Mixture of Experts in a Mixture of RL settings

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Willi, Timon, Obando-Ceron, Johan, Foerster, Jakob, Dziugaite, Karolina, Castro, Pablo Samuel
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917706002857984
author Willi, Timon
Obando-Ceron, Johan
Foerster, Jakob
Dziugaite, Karolina
Castro, Pablo Samuel
author_facet Willi, Timon
Obando-Ceron, Johan
Foerster, Jakob
Dziugaite, Karolina
Castro, Pablo Samuel
contents Mixtures of Experts (MoEs) have gained prominence in (self-)supervised learning due to their enhanced inference efficiency, adaptability to distributed training, and modularity. Previous research has illustrated that MoEs can significantly boost Deep Reinforcement Learning (DRL) performance by expanding the network's parameter count while reducing dormant neurons, thereby enhancing the model's learning capacity and ability to deal with non-stationarity. In this work, we shed more light on MoEs' ability to deal with non-stationarity and investigate MoEs in DRL settings with "amplified" non-stationarity via multi-task training, providing further evidence that MoEs improve learning capacity. In contrast to previous work, our multi-task results allow us to better understand the underlying causes for the beneficial effect of MoE in DRL training, the impact of the various MoE components, and insights into how best to incorporate them in actor-critic-based DRL networks. Finally, we also confirm results from previous work.
format Preprint
id arxiv_https___arxiv_org_abs_2406_18420
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Mixture of Experts in a Mixture of RL settings
Willi, Timon
Obando-Ceron, Johan
Foerster, Jakob
Dziugaite, Karolina
Castro, Pablo Samuel
Machine Learning
Artificial Intelligence
Mixtures of Experts (MoEs) have gained prominence in (self-)supervised learning due to their enhanced inference efficiency, adaptability to distributed training, and modularity. Previous research has illustrated that MoEs can significantly boost Deep Reinforcement Learning (DRL) performance by expanding the network's parameter count while reducing dormant neurons, thereby enhancing the model's learning capacity and ability to deal with non-stationarity. In this work, we shed more light on MoEs' ability to deal with non-stationarity and investigate MoEs in DRL settings with "amplified" non-stationarity via multi-task training, providing further evidence that MoEs improve learning capacity. In contrast to previous work, our multi-task results allow us to better understand the underlying causes for the beneficial effect of MoE in DRL training, the impact of the various MoE components, and insights into how best to incorporate them in actor-critic-based DRL networks. Finally, we also confirm results from previous work.
title Mixture of Experts in a Mixture of RL settings
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2406.18420