AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

Agent Learning via Early Experience

Kai Zhang, Xiangchao Chen, Bo Liu, Tianci Xue, Zeyi Liao, Zhihan Liu, Xiyao Wang, Yuting Ning, Zhaorun Chen, Xiaohan Fu, Jian Xie, Yuxuan Sun and 18 more

ICML 2026 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Intrinsically motivated exploration · 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
add implicit world modeling stage before imitation learningLlama-3.2-3B agent · WebShopsuccess rate (%): 41.8 -> 60.2Improved
add self-reflection data to imitation learningLlama-3.2-3B agent · TravelPlannersuccess rate (%): 19.4 -> 32.2Improved
add self-reflection data to imitation learningLlama-3.1-8B agent · ScienceWorldsuccess rate (%): 54.7 -> 68.0Improved
add early experience (IWM or SR) to imitation learningLlama-3.2-3B agent · WebArena-Litesuccess rate (%): 6.1 -> 8.5Improved
add early experience (IWM or SR) to imitation learningLlama-3.2-3B agent · SearchQAF1: 38.0 -> 39.0No clear effect
add early experience to imitation learning, evaluated out-of-domainLlama-3.1-8B agent · ALFWorld (out-of-domain split)success rate (%): 63.3 -> 78.1Improved
initialize GRPO RL from early-experience checkpoints instead of imitation learning (ablation)WebShop, ALFWorld, SearchQA agents · post-RL performancesuccess rate / F1Improved
run GRPO directly from raw pretrained model with no supervised stage (ablation)GRPO baseline · WebShop, ALFWorld, SearchQApost-RL performanceHurt
use only 1/8 of expert demonstrations with early experience (ablation)Llama-3.1-8B agent · WebShopsuccess rate (%)Improved
increase branching factor K for implicit world modeling (ablation)IWM · WebShop and ALFWorldsuccess rate (%)Improved
increase branching factor K for self-reflection to large values (ablation)SR · WebShop and ALFWorldsuccess rate (%)Mixed
add test-time long chain-of-thought to imitation-learned model (ablation)Llama-3.1-8B IL · WebShopsuccess rate (%): 47.3 -> 0.0Hurt

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