MisoDICE: Multi-Agent Imitation from Unlabeled Mixed-Quality Demonstrations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bui, The Viet, Mai, Tien, Nguyen, Hong Thanh
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915303339851776
author Bui, The Viet
Mai, Tien
Nguyen, Hong Thanh
author_facet Bui, The Viet
Mai, Tien
Nguyen, Hong Thanh
contents We study offline imitation learning (IL) in cooperative multi-agent settings, where demonstrations have unlabeled mixed quality - containing both expert and suboptimal trajectories. Our proposed solution is structured in two stages: trajectory labeling and multi-agent imitation learning, designed jointly to enable effective learning from heterogeneous, unlabeled data. In the first stage, we combine advances in large language models and preference-based reinforcement learning to construct a progressive labeling pipeline that distinguishes expert-quality trajectories. In the second stage, we introduce MisoDICE, a novel multi-agent IL algorithm that leverages these labels to learn robust policies while addressing the computational complexity of large joint state-action spaces. By extending the popular single-agent DICE framework to multi-agent settings with a new value decomposition and mixing architecture, our method yields a convex policy optimization objective and ensures consistency between global and local policies. We evaluate MisoDICE on multiple standard multi-agent RL benchmarks and demonstrate superior performance, especially when expert data is scarce.
format Preprint
id arxiv_https___arxiv_org_abs_2505_18595
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MisoDICE: Multi-Agent Imitation from Unlabeled Mixed-Quality Demonstrations
Bui, The Viet
Mai, Tien
Nguyen, Hong Thanh
Machine Learning
Artificial Intelligence
Multiagent Systems
We study offline imitation learning (IL) in cooperative multi-agent settings, where demonstrations have unlabeled mixed quality - containing both expert and suboptimal trajectories. Our proposed solution is structured in two stages: trajectory labeling and multi-agent imitation learning, designed jointly to enable effective learning from heterogeneous, unlabeled data. In the first stage, we combine advances in large language models and preference-based reinforcement learning to construct a progressive labeling pipeline that distinguishes expert-quality trajectories. In the second stage, we introduce MisoDICE, a novel multi-agent IL algorithm that leverages these labels to learn robust policies while addressing the computational complexity of large joint state-action spaces. By extending the popular single-agent DICE framework to multi-agent settings with a new value decomposition and mixing architecture, our method yields a convex policy optimization objective and ensures consistency between global and local policies. We evaluate MisoDICE on multiple standard multi-agent RL benchmarks and demonstrate superior performance, especially when expert data is scarce.
title MisoDICE: Multi-Agent Imitation from Unlabeled Mixed-Quality Demonstrations
topic Machine Learning
Artificial Intelligence
Multiagent Systems
url https://arxiv.org/abs/2505.18595