AI Paper Atlas

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

Stochastic Minimum-Cost Reach-Avoid Reinforcement Learning

Jingduo Pan, Taoran Wu, Yiling Xue, Bai Xue

ICML 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Constrained policy optimization · 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
use RAPCPO instead of RC-PPO under same iteration budgetRAPCPO · PointGoalreach rate: 62.29% -> 78.49%Improved
use RAPCPO instead of RC-PPO under same iteration budgetRAPCPO · FixedWingreach rate: 73.98% -> 88.67%Improved
add compensation factor to enhanced Bellman objective (ablation)RAPCPO · PointGoal, FixedWing, Safety Hopper, Safety HalfCheetahadditional cumulative costImproved
add compensation factor to fixed-gamma Bellman estimate (ablation)reach-avoid probability estimate · FrozenLakeestimate calibrationImproved
use enhanced Bellman formulation instead of fixed-gamma (ablation)RAPCPO · Safety Hopperreach rate: 0.3218 -> 0.9434Improved
use enhanced Bellman formulation instead of fixed-gamma (ablation)RAPCPO · PointGoalreach rate: 0.4526 -> 0.7821Improved
set p to 0 (ablation)RAPCPO · Safety Hopperreach rateHurt
vary p within 0.1 to 0.7 (ablation)RAPCPO · Safety Hopperreach rate and costImproved
raise p to 0.8-1.0 (ablation)RAPCPO · Safety Hopperadditional cumulative costHurt
use CVaR-constrained methods (Saute, CPPO) with state-wise costsSaute, CPPO · deterministic and stochastic MuJoCo taskscost and reach rateHurt
use expectation-based constraints instead of CVaR constraintsCMDP baselines · deterministic and stochastic MuJoCo taskscost and reach rateImproved

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