Machine Learning is a powerful and effective tool, but only if hyperparameters are well tuned to the problem at hand. This tuning process is difficult for non-experts and can be very costly. This talk will introduce the background of automated machine learning (AutoML) and hyperparameter tuning, the challenges in practice, and the research done in FLAML to address some of the challenges. It will present a new cost-effective hyperparameter optimization method with theoretical guarantee, and the AutoML system built based on it with superior empirical performance and easy to use. Example use cases will be discussed, as well as open problems.
Please email for a
Chi Wang is a principal researcher in Microsoft Research at Redmond. He has worked on automated machine learning, machine learning for systems, scalable solutions for data science and data analytics, and knowledge mining from text data and graph data (with a SIGKDD Data Science/Data Mining PhD Dissertation Award). Chi is the creator of FLAML, a fast open source library for AutoML & tuning used widely inside and outside Microsoft.