Smart Testing with Vaccination: A Bandit Algorithm for Active Sampling for Managing COVID-19
Abstract:
This paper presents methods to choose individuals to test for infection during a pandemic such as COVID-19, characterized by high contagion and presence of asymptomatic carriers. The smart-testing ideas presented here are motivated by active learning and multi-armed bandit techniques in machine learning. Our active sampling method works in conjunction with vaccination and quarantine policies and is adaptive to changes in real-time data. Using a data-driven agent-based model simulating New York City we show that the algorithm samples individuals to test in a manner that rapidly traces infected individuals. The results show that smart-testing is effective in significantly reducing infection and death rates as compared to current policies, with or without vaccination.
Bio:
Yingfei Wang is an Assistant Professor at Department of Information Systems and Operations Management, University of Washington. Her research lies at the intersection of data analytics, statistics, machine learning and management information systems, exploring the ways where efficient information collection influences and improves decision-making strategies, with potential applications as diverse as healthcare management, e-commerce, recommendation systems, auctions and revenue management. She is devoted to solving machine learning challenges to provide efficient business solutions, using technologies from deep learning, multi-armed bandits, Bayesian optimization, natural language processing and beyond. Her work has been published in Management Science, Journal on Computing, Siam Journals, IEEE Transactions on Knowledge and Data Engineering, and leading CS conferences. Yingfei holds a Ph.D. degree from Department of Computer Science at Princeton University, and a BS degree in Computer Science from Peking University.