Beyond Simply Environment Scaling: Designing Effective Environment Distributions for Multimodal Agent Learning

Kejian Zhu; Zhuoran Jin; Dongqi Huang; Hongbang Yuan; Yupu Hao; Kang Liu; Jun Zhao

cs.CVAug 4, 2026

Beyond Simply Environment Scaling: Designing Effective Environment Distributions for Multimodal Agent Learning

Authors: Kejian Zhu, Zhuoran Jin, Dongqi Huang, Hongbang Yuan, Yupu Hao, Kang Liu, Jun Zhao

Organizations: The Key Laboratory of Cognition and Decision Intelligence for Complex Systems, Institute of Automation, Chinese Academy of Sciences, Beijing, China · School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China

Abstract

Recent works train agents by constructing large-scale multimodal environment pools. However, we find that simply increasing the number of multimodal environments does not always benefit. We further analyze the limitations in current multimodal environment distributions through a series of experiments. Based on these findings, we study how to build more effective training environment distributions from two dimensions: diversity and difficulty structure. For diversity, we propose Ability-aware Environment Selection (AES) to obtain diverse environment sets. For difficulty structure, we propose Hierarchical Difficulty Curriculum (HDC), which organizes curriculum learning through two difficulty levels: harness weakening and state-scale progression. Experiments show that AES and HDC effectively improve multimodal agent training.

Explore similar work

Scalable Environments Drive Generalizable Agents
May 18, 2026Jiayi Zhang, Fanqi Kong, Guibin Zhang +7Agentic Control Scalability
Multi-Modal, Multi-Environment Machine Teaching for Robust Reward Learning
Jul 9, 2026Ali Larian, Qian Lin, Chang Zong Wu +1Heterogeneous Modalities
Environment Evolution for Terminal Agents
Sep 3, 2026Zhiyuan Fan, Tinghao Yu, Yuanjun Cai +9Self-Evolving Agents Evolution

May 18, 2026 · cs.AIJ/K move · Enter open · S save

Scalable Environments Drive Generalizable Agents

Jiayi Zhang, Fanqi Kong, Guibin Zhang, Maojia Song+6

HKUST(GZ) · DeepWisdom · PKU · NUS · SUTD

Generalizable agents should adapt to diverse tasks and unseen environments beyond their training distribution. This position paper argues that such generalization requires environment scaling: expanding the distribution of executable rule-sets that agents interact with, rather than only increasing trajectories or tasks within fixed benchmarks. Current scaling practices largely focus on collecting more experience or broader task sets under fixed interaction rules, leaving agents brittle when underlying interfaces, dynamics, observations, or feedback signals change. The core challenge is therefore a world-level distribution shift: agents need systematic exposure to environments with meaningfully different executable rule-sets. To clarify this challenge, we propose a unified taxonomy that separates trajectory scaling, task scaling, and environment scaling by their primary deliverables and by what changes in the executable rule-set. Building on this taxonomy, we synthesize construction paradigms for scalable environments, contrasting programmatic generators that prioritize controllability and verifiability with generative world models that offer broader coverage and open-endedness. We further outline how environment scaling can be coupled with stateful learning mechanisms, emphasizing learned update rules for cross-environment adaptation. We conclude by discussing alternative perspectives and argue that scalable environments provide the essential substrate for measurable and controllable progress toward robust general agents.

Agentic Control Scalability Environment Trajectory Prediction

Open paper Ask about it PDF

Abstract

Explore similar work

Scalable Environments Drive Generalizable Agents

Multi-Modal, Multi-Environment Machine Teaching for Robust Reward Learning

Environment Evolution for Terminal Agents