Winning solutions and post-challenge analyses of the ChaLearn AutoDL challenge 2019

TL;DR

This paper analyzes the results of the 2019 ChaLearn AutoDL challenge, highlighting the dominance of deep learning methods, the modular solution architecture, and the importance of meta-learning and ensembling, while providing benchmarks and open-source tools.

Contribution

It provides a comprehensive post-challenge analysis, emphasizing the modular approach and key techniques that led to successful AutoDL solutions, and establishes benchmarks and open resources.

Findings

Deep learning dominated solutions.

Modular architecture enabled ablation studies.

Meta-learning and ensembling were crucial for performance.

Abstract

This paper reports the results and post-challenge analyses of ChaLearn's AutoDL challenge series, which helped sorting out a profusion of AutoML solutions for Deep Learning (DL) that had been introduced in a variety of settings, but lacked fair comparisons. All input data modalities (time series, images, videos, text, tabular) were formatted as tensors and all tasks were multi-label classification problems. Code submissions were executed on hidden tasks, with limited time and computational resources, pushing solutions that get results quickly. In this setting, DL methods dominated, though popular Neural Architecture Search (NAS) was impractical. Solutions relied on fine-tuned pre-trained networks, with architectures matching data modality. Post-challenge tests did not reveal improvements beyond the imposed time limit. While no component is particularly original or novel, a high level modular organization emerged featuring a "meta-learner", "data ingestor", "model selector", "model/learner", and "evaluator". This modularity enabled ablation studies, which revealed the importance of (off-platform) meta-learning, ensembling, and efficient data management. Experiments on heterogeneous module combinations further confirm the (local) optimality of the winning solutions. Our challenge legacy includes an ever-lasting benchmark (http://autodl.chalearn.org), the open-sourced code of the winners, and a free "AutoDL self-service".