Work

Similarity-based policy transfer for Multi-Task Reinforcement Learning

Year: 2024

Type: dissertation

Abstract: This thesis proposes a new Multi Task Reinforcement Learning approach to improve the learning of a new task by using previously acquired knowledge (source policies). This method leverage a common latent space for frames and task descriptions, using an Autoencoder and BERT transformer, to capture the similarity between source and target tasks via cosine similarity. The closest source policy is selected to guide the learning process of the target task. The experiments were carried out on two different game environments: Atari and Highway-env.

Source: Electronic Theses and Dissertations Repository (University of Pisa)

Author: CAPURSO, ALESSANDRO

Language: English

Topic: Reinforcement Learning in Robotics

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7071400690

Cites:

Cited by:

Related to: