Learning to Achieve Goals

  title={Learning to Achieve Goals},
  author={Leslie Pack Kaelbling},
Temporal diierence methods solve the temporal credit assignment problem for reinforcement learning. An important subproblem of general reinforcement learning is learning to achieve dynamic goals. Although existing temporal diierence methods, such as Q learning, can be applied to this problem, they do not take advantage of its special structure. This paper presents the DG-learning algorithm, which learns eeciently to achieve dynamically changing goals and exhibits good knowledge transfer between… CONTINUE READING


Publications referenced by this paper.
Showing 1-7 of 7 references

A role for anticipation in reactive systems that learn

Steven D. Whiteheadand Dana H. Ballard
Proceedings of the SixthInternational Workshop on Machine Learning • 1989
View 2 Excerpts
Highly Influenced

Learning to predict bythe method of temporal di erences

Richard S. Sutton.

The role of explo- ration in learning control

Richard S. Sutton
Seventh International Conference on Machine • 1988
View 2 Excerpts

Howard . Dynamic Programming and Markov Processes

A. Ronald

Thrun . The role of exploration in learning control

B. Sebastian

Similar Papers

Loading similar papers…