On The Transferability of Deep-Q Networks

TL;DR

This paper investigates the transferability of Deep-Q Networks in Deep Reinforcement Learning, revealing that transfer often leads to negative results and providing insights into the training dynamics affecting transfer success.

Contribution

It systematically evaluates transferability of Deep-Q Networks across benchmarks and control tasks, highlighting challenges and offering new understanding of training dynamics in DRL.

Findings

Transfer in DRL often results in negative transfer.

Deep-Q Networks exhibit poor transferability across tasks.

Insights into training dynamics explain transfer difficulties.

Abstract

Transfer Learning (TL) is an efficient machine learning paradigm that allows overcoming some of the hurdles that characterize the successful training of deep neural networks, ranging from long training times to the needs of large datasets. While exploiting TL is a well established and successful training practice in Supervised Learning (SL), its applicability in Deep Reinforcement Learning (DRL) is rarer. In this paper, we study the level of transferability of three different variants of Deep-Q Networks on popular DRL benchmarks as well as on a set of novel, carefully designed control tasks. Our results show that transferring neural networks in a DRL context can be particularly challenging and is a process which in most cases results in negative transfer. In the attempt of understanding why Deep-Q Networks transfer so poorly, we gain novel insights into the training dynamics that characterizes this family of algorithms.