arXiv：cs.AI· Mark K. Ho, Michael L. Littman, James MacGlashan, Fiery Cushman, Joseph L. Austerweil·· 4 小时前
演示即教学：IRL 从“做”与“教”中学习的差异
Showing versus Doing: Teaching by Demonstration
AI 导读
研究者提出一个面向“示范教学”的贝叶斯模型，区分演示者有意教学（showing）与单纯执行任务（doing）两种情形。两项人类实验显示，参与者的教学行为会按该模型预测的方式改变。标准 IRL 算法在从 showing（而非 doing）中学习时表现也能获得提升。
正文
Abstract:People often learn from others' demonstrations, and inverse reinforcement learning (IRL) techniques have realized this capacity in machines. In contrast, teaching by demonstration has been less well studied computationally. Here, we develop a Bayesian model for teaching by demonstration. Stark differences arise when demonstrators are intentionally teaching (i.e. showing) a task versus simply performing (i.e. doing) a task. In two experiments, we show that human participants modify their teaching behavior consistent with the predictions of our model. Further, we show that even standard IRL algorithms benefit when learning from showing versus doing.
| Comments: | Published at NeurIPS 2016 (oral presentation). 12 pages, including supplementary materials |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC) |
| Cite as: | arXiv:2610.14004 [cs.LG] |
| (or arXiv:2610.14004v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.14004 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | Advances in Neural Information Processing Systems 29 (NIPS 2016) |
Submission history
From: Mark Ho [view email]
[v1]
Sun, 27 Sep 2026 17:58:27 UTC (2,189 KB)
来源：arXiv：cs.AI · arxiv.org