Reward-aware Exploration in Reinforcement Learning

Student: Meltzer Sergey

Supervisor: Boris Novikov

Faculty: St. Petersburg School of Physics, Mathematics, and Computer Science

Educational Programme: Enterprise Software Development (Master)

Year of Graduation: 2021

This paper explores the use of knowledge about the distribution of the reward function in the environment to better explore the environment by a reinforcement learning agent. To do this, we propose to use one possible way of obtaining a vector representation of the states of the environment, as well as Bayesian methods for approximating the distribution of the reward function. Studies are conducted on ten Atari game environments in comparison with existing Random Network Distillation and Intrinsic Curiosity Module algorithms. In half of the environments, the new Intrinsic Bayesian Module method demonstrates its superiority.

Student Theses at HSE must be completed in accordance with the University Rules and regulations specified by each educational programme.

Summaries of all theses must be published and made freely available on the HSE website.

The full text of a thesis can be published in open access on the HSE website only if the authoring student (copyright holder) agrees, or, if the thesis was written by a team of students, if all the co-authors (copyright holders) agree. After a thesis is published on the HSE website, it obtains the status of an online publication.

Student theses are objects of copyright and their use is subject to limitations in accordance with the Russian Federation’s law on intellectual property.

In the event that a thesis is quoted or otherwise used, reference to the author’s name and the source of quotation is required.

Search all student theses