AI Paper Atlas

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

Enhancing Control Policy Smoothness by Aligning Actions with Predictions from Preceding States

Kyoleen Kwak, Hyoseok Hwang

AAAI 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Regularized scalable deep value learning · 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 ASAP loss (transition-induced spatial term plus second-order temporal term)PPO · Gymnasium Hoppersmoothness score: 1.709 -> 0.179Improved
add ASAP lossPPO · Gymnasium Walkercumulative return: 2654 -> 3128Improved
add ASAP lossSAC · Gymnasium Walkersmoothness score: 0.823 -> 0.578Improved
add ASAP lossPPO · Isaac-Lab Reach-Frankasmoothness score: 0.959 -> 0.658Improved
add ASAP lossPPO · Isaac-Lab Lift-Cube-Frankacumulative return: 136.1 -> 134.0No clear effect
add ASAP lossPPO · Isaac-Lab Repose-Cube-Allegrocumulative return: 31.90 -> 32.82No clear effect
add ASAP loss on top of LipsNetSAC LipsNet · Hoppersmoothness score: 0.838 -> 0.381Improved
replace ASAP spatial term with CAPS spatial term (GRAD temporal fixed) (ablation)PPO · Walkercumulative return: 3128 -> 2303Hurt
replace ASAP spatial term with L2C2 spatial term (GRAD temporal fixed) (ablation)PPO · Hoppersmoothness score: 0.179 -> 0.227Hurt
add ASAP spatial term to GRAD temporal term (ablation)PPO · Walkersmoothness score: 0.541 -> 0.345Improved
add ASAP spatial term to CAPS temporal term (ablation)PPO · Walkersmoothness score: 0.872 -> 0.389Improved
add ASAP spatial term to GRAD temporal term (ablation)PPO · Hoppercumulative return: 2963 -> 2691Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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