Learning Policies for Neural Network Architecture Optimization Using Reinforcement Learning

Deep learning systems tend to be very sensitive to the specific network architecture both in terms of learning ability and performance of the learned solution. This, together with the difficulty of tuning neural network architectures leads to a need for automatic network optimization. Previous work...

Full description

Saved in:
Bibliographic Details
Main Authors: Raghav Vadhera, Manfred Huber
Format: Article
Language:English
Published: LibraryPress@UF 2023-05-01
Series:Proceedings of the International Florida Artificial Intelligence Research Society Conference
Online Access:https://journals.flvc.org/FLAIRS/article/view/133380
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849734593235648512
author Raghav Vadhera
Manfred Huber
author_facet Raghav Vadhera
Manfred Huber
author_sort Raghav Vadhera
collection DOAJ
description Deep learning systems tend to be very sensitive to the specific network architecture both in terms of learning ability and performance of the learned solution. This, together with the difficulty of tuning neural network architectures leads to a need for automatic network optimization. Previous work largely optimizes a network for one specific problem using architecture search, requiring significant amounts of time training different architectures during optimization. To address this and to open up the potential for transfer across tasks, this paper presents a novel approach that uses Reinforcement Learning to learn a policy for network optimization in a derived architecture embedding space that incrementally optimizes the network for the given problem. By utilizing policy learning and an abstract problem embedding, this approach brings the promise of transfer of the policy across problems and thus the potential optimization of networks for new problems without the need for excessive additional training. For an initial evaluation of the base capabilities, experiments for a standard classification problem are performed in this paper, showing the ability of the approach to optimize the architecture for a specific problem within a given rang of fully connected networks, and indicating its potential for learning effective policies to automatically improve network architectures.
format Article
id doaj-art-ab1d101e29734ddca8e91635ca4a1fe9
institution DOAJ
issn 2334-0754
2334-0762
language English
publishDate 2023-05-01
publisher LibraryPress@UF
record_format Article
series Proceedings of the International Florida Artificial Intelligence Research Society Conference
spelling doaj-art-ab1d101e29734ddca8e91635ca4a1fe92025-08-20T03:07:44ZengLibraryPress@UFProceedings of the International Florida Artificial Intelligence Research Society Conference2334-07542334-07622023-05-013610.32473/flairs.36.13338069686Learning Policies for Neural Network Architecture Optimization Using Reinforcement LearningRaghav Vadhera0Manfred Huber1https://orcid.org/0009-0007-0294-9147RAYTHEONUniversity of Texas at ArlingtonDeep learning systems tend to be very sensitive to the specific network architecture both in terms of learning ability and performance of the learned solution. This, together with the difficulty of tuning neural network architectures leads to a need for automatic network optimization. Previous work largely optimizes a network for one specific problem using architecture search, requiring significant amounts of time training different architectures during optimization. To address this and to open up the potential for transfer across tasks, this paper presents a novel approach that uses Reinforcement Learning to learn a policy for network optimization in a derived architecture embedding space that incrementally optimizes the network for the given problem. By utilizing policy learning and an abstract problem embedding, this approach brings the promise of transfer of the policy across problems and thus the potential optimization of networks for new problems without the need for excessive additional training. For an initial evaluation of the base capabilities, experiments for a standard classification problem are performed in this paper, showing the ability of the approach to optimize the architecture for a specific problem within a given rang of fully connected networks, and indicating its potential for learning effective policies to automatically improve network architectures.https://journals.flvc.org/FLAIRS/article/view/133380
spellingShingle Raghav Vadhera
Manfred Huber
Learning Policies for Neural Network Architecture Optimization Using Reinforcement Learning
Proceedings of the International Florida Artificial Intelligence Research Society Conference
title Learning Policies for Neural Network Architecture Optimization Using Reinforcement Learning
title_full Learning Policies for Neural Network Architecture Optimization Using Reinforcement Learning
title_fullStr Learning Policies for Neural Network Architecture Optimization Using Reinforcement Learning
title_full_unstemmed Learning Policies for Neural Network Architecture Optimization Using Reinforcement Learning
title_short Learning Policies for Neural Network Architecture Optimization Using Reinforcement Learning
title_sort learning policies for neural network architecture optimization using reinforcement learning
url https://journals.flvc.org/FLAIRS/article/view/133380
work_keys_str_mv AT raghavvadhera learningpoliciesforneuralnetworkarchitectureoptimizationusingreinforcementlearning
AT manfredhuber learningpoliciesforneuralnetworkarchitectureoptimizationusingreinforcementlearning