Reinforcement Learning and Inverse Reinforcement Learning with System 1 and System 2

@article{Peysakhovich2019ReinforcementLA,
  title={Reinforcement Learning and Inverse Reinforcement Learning with System 1 and System 2},
  author={Alexander Peysakhovich},
  journal={Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society},
  year={2019}
}
  • Alexander Peysakhovich
  • Published 2019
  • Computer Science, Mathematics
  • Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society
  • Inferring a person's goal from their behavior is an important problem in applications of AI (e.g. automated assistants, recommender systems). The workhorse model for this task is the rational actor model - this amounts to assuming that people have stable reward functions, discount the future exponentially, and construct optimal plans. Under the rational actor assumption techniques such as inverse reinforcement learning (IRL) can be used to infer a person's goals from their actions. A competing… CONTINUE READING
    5 Citations

    References

    SHOWING 1-4 OF 4 REFERENCES
    Maximum Entropy Inverse Reinforcement Learning
    • 1,517
    • Highly Influential
    • PDF
    Planning Complexity Registers as a Cost in Metacontrol
    • 15
    • Highly Influential
    • PDF
    Doing It Now or Later
    • 2,744
    • Highly Influential
    • PDF
    A Dual-Self Model of Impulse Control.
    • 944
    • Highly Influential
    • PDF