AI Paper Atlas

Research map / Multi-agent systems and games / LLM agents

Hybrid Advantage Estimation with Unified Critic for VLM Agentic Reinforcement Learning

Wenxuan Zhang, Yuhui Wang, Donggang Jia, Xiaoqian Shen, Jian Ding, Ivan Viola, Jürgen Schmidhuber, Mohamed Elhoseiny

ECCV 2026 · Multi-agent systems and games › LLM agents · approach: Self-evolving language agent 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 turn-level advantage to token-level advantage (alpha 0 -> 0.5) (ablation)HyGAE · Sokobansuccess rate: 0.59 -> 0.83Improved
add token-level advantage to turn-level advantage (alpha 1 -> 0.5) (ablation)HyGAE · Sokobansuccess rate: 0.38 -> 0.83Improved
mix turn and token advantage (alpha 0 -> 0.5) (ablation)HyGAE · FrozenLakesuccess rate: 0.72 -> 0.80Improved
mix turn and token advantage (alpha 1 -> 0.5) (ablation)HyGAE · FrozenLakesuccess rate: 0.70 -> 0.80Improved
mix turn and token advantage (alpha 0 -> 0.5) (ablation)HyGAE · Primitive Skill (avg.)success rate: 0.98 -> 1.00Improved
mix turn and token advantage (alpha 1 -> 0.5) (ablation)HyGAE · Primitive Skill (avg.)success rate: 0.25 -> 1.00Improved
use unbalanced mixing alpha 0.25 or 0.75 instead of 0.5 (ablation)HyGAE · Primitive Skill (avg.)success rate: 1.00 -> 0.73Hurt
replace PPO with hybrid advantage and unified critic (HyGAE)PPO · Sokobansuccess rate: 0.5797 ± 0.2105 -> 0.8222 ± 0.0711Improved
use hybrid advantage and unified critic instead of best token-wise baselineToken-PPO · Average over five benchmarkssuccess rate: 0.81 -> 0.91Improved
broadcast turn-level return uniformly across tokens, or drop the value model (REINFORCE), instead of the mixed return (ablation)HyGAE · Sokobantraining stabilityHurt
aggregate small token-level rewards (KL penalties) into the turn reward (ablation)HyGAE · Sokobansuccess rateNo clear effect
use turn-level value as constant baseline instead of token-specific value for broadcast turn advantage (ablation)HyGAE · Sokobansuccess rateNo clear effect

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