AI Paper Atlas

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

Learning Reward–Cost Balance in Safe RL via Score-Based World Models

Yuetian Wang, Dianxi Shi, Yuanze Wang, Huanhuan Yang, Shiming Song, Chunping Qiu

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Safety layers and shielding

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
remove score distillation into the world model (query external ScoreNet instead) (ablation)USB-RL · PointButton1 (ablation)rewardImproved
use within-root pair construction instead of uniformly random pairs for ScoreNet (ablation)USB-RL · PointButton1 (ablation)reward and costImproved
add ScoreCritic temporal score propagation (ablation)USB-RL · PointButton1 (ablation)reward and costImproved
replace linear reward-cost penalty with learned score rankingSafeDreamer · Safety-Gymnasium (PointGoal2, PointButton1, PointPush1, CarGoal1)average episode reward and costImproved
replace linear reward-cost penalty with learned score rankingSafeDreamer · MetaDrivearrival rate and episode cost 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