讲座时间：2019.5.7
讲座地点：曹光彪信息楼313
报告人：王伟，Dr. Wang Wei is an assistant professor at School of Computing, National University of Singapore. His research interests include machine learning system optimization and multi-modal data analysis. He is the core developer of Apache SINGA project. He has been the reviewer of top conferences and journals like ACM Multimedia, VLDB, TKDE, and ACM transactions on Data Science. His papers have been nominated for the best paper award of ACM Multimedia 2015 and VLDB 2014.
报告内容：The life-cycle of developing a machine learning application includes multiple phases, including data preparation, model construction, training, inference and model adaption. Automated machine learning (a.k.a AutoML) tries to automate the whole process to make the life of developers easier. A lot of work has been done to optimize each stage of the life-cycle, especially on accuracy. In this talk, I am going to focus on the efficiency of AutoML. In particular, I will introduce a system developed at NUS, called Rafiki, to support AutoML. Rafiki provides the training and inference service for machine learning models. Specific efficiency optimization for distributed hyper-parameter tuning, model architecture search, and model adaption will be discussed.
