AI Paper Atlas

Research map / Reinforcement learning / Imitation learning

Residual Q-Learning: Offline and Online Policy Customization without Value

Chenran Li, Chen Tang, Haruki Nishimura, Jean Mercat, Masayoshi Tomizuka, Wei Zhan

NeurIPS 2023 · Reinforcement learning › Imitation learning · approach: Behavior cloning theory and models · 3 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
customize RL prior with residual Qsoft Q-learning prior · CartPolemean abs cart position error e_x: 0.15 -> 0.05Improved
customize IL prior with residual QGAIL prior · CartPolemean abs cart position error e_x: 0.19 -> 0.07Improved
customize RL prior with residual soft actor-criticSAC prior · Continuous Mountain Carnegative actions per episode n_neg: 42.26 -> 37.90Improved
customize IL prior with residual soft actor-criticGAIL prior · Continuous Mountain Carnegative actions per episode n_neg: 64.40 -> 41.08Improved
customize RL prior with residual soft Qsoft Q-learning prior · Highwayavg normalized lane index: 0.46 -> 0.97Improved
customize IL prior with residual soft QGAIL prior · Highwayavg normalized lane index: 0.50 -> 0.99Improved
customize RL prior with residual soft actor-criticSAC prior · Parkingnon-violation rate (%): 57.61 -> 96.09Improved
customize IL prior with residual soft actor-criticGAIL prior · Parkingnon-violation rate (%): 11.27 -> 66.07Improved
customize IL prior with residual SAC (basic task side effect)GAIL prior · Parkingbasic task success rate (%): 70.91 -> 83.59Improved
customize RL prior with residual soft Qsoft Q-learning prior · Highwaybasic task reward: 44.67 -> 42.23Mixed
customize RL prior with residual SACSAC prior · Parkingbasic task success rate (%): 99.09 -> 98.73No clear effect
zero-shot residual max-entropy MCTS instead of offline residual QRL prior · Highwayavg normalized lane index: 0.97 -> 0.96No 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