AI Paper Atlas

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

Stabilizing the Q-Gradient Field for Policy Smoothness in Actor-Critic Methods

Jeong Woon Lee, Kyoleen Kwak, Daeho Kim, Hyoseok Hwang

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Actor-critic theory · 0 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 PAVE critic regularization (no actor change)TD3 · Hoppersmoothness score (lower better): 2.715 -> 0.950Improved
add PAVE critic regularizationTD3 · Walker2dcumulative return: 4834 -> 5563Improved
add PAVE critic regularizationSAC · Reachersmoothness score (lower better): 0.053 -> 0.052No clear effect
add PAVE critic regularizationSAC · Walker2dsmoothness score (lower better): 0.748 -> 0.584Improved
add Curvature Preservation loss to MPR + VFC (ablation)PAVE (TD3) · Walker2dcumulative return: 4648.2 -> 5562.9Improved
add VFC to MPR (ablation)PAVE (TD3) · LunarLandersmoothness score (lower better): 1.748 -> 1.177Improved
add MPR alone to base TD3 (ablation)TD3 · LunarLandersmoothness score (lower better): 1.809 -> 1.748No clear effect
add Curv to MPR + VFC (ablation)PAVE (TD3) · Walker2d (factorial)smoothness score (lower better): 2.095 -> 1.483Improved
add critic spectral normalization (ablation)TD3 critic · LunarLandercumulative return: 228 -> −970Hurt
add critic gradient penalty (ablation)TD3 critic · LunarLandersmoothness score (lower better): 1.81 -> 2.18Hurt
add actor-side CAPS on top of PAVE (ablation)PAVE (TD3) · LunarLandersmoothness score (lower better): 0.541 -> 0.543No clear effect
add PAVE to enable learningTD3 · Humanoid-v5cumulative return: 309 -> 2894Improved

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