2019

Inverse Reinforcement Learning with Multiple Ranked Experts

Castro, Pablo Samuel, Li, Shijian, Zhang, Daqing

Understand

We consider the problem of learning to behave optimally in a Markov Decision Process when a reward function is not specified, but instead we have access to a set of demonstrators of varying performance.

  • We assume the demonstrators are classified into one of k ranks, and use ideas from ordinal regression to find a reward function that maximizes the margin between the different ranks.
  • This approach is based on the idea that agents should not only learn how to behave from experts, but also how not to behave from non-experts.
  • We show there are MDPs where important differences in the reward function would be hidden from existing algorithms by the behaviour of the expert.

Built on

  • Extrapolating beyond suboptimal demonstrations via inverse reinforcement learning from observations

    Original

    Brown, D. S., Goo, W., Nagarajan, P., and Niekum, S. (2019) · 1904

    Earlier work this paper cites.

  • ALVINN: An Autonomous Land Vehicle in a Neural Network

    Pomerleau, D. (1989) · 1989

    Earlier work this paper cites.

  • A robot controller using learning by imitation

    Hayes, G. and Demiris, J. (1994) · 1994

    Earlier work this paper cites.

  • Uncovering cabdrivers’ behavior patterns from their digital traces

    Liu, L., Andris, C., and Ratti, C. (1994) · 1994

    Earlier work this paper cites.

  • Markov Decision Processes: Discrete Stochastic Dynamic Programming

    Puterman, M. L. (1994) · 1994

    Earlier work this paper cites.

  • Introduction to Reinforcement Learning

    Sutton, R. S. and Barto, A. G. (1998) · 1998

    Earlier work this paper cites.

Similar

  • Algorithms for inverse reinforcement learning

    Ng, A. Y. and Russell, S. (2000) · 2000

    Cited alongside, same era.

  • Learning movement sequences from demonstration

    Amit, R. and Mataric, M. (2002) · 2002

    Cited alongside, same era.

  • Ranking with Large Margin Principle: Two Approaches

    Shashua, A. and Levin, A. (2002) · 2002

    Cited alongside, same era.

  • Apprenticeship learning via inverse reinforcement learning

    Abbeel, P. and Ng, A. Y. (2004) · 2004

    Cited alongside, same era.

  • Maximum margin planning

    Ratliff, N. D., Bagnell, J. A., and Zinkevich, M. A. (2006) · 2006

    Cited alongside, same era.

Then

  • Maximum entropy inverse reinforcement learning

    Ziebart, B. D., Maas, A., Bagnell, J. A., and Dey, A. K. (2008) · 2008

    Later among the works it cites.

  • Taxi-Aware Map: Identifying and predicting vacant taxis in the city

    Phithakkitnukoon, S., Veloso, M., Biderman, A., Bento, C., and Ratti, C. (2010) · 2010

    Later among the works it cites.

  • Hunting or waiting? discovering passenger-finding strategies from a large-scale real-world taxi dataset

    Li, B., Zhang, D., Sun, L., Chen, C., Li, S., Qi, G., and Yang, Q. (2011) · 2011

    Later among the works it cites.

  • Waiting/cruising location recommendation for efficient taxi business

    Takayama, T., Matsumoto, K., and Kumagai, A. (2011) · 2011

    Later among the works it cites.

  • Efficient probabilistic performance bounds for inverse reinforcement learning

    Brown, D. S. and Niekum, S. (2018) · 2018

    Later among the works it cites.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…