AutoHAS: Efficient Hyperparameter and Architecture Search

TL;DR

AutoHAS introduces a unified, efficient method for simultaneously searching optimal hyperparameters and architectures using reinforcement learning, improving model accuracy across various datasets and architectures.

Contribution

A novel unified pipeline that efficiently searches for both hyperparameters and architectures using an RL controller and shared weights.

Findings

AutoHAS improves accuracy on ResNet and EfficientNet.

It is efficient and generalizable across different search spaces and datasets.

AutoHAS outperforms baseline methods in hyperparameter and architecture search.

Abstract

Efficient hyperparameter or architecture search methods have shown remarkable results, but each of them is only applicable to searching for either hyperparameters (HPs) or architectures. In this work, we propose a unified pipeline, AutoHAS, to efficiently search for both architectures and hyperparameters. AutoHAS learns to alternately update the shared network weights and a reinforcement learning (RL) controller, which learns the probability distribution for the architecture candidates and HP candidates. A temporary weight is introduced to store the updated weight from the selected HPs (by the controller), and a validation accuracy based on this temporary weight serves as a reward to update the controller. In experiments, we show AutoHAS is efficient and generalizable to different search spaces, baselines and datasets. In particular, AutoHAS can improve the accuracy over popular network architectures, such as ResNet and EfficientNet, on CIFAR-10/100, ImageNet, and four more other datasets.