Paper Detail
Chenghao Li, Jun Liu, Songbo Zhang, Huadong Jian, Hao Ni, Lik-Hang Lee, Sung-Ho Bae, Guoqing Wang, Yang Yang, Chaoning Zhang
Multimodal LLM agents operating in complex game environments must continually reuse past experience to solve new tasks efficiently. In this work, we propose Echo, a transfer-oriented memory framework that enables agents to derive actionable knowledge from prior interactions rather than treating memory as a passive repository of static records. To make transfer explicit, Echo decomposes reusable knowledge into five dimensions: structure, attribute, process, function, and interaction. This formulation allows the agent to identify recurring patterns shared across different tasks and infer what prior experience remains applicable in new situations. Building on this formulation, Echo leverages In-Context Analogy Learning (ICAL) to retrieve relevant experiences and adapt them to unseen tasks through contextual examples. Experiments in Minecraft show that, under a from-scratch learning setting, Echo achieves a 1.3x to 1.7x speed-up on object-unlocking tasks. Moreover, Echo exhibits a burst-like chain-unlocking phenomenon, rapidly unlocking multiple similar items within a short time interval after acquiring transferable experience. These results suggest that experience transfer is a promising direction for improving the efficiency and adaptability of multimodal LLM agents in complex interactive environments.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@misc{li2026experience,
title = {Experience Transfer for Multimodal LLM Agents in Minecraft Game},
author = {Chenghao Li and Jun Liu and Songbo Zhang and Huadong Jian and Hao Ni and Lik-Hang Lee and Sung-Ho Bae and Guoqing Wang and Yang Yang and Chaoning Zhang},
year = {2026},
abstract = {Multimodal LLM agents operating in complex game environments must continually reuse past experience to solve new tasks efficiently. In this work, we propose Echo, a transfer-oriented memory framework that enables agents to derive actionable knowledge from prior interactions rather than treating memory as a passive repository of static records. To make transfer explicit, Echo decomposes reusable knowledge into five dimensions: structure, attribute, process, function, and interaction. This formula},
url = {https://huggingface.co/papers/2604.05533},
keywords = {multimodal LLM agents, transfer-oriented memory framework, In-Context Analogy Learning, knowledge decomposition, experience transfer, recurrent patterns, contextual examples, Minecraft, object-unlocking tasks, chain-unlocking phenomenon, huggingface daily},
eprint = {2604.05533},
archiveprefix = {arXiv},
}
{}