AutoCLINT: The Winning Method in AutoCV Challenge 2019
This work addresses the need for efficient and automated machine learning methods in visual domains, but it is incremental as it builds on existing techniques like Fast AutoAugment.
The authors tackled the problem of automated machine learning for visual classification tasks by introducing AutoCLINT, which won the AutoCV Challenge 2019 by implementing an autonomous training strategy with efficient code optimization and automated data augmentation, achieving fast adaptation of pretrained networks.
NeurIPS 2019 AutoDL challenge is a series of six automated machine learning competitions. Particularly, AutoCV challenges mainly focused on classification tasks on visual domain. In this paper, we introduce the winning method in the competition, AutoCLINT. The proposed method implements an autonomous training strategy, including efficient code optimization, and applies an automated data augmentation to achieve the fast adaptation of pretrained networks. We implement a light version of Fast AutoAugment to search for data augmentation policies efficiently for the arbitrarily given image domains. We also empirically analyze the components of the proposed method and provide ablation studies focusing on AutoCV datasets.