AI Paper Atlas

Research map / Reinforcement learning / Policy gradient and actor-critic

PolicyFlow: Policy Optimization with Continuous Normalizing Flow in Reinforcement Learning

Shunpeng Yang, Ben Liu, Hua Chen

ICLR 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Tree search and policy optimization · 4 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
replace PPO Gaussian policy with PolicyFlowPPO · IsaacLab G1terminal episodic reward: 25.4 -> 30.0Improved
replace PPO Gaussian policy with PolicyFlowPPO · IsaacLab Navigationterminal episodic reward: 3.5 -> 4.2Improved
replace PPO Gaussian policy with PolicyFlowPPO · IsaacLab H1terminal episodic reward: 29.3 -> 27.3Hurt
replace PPO Gaussian policy with PolicyFlowPPO · IsaacLab Lift-Cubeterminal episodic reward: 153.1 -> 154.6No clear effect
replace PPO Gaussian policy with PolicyFlowPPO · IsaacLab Quadcopterterminal episodic reward: 141.8 -> 141.0No clear effect
replace PPO Gaussian policy with PolicyFlowPPO · IsaacLab Open-Drawerterminal episodic reward: 99.8 -> 99.1No clear effect
replace PPO Gaussian policy with PolicyFlowPPO · IsaacLab per-iteration training timeper-iteration time (ms): 43.0 -> 57.7Hurt
increase embedding dimensions to 512 for PolicyFlowPolicyFlow · IsaacLab H1 training timeper-iteration time (ms): 63.4 -> 115.5Hurt
add Brownian regularizer to PolicyFlow (ablation)PolicyFlow · MultiGoal / PointMaze explorationgoal coverageImproved
add Brownian regularizer on top of Gaussian entropy regularization (w_g=0.001) (ablation)PolicyFlow · MultiGoalmode coverageImproved
replace uniform noise injection (weight 0.05) with Brownian regularizer (ablation)PolicyFlow · MultiGoalmode coverageImproved
use smaller clipping range (epsilon 0.1 instead of 0.2) (ablation)PolicyFlow · IsaacLab ANYmal-Depisodic rewardHurt

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