AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Evaluating Model-Based Planning and Planner Amortization for Continuous Control

Arunkumar Byravan, Leonard Hasenclever, Piotr Trochim, Mehdi Mirza, Alessandro Davide Ialongo, Yuval Tassa, Jost Tobias Springenberg, Abbas Abdolmaleki, Nicolas Heess, Josh Merel, Martin Riedmiller

ICLR 2022 · Reinforcement learning › Visual and representation RL · approach: RL for combinatorial routing · 1 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (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 changedWhereMeasuredEffect
Use a pre-trained task-agnostic proposal instead of a zero-mean Gaussian proposal for MPCMPC with learned or ground-truth dynamics · Ant and OP3 locomotion tasksreward: 56± 1 -> 420±14 (Ant GTTP); 36± 1 -> 178±4 (OP3 GTTP)Improved
Combine MPC data collection with MPO policy learningMPC+MPO · Ant and OP3 GTTPactor performanceImproved
Add behavioral cloning of planner-generated data to MPOMPO · GTTP tasksproposal performanceImproved
Combine MPC data collection with MPO and behavioral cloningMPC+MPO+BC · GTTP taskslearning speedImproved
Use MPC data collection with MPO without behavioral cloningMPC+MPO · GTTP tasksproposal performanceNo clear effect
Add behavioral cloning to MPC+MPOMPC+MPO · Ant and OP3 walkingperformanceNo clear effect
Transfer a learned model to the target task, either frozen or finetunedMPC+MPO+BC · OP3 GTTP and cross-task transferslearning speedNo clear effect
Keep the transferred model frozen rather than finetuning itMPC+MPO+BC · OP3 GTTPperformanceHurt
Transfer a proposal from forward walking to GTTP or from GTTP to backward walkingMPO+BC and MPC+MPO+BC · Cross-task transferlearning speedHurt
Always plan rather than interleave planning and proposal actions (p_plan=1.0 instead of p_plan=0.5) (ablation)MPC+MPO+BC · OP3 GTTPperformanceHurt
Increase planner sample count above the default N=250 (ablation)MPC+MPO+BC · OP3 GTTPperformanceNo clear effect
Use a longer planning horizon, H=20, instead of the default H=10 (ablation)MPC+MPO+BC · OP3 GTTPperformanceNo clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy