arXiv：cs.LG· Ziyad Sheebaelhamd, Luca Viano, Volkan Cevher, Claire Vernade·· 7 小时前
MA-BC：可证明高效的多目标模仿学习算法
Split the Differences, Pool the Rest: Provably Efficient Multi-Objective Imitation
AI 导读
研究提出 MA-BC（Multi-Output Augmented Behavioral Cloning），通过拆分冲突专家数据、合并无冲突状态-动作对，从多个 Pareto 最优专家的演示中恢复 Pareto 前沿策略。
正文
Abstract:This work investigates multi-objective imitation learning: the problem of recovering policies that lie on the Pareto front given demonstrations from multiple Pareto-optimal experts in a Multi-Objective Markov Decision Process (MOMDP). Standard imitation approaches are ill-equipped for this regime, as naively aggregating conflicting expert trajectories can result in dominated policies. To address this, we introduce Multi-Output Augmented Behavioral Cloning (MA-BC), an algorithm that systematically partitions divergent expert data while pooling state-action pairs where no behavior conflict is observed. Theoretically, we prove that MA-BC converges to Pareto-optimal policies at a faster statistical rate than any learner that considers each expert dataset independently. Furthermore, we establish a novel lower bound for multi-objective imitation learning, demonstrating that MA-BC is minimax optimal. Finally, we empirically validate our algorithm across diverse discrete environments and, guided by our theoretical insights, extend and evaluate MA-BC on a continuous Linear Quadratic Regulator (LQR) control task.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.12000 [cs.LG] |
| (or arXiv:2605.12000v3 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2605.12000 arXiv-issued DOI via DataCite |
Submission history
From: Ziyad Sheebaelhamd [view email]
[v1]
Tue, 12 May 2026 11:49:08 UTC (1,968 KB)
[v2]
Mon, 18 May 2026 14:32:07 UTC (1,969 KB)
[v3]
Wed, 14 Oct 2026 10:51:06 UTC (1,977 KB)
来源：arXiv：cs.LG · arxiv.org