We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Statistics > Machine Learning

Title: A Bayesian Approach to Policy Recognition and State Representation Learning

Abstract: Learning from demonstration (LfD) is the process of building behavioral models of a task from demonstrations provided by an expert. These models can be used e.g. for system control by generalizing the expert demonstrations to previously unencountered situations. Most LfD methods, however, make strong assumptions about the expert behavior, e.g. they assume the existence of a deterministic optimal ground truth policy or require direct monitoring of the expert's controls, which limits their practical use as part of a general system identification framework. In this work, we consider the LfD problem in a more general setting where we allow for arbitrary stochastic expert policies, without reasoning about the optimality of the demonstrations. Following a Bayesian methodology, we model the full posterior distribution of possible expert controllers that explain the provided demonstration data. Moreover, we show that our methodology can be applied in a nonparametric context to infer the complexity of the state representation used by the expert, and to learn task-appropriate partitionings of the system state space.
Comments: 17 pages, 8 figures; ### Version 4 ### to appear in IEEE Transactions on Pattern Analysis and Machine Intelligence
Subjects: Machine Learning (stat.ML); Machine Learning (cs.LG); Systems and Control (eess.SY); Dynamical Systems (math.DS); Probability (math.PR)
DOI: 10.1109/TPAMI.2017.2711024
Cite as: arXiv:1605.01278 [stat.ML]
  (or arXiv:1605.01278v4 [stat.ML] for this version)

Submission history

From: Adrian Šošić [view email]
[v1] Wed, 4 May 2016 13:44:53 GMT (711kb,D)
[v2] Mon, 30 May 2016 15:05:59 GMT (711kb,D)
[v3] Fri, 19 May 2017 14:13:55 GMT (1485kb,D)
[v4] Fri, 4 Aug 2017 12:50:01 GMT (2469kb,D)

Link back to: arXiv, form interface, contact.