TY - GEN

T1 - Provably efficient maximum entropy exploration

AU - Hazan, Elad

AU - Kakade, Sham M.

AU - Singh, Karan

AU - van Soest, Abby

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Suppose an agent is in a (possibly unknown) Markov Decision Process in the absence of a reward signal, what might we hope that an agent can efficiently learn to do? This work studies a broad class of objectives that are defined solely as functions of the state-visitation frequencies that are induced by how the agent behaves. For example, one natural, intrinsically defined, objective problem is for the agent to learn a policy which induces a distribution over state space that is as uniform as possible, which can be measured in an entropie sense. We provide an efficient algorithm to optimize such such intrinsically defined objectives, when given access to a black box planning oracle (which is robust to function approximation). Furthermore, when restricted to the tabular setting where we have sample based access to the MDP, our proposed algorithm is provably efficient, both in terms of its sample and computational complexities. Key to our algorithmic methodology is utilizing the conditional gradient method (a.k.a. the Frank-Wolfe algorithm) which utilizes an approximate MDP solver.

AB - Suppose an agent is in a (possibly unknown) Markov Decision Process in the absence of a reward signal, what might we hope that an agent can efficiently learn to do? This work studies a broad class of objectives that are defined solely as functions of the state-visitation frequencies that are induced by how the agent behaves. For example, one natural, intrinsically defined, objective problem is for the agent to learn a policy which induces a distribution over state space that is as uniform as possible, which can be measured in an entropie sense. We provide an efficient algorithm to optimize such such intrinsically defined objectives, when given access to a black box planning oracle (which is robust to function approximation). Furthermore, when restricted to the tabular setting where we have sample based access to the MDP, our proposed algorithm is provably efficient, both in terms of its sample and computational complexities. Key to our algorithmic methodology is utilizing the conditional gradient method (a.k.a. the Frank-Wolfe algorithm) which utilizes an approximate MDP solver.

UR - http://www.scopus.com/inward/record.url?scp=85078083938&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85078083938&partnerID=8YFLogxK

M3 - Conference contribution

T3 - 36th International Conference on Machine Learning, ICML 2019

SP - 4774

EP - 4786

BT - 36th International Conference on Machine Learning, ICML 2019

PB - International Machine Learning Society (IMLS)

T2 - 36th International Conference on Machine Learning, ICML 2019

Y2 - 9 June 2019 through 15 June 2019

ER -