arXiv cs.AI· Luca Viano, Antoine Moulin, Audrey Huang, Volkan Cevher, Philip Amortila, Dylan J. Foster·· 4 小时前AI 评分40
何时在线交互有助于模仿学习?基于价值的模仿学习中的表征权衡分析
When Does On-Policy Interaction Help? Representational Tradeoffs in Value-Based Imitation Learning
AI 导读
一项模仿学习理论研究指出,专家交互之所以能提升学习效果,是因为它放宽了对学习器的表征要求——只需能实现专家的价值函数,无需实现专家策略本身。论文提出交互式在线策略算法 OVI,在可表示专家价值函数时统计与计算上均高效;并证明仅凭专家价值可实现性假设,任何离线 IL 算法复杂度必然随专家策略类规模增长。实验中 OVI 优于 BC、DAgger 及离线价值类方法,学习器网络远弱于专家时增益最大。
正文 · 原文
Abstract:Imitation learning (IL)---training an agent to replicate expert behavior from demonstrations---underpins applications from robotics to language model training. Standard approaches such as Behavior Cloning (BC) are known to suffer from compounding errors and performance plateaus, particularly when the learner cannot perfectly represent the expert's policy (as is typical, e.g., in distillation). Two interventions are widely understood empirically to improve performance: querying the expert interactively along the learner's own trajectories, and using value function estimation en route to generating a policy.
We investigate the nature of these improvements and their potentially surprising interplay. Our main finding is that expert interaction relaxes the representational demands on the learner: one only needs a model capable of realizing the expert's value function, bypassing the (often stricter) requirement of realizing the expert's policy itself. Concretely, we introduce OVI, an interactive on-policy IL algorithm that is statistically and computationally efficient whenever the learner can represent the expert's value function. We complement this with a negative result showing that interaction is necessary. Namely, without stronger assumptions beyond expert-value realizability alone, any offline IL algorithm must scale with the complexity of the expert policy class. Our findings bear out empirically. OVI outperforms offline policy-based (BC), interactive policy-based (DAgger), and offline value-based IL methods, with the largest gains when the learner network is substantially less expressive than the expert's.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML) |
| Cite as: | arXiv:2607.29617 [cs.LG] |
| (or arXiv:2607.29617v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2607.29617 arXiv-issued DOI via DataCite |
Submission history
From: Antoine Moulin [view email]
[v1]
Fri, 31 Jul 2026 16:52:47 UTC (527 KB)
[v2]
Fri, 13 Nov 2026 07:28:38 UTC (559 KB)
来源:arXiv cs.AI · arxiv.org