Chaochao Chen, Assistant Professor
Zhejiang University

Bio:

Chaochao Chen is currently an Assistant Professor at the College of Computer Science and Technology, Zhejiang University. Before that, He was a Staff Algorithm Engineer at Ant Group. He was also a visiting scholar at the Computer Science Department of UIUC. His research mainly focuses on privacy preserving machine learning, privacy-preserving data mining, and machine learning. Prof. Chen has published more than 60 papers in peer reviewed journals and conferences. He has also applied for more than 200 International and Chinese patents, and more than 100 of them have been issued.

 

Title:

Privacy Preserving Recommender Systems

Abstract:

Recommender systems aim to provide users with personalized information filtering services, which inevitably uses a large amount of personal information during the process. With the promulgation of data security related regulations in recent years, how to protect user privacy and data security when building recommender systems has become a hot research problem. In this talk, I will focus on three privacy preserving recommendation scenarios, i.e., cross-device privacy preserving recommendation, cross-silo privacy preserving recommendation, and privacy preserving recommendation unlearning. I will introduce the corresponding solutions and representative technologies used in these scenarios.

 

 

Yanchi Liu, Research Staff Member
NEC Labs ,America

Bio

Yanchi Liu is a Research Staff Member in the Department of Data Science and System Security at the NEC Labs America. He received his Ph.D. in Information Technology from Rutgers, the State University of New Jersey. He has broad interests in data science, machine learning, natural language processing, and their applications to solve real problems. His current research develops efficient and effective solutions to address challenges in various domains, including e-commerce, social networks, complex system management, finance, cybersecurity, and aerospace. He is the author of over 70 technical papers and a recipient of the Baidu Research Fellowship.

 

Title

Domain-oriented Textual Representation Learning

Abstract:

Representation learning serves as an important component for recommendation. However, most works on Recommender Systems pay little attention to textual information and do not fully take advantage of recent natural language processing (NLP) techniques. Moreover, each recommendation task usually requires accurate understanding of domain knowledge, and such fine-grained knowledge is hard to be captured by existing scheme. In this talk, we will introduce how to utilize NLP for effective representation learning in specific domains like e-commerce, from formalizing concrete problems to developing algorithmic methods, motivated by real applications.