AI & ML
From Self-Play RL to Constrained Control: Distilling DiffMPC Students with DAgger
arXiv cs.RO0 views
arXiv:2611.25197v1 Announce Type: new Abstract: Self-play reinforcement learning (RL) produces strong driving policies, but these policies are hard to verify or interpret. Differentiable model predictive control (DiffMPC) can enforce explicit constraints, but it is usually trained by imitating logged data or with costly RL. To add constraints to an already trained policy, we keep a PPO teacher's backbone and replace its action head with a DiffMPC that enforces road boundary half-planes and agent ellipses. We then distill the student with DAgger in the teacher's self-play simulator. Over 9 seeds, the constrained student of a smaller teacher matches its teacher on nuPlan and beats it on CARLA and on out-of-distribution Waymo scenes. For a stronger teacher, the student has fewer failures on nuPlan and Waymo, while staying level on CARLA. Its distillation takes no more than 30% of the teacher's training time.
Source:https://arxiv.org/abs/2611.25197PreviousOld-School Planners Drive (Sometimes) Better Than Learned OnesNextVG-Steer: Visual-Geometric Policy Steering from In-Context Aligned Abstractions
Related posts

Roast My Dev Desktop: Zero Minimalism, Maximum Chaos
0dev.to
랩트 모네로 붕괴: 익숙한 신뢰 배신 이야기
0dev.to
Why has it become harder and harder to type one-handed on smartphones these days?
0dev.to
내 PDF 문서를 로컬 LLM에게 검색시키기 — RAG 실전 구축
0dev.to
How we tell parents what a teen's AI chat was about, without showing them the chat
0dev.to
GitHub Actions GITHUB_TOKEN: 11 Tags Pushed, 0 Release Workflows Ran
0dev.to