AI Paper Atlas

Research map / Reinforcement learning / Imitation learning

Receding Horizon Inverse Reinforcement Learning

Yiqing Xu, Wei Gao, David Hsu

NeurIPS 2022 · Reinforcement learning › Imitation learning · approach: Theory of inverse RL · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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 local receding-horizon control-sequence matching instead of global matchingRHIRL · CarRacingreturn ratio: 0.09 -> 0.40Improved
Use a noise-disentangled state cost with explicit control-noise modelingRHIRL · Hopper, mild and high control noisereturn ratio: 0.82 -> 0.98; 0.43 -> 0.96Improved
Use a noise-disentangled state cost with explicit control-noise modelingRHIRL · Walker2d, mild and high control noisereturn ratio: 0.82 -> 0.99; 0.78 -> 0.95Improved
Increase receding horizon from K = 5 to K = 20 (ablation)RHIRL · Hopper-v2final score: 1100 -> 3071.68Improved
Increase receding horizon from K = 20 to K = 100 (ablation)RHIRL · Hopper-v2final score: 3071.68 -> 3083Improved
Increase receding horizon from K = 20 to K = 100 (ablation)RHIRL · Hopper-v2environment steps to stabilize above 3000: fewer than 1e6 -> more than 1e7Hurt
Use RHIRL rather than f-IRLIRL methods · Pendulum, no noisereturn ratio: 1.07 -> 1.06No clear effect
Use RHIRL rather than f-IRLIRL methods · LunarLander, no noisereturn ratio: 0.76 -> 1.05Improved
Use RHIRL rather than f-IRLIRL methods · Ant, no noisereturn ratio: 0.87 -> 0.86No clear effect
Increase control noise in the evaluation environmentRHIRL learned cost · Pendulumreturn ratio: 1.06 -> 1.07 -> 1.06No clear effect
Increase control noise in the evaluation environmentRHIRL learned cost · Antreturn ratio: 0.86 -> 0.55 -> 0.15Hurt

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