Distinguished Professor Jian Pei Sheds Light on Data Valuation in Federated Learning

Professor Jian Pei, an esteemed expert in the fields of data science, big data, data mining, and database systems from Duke University, delivered an insightful lecture titled “Data Valuation in Federated Learning” on 28 June 2024.

In his sharing, Professor Pei explored the complexities surrounding data valuation in federated learning and shared his research methodologies in addressing critical incentive and fairness issues. Federated Learning, a decentralised machine learning approach, enables multiple parties to collaboratively build models. Yet, there are two pivotal challenges in current federated learning development: privacy and fairness concerns. To address these issues, Professor Pei highlighted the importance of Data Valuation in Federated Learning. While the existing Federated Shapley Value framework measures data value effectively, there are still elements of potential unfairness. In response, Professor Pei introduced his research work on Completed Federated Shapley Value to refine the impartial and evaluate data effectively. Furthermore, his research work on personalised federated learning models shedding light on methodologies to enhance fairness in learning outcomes. Professor Pei's sharing not only offered valuable insights into the practical implementation of federated learning but also paved the way for more equitable and efficient solutions for future machine learning models.

Professor Jian Pei, the Arthur S. Pearse Distinguished Professor at Duke University, is a renowned expert in data science, big data, data mining, and database systems. He is also a Fellow of the Royal Society of Canada, the Canadian Academy of Engineering, ACM, and IEEE. Professor Pei also received many prestigious awards, including the 2017 ACM SIGKDD Innovation Award and the 2015 ACM SIGKDD Service Award. His algorithms are widely adopted in industry and popular open source software suites.

More Photos