Leveraging class abstraction for commonsense reinforcement learning via residual policy gradient methods

Autor:	Höpner, Niklas, Tiddi, Ilaria, van Hoof, Herke
Rok vydání:	2022
Předmět:	Computer Science - Artificial Intelligence Computer Science - Machine Learning
Druh dokumentu:	Working Paper
Popis:	Enabling reinforcement learning (RL) agents to leverage a knowledge base while learning from experience promises to advance RL in knowledge intensive domains. However, it has proven difficult to leverage knowledge that is not manually tailored to the environment. We propose to use the subclass relationships present in open-source knowledge graphs to abstract away from specific objects. We develop a residual policy gradient method that is able to integrate knowledge across different abstraction levels in the class hierarchy. Our method results in improved sample efficiency and generalisation to unseen objects in commonsense games, but we also investigate failure modes, such as excessive noise in the extracted class knowledge or environments with little class structure.
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2201.12126 Zobrazit plný text záznamu View this record from Arxiv