- All Categories
- Featured Events
- Alumni
- Application Deadline
- Arts
- Campus Discourse
- Careers
- BU Central
- Center for the Humanities
- Charity & Volunteering
- Kilachand Center
- Commencement
- Conferences & Workshops
- Diversity & Inclusion
- Examinations
- Food & Beverage
- Global
- Health & Wellbeing
- Keyword Initiative
- Lectures
- LAW Community
- Meetings
- Orientation
- Other Events
- Religious Services & Activities
- Special Interest to Women
- Sports & Recreation
- Social Events
- Study Abroad
- Weeks of Welcome
- Conference for Undergraduate Women in PhysicsAll day
- Grad Welcome All day
- Structuring and Delivering Presentations with Jean-luc Doumont9:00 am
- IS&T RCS Tutorial - Introduction to Linux9:30 am
- Mathemalchemy Exhibition 11:00 am
- IS&T RCS Tutorial - Introduction to BU's Shared Computing Cluster12:00 pm
- LeaRRn Grand Rounds: Jacob Kean, PhD, CCC-SLP presents Electronic Health Record Repositories: Benefits and Challenges of Data Standardization in Rehabilitation1:30 pm
- Workshop: Fellowship Applications (Zoom)2:00 pm
- Graduate Students Support Group3:00 pm
- CISE Seminar: Wanning Chen, Assistant Professor, University of Washington3:00 pm
- Mathemalchemy Opening Reception 4:30 pm
- Legos and Eggos7:00 pm
- Welcome Back Shabbat Dinner7:00 pm
- GradHIllel Welcome Back Shabbat Dinner 7:00 pm
CISE Seminar: Wanning Chen, Assistant Professor, University of Washington
Speed Up the Cold-Start Learning in Two-Sided Bandits with Many Arms
Multi-armed bandit (MAB) algorithms are efficient approaches to reduce the opportunity cost of online experimentation and are used by companies to find the best product from periodically refreshed product catalogs. However, these algorithms face the so-called cold-start at the onset of the experiment due to a lack of knowledge of customer preferences for new products, requiring an initial data collection phase known as the burn-in period. During this period, MAB algorithms operate like randomized experiments, incurring large burn-in costs which scale with the large number of products. We attempt to reduce the burn-in by identifying that many products can be cast into two-sided products, and then naturally model the rewards of the products with a matrix, whose rows and columns represent the two sides respectively. Next, we design two-phase bandit algorithms that first use subsampling and low-rank matrix estimation to obtain a substantially smaller targeted set of products and then apply a UCB procedure on the target products to find the best one. We theoretically show that the proposed algorithms lower costs and expedite the experiment in cases when there is limited experimentation time along with a large product set. Our analysis also reveals three regimes of long, short, and ultra-short horizon experiments, depending on dimensions of the matrix. Empirical evidence from both synthetic data and a real-world dataset on music streaming services validates this superior performance.
Wanning Chen is an assistant professor in the Information Systems group at Foster School of Business. She completed her Ph.D. in business from Stanford University in 2022. Her research focuses on developing statistical methodology for data-driven decision making. In particular, She designs novel machine learning algorithms for problems where the underlying data has a natural matrix structure.
Faculty Host: Jinglong ZhaoStudent Host: Zili WangWhen | 3:00 pm to 4:00 pm on Friday, January 20, 2023 |
---|---|
Location | 8 St. Mary's Street, PHO 203 |