Since the dawn of machine learning, the dream of many researchers has been to create the “perfect black box”: a learning machine capable of learning from examples without any human intervention? How far have we got in 2015? This is the question raised by the AutoML challenge, running from December 2014 to May 2015.
We are progressively introducing 30 classification and regression tasks, with datasets pre-formatted in feature representations. There is a broad diversity of data types and distributions and the problems are drawn from a wide variety of domains and include medical diagnosis from laboratory analyses, speech recognition, credit rating, prediction or drug toxicity or efficacy, classification of text, prediction of customer satisfaction, object recognition, protein structure prediction, action recognition in video data, etc. While there exist machine learning toolkits including methods that can solve all these problems, it is still considerable human effort to find, for a given combination of dataset, task, metric of evaluation, and available computational time, the combination of methods and hyper-parameter setting that is best suited.
This presentation will explain the protocol of the challenge and give pointers to get started. There is a prize pool of $30,000 donated by Microsoft.