In this talk, I'll present several case studies demonstrating the risks involved in deploying personalized machine learning algorithms, especially recommender systems. Examples range from problems of mainstream interest (such as filter bubbles and 'extremification'), to more subtle forms of unfairness, such as the loss of utility felt by users whose demographics are poorly represented by marketing content. I'll also discuss various strategies to mitigate these undesired effects, by adapting ideas from 'fair ML' to recommendation scenarios.

Speaker Bio: Julian McAuley is a Professor at UC San Diego, where he works on applications of machine learning to problems involving personalization, and teaches classes on personalized recommendation. He likes bicycling and baroque keyboard.

In this talk I'll discuss the fundamentals of personalized shopping in eCommerce using voice-based AI assistants like Amazon's Alexa. I'll cover some of the recent works done in our lab on personalized shopping experience with Alexa, with the focus on product search and product question answering.

Speaker Bio: David is a Principal Applied Scientist at Amazon, Israel and an ACM Distinguished Engineer. His current research is focused on improving product search and product question answering through Amazon’s Alexa. David has published more than 150 papers in IR and Web journals and conferences. He served as a general co-chair for WSDM 2021, a technical program co-chair for CIKM 2019, and an area chair and SPC in many IR and Web conferences. David earned his PhD in Computer Science from the Technion, Israel Institute of Technology in 1997.

Despite the somewhat different techniques used in developing search engines and recommender systems, they generally follow a common goal: helping people to get the information they need at the right time. Therefore, search and recommendation models can potentially benefit from each other. The recent advances in neural network technologies make them effective and easily extendable to various tasks, including retrieval and recommendation. This raises the possibility of jointly modeling and optimizing search ranking and recommendation algorithms, with potential benefits to both. In this talk, I will highlight challenges we face in joint modeling of search and recommendation as well as the opportunities it introduces, such as improving model generalization, interpretability, and explainability. I will further discuss the applications of joint search and recommendation models to conversational AI systems.

Speaker Bio: Hamed Zamani is an Assistant Professor in the Manning College of Information and Computer Sciences (CICS) at the University of Massachusetts Amherst (UMass), where he also serves as the Associate Director of the Center for Intelligent Information Retrieval (CIIR). Prior to UMass, he was a Researcher at Microsoft. In 2019, he received his Ph.D. from UMass under supervision of W. Bruce Croft and received the CICS Outstanding Dissertation Award. His research focuses on developing and evaluating statistical and machine learning models with application to (interactive) information access systems including search engines, recommender systems, and question answering. He is an active member of the information retrieval community and has published over 70 peer-reviewed articles. He is mostly known for his work on neural information retrieval and conversational search. His papers have received awards from ICTIR 2019 and CIKM 2020. Most recently, his research group has been selected for the Alexa Prize Challenge 2021. He has organized multiple workshops at SIGIR, WSDM, and RecSys and is currently serving as the Program Committee Co-Chair for SIGIR 2022 - Short Paper Track.