AI Paper Atlas

Research map / Reinforcement learning / RL from feedback

Towards Adaptive Humanoid Control via Multi-Behavior Distillation and Reinforced Fine-Tuning

Yingnan Zhao, Xinmiao Wang, Dewei Wang, Xinzhe Liu, Dan Lu, Qilong Han, Peng Liu, Chenjia Bai

AAAI 2026 · Reinforcement learning › RL from feedback · approach: RL fine-tuning of generative 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
add RL fine-tuning on diverse terrains after distillationMulti-Behaviors Policy · Locomotion, Slope terraintraversal success rate: 0.160 -> 0.975Improved
add RL fine-tuning on diverse terrains after distillationMulti-Behaviors Policy · Locomotion, Discrete terraintraversal success rate: 0.702 -> 0.969Improved
add RL fine-tuning on diverse terrains after distillationMulti-Behaviors Policy · Locomotion, Hurdle terraintraversal success rate: 0.756 -> 0.922Improved
add RL fine-tuning on diverse terrains after distillationMulti-Behaviors Policy · Locomotion, Plane terraintraversal success rate: 0.993 -> 0.992No clear effect
add RL fine-tuning on diverse terrains after distillationMulti-Behaviors Policy · Fail recovery, Slope terrainrecovery success rate: 0.904 -> 1.000Improved
add RL fine-tuning on diverse terrains after distillationMulti-Behaviors Policy · Fail recovery, Hurdle terrainrecovery success rate: 0.997 -> 0.985Hurt
distill walking and recovery into one multi-behavior policyWalking Policy · Locomotion, Hurdle terraintraversal success rate: 0.161 -> 0.756Improved
distill walking and recovery into one multi-behavior policyFail Recovery Policy · Fail recovery, Slope terrainrecovery success rate: 0.757 -> 0.904Improved
add PCGrad gradient projection with a shared critic (ablation)AHC · Gradient cosine similaritycosine similarity: 0.247 -> 0.519Improved
add PCGrad gradient projection with behavior-specific critics (ablation)AHC · Gradient cosine similaritycosine similarity: 0.334 -> 0.535Improved
use behavior-specific critics instead of a shared critic (ablation)AHC · Gradient cosine similarity (no PCGrad)cosine similarity: 0.247 -> 0.334Improved
use behavior-specific critics instead of a shared critic (ablation)AHC · Stage-2 value loss and returnsvalue loss / episode returnImproved

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