Go back to publications View external

Towards Automated Deep Learning: Analysis of the AutoDL Challenge Series 2019

June 2020 Zhengying Liu, Zhen Xu, Shangeth Rajaa, Meysam Madadi, Julio C. S. Jacques Junior, Sergio Escalera, Adrien Pavao, Sebastien Treguer, Wei-Wei Tu, Isabelle Guyon Proceedings of Machine Learning Research (PMLR), NeurIPS 2019 Competition Track, Vol. 123, pp. 242–252
#AutoML #Deep Learning #NeurIPS

Abstract

We present the design and results of recent competitions in Automated Deep Learning (AutoDL). In the AutoDL challenge series 2019, we organized 5 machine learning challenges: AutoCV, AutoCV2, AutoNLP, AutoSpeech and AutoDL, covering computer vision, natural language processing, and speech recognition domains.

Key contributions include a benchmark suite of baseline AutoML solutions and a repository of around 100 datasets for meta-learning research. Results demonstrate that winning solutions generalize to new unseen datasets, supporting development toward universal AutoML approaches.