Files

711649.pdf (1.95 MB)

Document type

Article

Version

Accepted version

Publication date

2021-09-01

All rights reserved

Please use this identifier to cite or link to this item: https://hdl.handle.net/2445/190660

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

Journal Title

Authors

Director/Tutor

Additional contribution

Journal ISSN

Volume Title

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."

Related resource

https://doi.org/10.1109/TPAMI.2021.3075372

Subject

Aprenentatge automàtic, Sistemes classificadors (Intel·ligència artificial), Xarxes neuronals (Informàtica)

Subject (English)

Machine learning, Learning classifier systems, Neural networks (Computer science)

Citation

Collections

Articles publicats en revistes (Matemàtiques i Informàtica)

Citation

LIU, Zhengying, et al. Winning solutions and post-challenge analyses of the ChaLearn AutoDL challenge 2019. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2021. Vol. 43, num. 9, pags. 3108-3125. ISSN 0162-8828. [consulted: 9 of October of 2026]. Available at: https://hdl.handle.net/2445/190660

Export metadata

JSON - METS

Share record