modeling in End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.84)
- HMARL-CBF – Hierarchical Multi-Agent Reinforcement Learning with Control Barrier Functions for Safety-Critical Autonomous Systems (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.