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

Your guide to containers at AWS re:Invent 2026
0AWS Containers Blog

Three Asian forwarders, one corridor: what Nippon Express, Yusen and CJ Logistics actually publish
0dev.to Data
Tương quan số liệu Nice vs PSG: Phép thử phòng thủ tại Allianz Riviera
0dev.to Data
Jacksonville, Florida Property Listings Dataset (2026-11-04): 6534 Rows, Excel Spreadsheet
0dev.to Data
Tương quan phong độ Parma vs AS Roma: Bài toán kiểm soát hỏa lực đội khách
0dev.to Data
Schalke 04 chạm trán Stuttgart: Bài toán hiệu suất chuyển hóa tại VELTINS-Arena
0dev.to Data