Machine Learning as a Service: Challenges and Opportunities
Machine Learning as a Service: Challenges and Opportunities

Abstract: 

Thanks to the recent development of cutting-edge machine learning (ML) techniques, there is a growing number of products that are turning ML models into a long-term sustainable service with the hope that it can evolve and improve with time. In this talk, Dr. Shou-De Lin, Chief Machine Learning Scientst at Appier, will share challenges and opportunities of deploying machine learning models as a long-term service. Typically, accuracy and efficiency have been considered as key factors for evaluating the quality of ML models, but there are many other factors to consider to ensure long-term success. This talk will cover several other important aspects including the data health; modle robustness; the gap between the objective and the ultimate goal; and managing bias throughout the process.

Bio: 

Dr. Shou-de Lin is Appier’s Chief Machine Learning (ML) Scientist. He joined Appier from National Taiwan University (NTU), where he served as a full-time professor in the Department of Computer Science and Information Engineering. Prior to joining NTU, Dr. Lin was a postdoctoral research fellow at the Los Alamos National Lab. Dr. Lin specializes in areas including artificial intelligence, knowledge discovery and natural language processing. He is an expert in solving practical challenges in machine learning applications. Dr. Lin joined Appier in February 2020 and leads the AI research team. His focus is on the development and application of advanced machine learning technology to help customers smoothly implement AI solutions and optimize performance.
Dr. Lin has collaborated with more than 50 companies and has won several awards in recognition of his work, including the 2007 Google Research Awards; the Microsoft Research Awards (which Dr. Lin won three times); and the IBM Research Awards.
During his time at NTU, Dr. Lin established the Machine Discovery and Social Network Mining Lab. He is the all-time winner of the ACM KDD Cup, for which he either led or co-led the NTU team to win 6 championships. He also led a team to win the WSDM Cup in 2016. Dr. Lin is passionate about developing young AI talent in Taiwan, and has supervised more than 100 graduate students, and several of which have gone on to become university professors, founders of start-up companies, or key people at software and hardware companies.