AI Paper Atlas

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

Scalable Multi-Objective and Meta Reinforcement Learning via Gradient Estimation

Zhenshuo Zhang, Minxuan Duan, Youran Ye, Hongyang R. Zhang

AAAI 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Variance-reduced policy gradients · 7 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 affinity-based grouping from first-order gradient estimation instead of random grouping (ablation)soft modularization on MT10 · Meta-World MT10average success rate (%): 58.2 -> 94.0Improved
use affinity-based grouping instead of gradient-similarity grouping (ablation)soft modularization on MT10 · Meta-World MT10average success rate (%): 69.6 -> 94.0Improved
group tasks with PolicyGradEx rather than training one multitask policysoft modularization on MT10 · Meta-World MT10average success rate (%): 71.3 -> 94.0Improved
select meta-training tasks via PolicyGradEx groups (MAML)MAML on CartPole · CartPolereward after adaptation: 145.9 -> 159.2Improved
select meta-training tasks via PolicyGradEx groups instead of gradient-similarity (ablation)MAML on LunarLander · LunarLanderreward after adaptation: 80.8 -> 82.8No clear effect
use 3 groups instead of 2 (ablation)PolicyGradEx · Meta-World MT10average success rate (%): 89.5 -> 94.0Improved
use 4 groups instead of 3 (ablation)PolicyGradEx · Meta-World MT10average success rate (%): 94.0 -> 95.1No clear effect
increase random projection dimension beyond 400 (ablation)PolicyGradEx · Meta-World / controlclustering qualityNo clear effect
increase number of groups k beyond 3 (ablation)PolicyGradEx · control environmentssuccess rateNo clear effect
increase network depth from 2 to 8 MLP layers (ablation)PolicyGradEx · Meta-World / LunarLanderFLOP speedup: 21x -> 26xImproved
let policy parameters drift farther from meta-initialization (ablation)first-order approximation · RL environmentsapproximation error: <2% -> up to 10%Hurt
increase subset size alpha in Meta-World groups (ablation)multitask policy · Meta-Worldgeneralization errorMixed

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