Presented By: U-M Industrial & Operations Engineering
PHD SEMINAR: "Personalized Data-Driven Learning and Optimization: Theory and Applications to Healthcare" — Esmaeil Keyvanshokooh
This event is designed for U-M IOE PhD students and faculty and is also open to all U-M students, faculty and staff.
Title:
Personalized Data-Driven Learning and Optimization: Theory and Applications to Healthcare
Abstract:
The rapid growth of information and accessibility to big data provide a unique opportunity to shift toward personalized data-driven decision-making analytics. Healthcare presents many decision support opportunities for personalized/precision treatment choices based on patient biomarkers and clinical history. In marketing, these same methods can increase the click-through rate through ads and promotions tailored to the user’s demographics and interests. These real-time personalized decision-making paradigms (i) adaptively learn a model that predicts a user-specific outcome for each available decision as a function of the user's known contextual information (prediction), and (ii) harness this model to make optimize personalized decisions for subsequent users (prescription). In this talk, I introduce critical challenges in the development of today’s real-time personalized decision making paradigms: the need for making "nested" personalized decisions jointly and accounting for limited resource capacities. I then present new personalized data driven predictive and prescriptive analytical methods with provable performance guarantee to deal with these challenges. In addition to provable performance guarantees, the effectiveness of these new methods is illustrated through case studies using real-word medical/healthcare data.
Bio:
Esmaeil Keyvanshokooh is broadly interested in developing personalized data driven analytical methods for a wide range of business analytics applications. To address unmet real-world needs in healthcare and operations engineering, he generates novel state-of-the-art analytics techniques to yield insights and new functionality. For the 2020 INFORMS Decision Analysis Society Best Paper Award, he was a finalist (2nd place). He won both the 2020 Katta G. Murty Best Paper Award on Optimization, and 2019 Richard Wilson Best Paper Award on Service Operations. He has received several other awards, including the 2017 IOE Bonder Fellowship in Applied Operations Research and the prestigious 2020 University of Michigan Rackham Predoctoral Fellowship. Esmaeil is a Ph.D. candidate in Operations Research at the Department of Industrial and Operations Engineering (IOE) at University of Michigan, working under the supervision of Prof. Mark Van Oyen and Prof. Cong Shi. He received his M.Sc. degrees in Statistics from University of Michigan, and in Industrial Engineering and Operations Research from Iowa State University.
Title:
Personalized Data-Driven Learning and Optimization: Theory and Applications to Healthcare
Abstract:
The rapid growth of information and accessibility to big data provide a unique opportunity to shift toward personalized data-driven decision-making analytics. Healthcare presents many decision support opportunities for personalized/precision treatment choices based on patient biomarkers and clinical history. In marketing, these same methods can increase the click-through rate through ads and promotions tailored to the user’s demographics and interests. These real-time personalized decision-making paradigms (i) adaptively learn a model that predicts a user-specific outcome for each available decision as a function of the user's known contextual information (prediction), and (ii) harness this model to make optimize personalized decisions for subsequent users (prescription). In this talk, I introduce critical challenges in the development of today’s real-time personalized decision making paradigms: the need for making "nested" personalized decisions jointly and accounting for limited resource capacities. I then present new personalized data driven predictive and prescriptive analytical methods with provable performance guarantee to deal with these challenges. In addition to provable performance guarantees, the effectiveness of these new methods is illustrated through case studies using real-word medical/healthcare data.
Bio:
Esmaeil Keyvanshokooh is broadly interested in developing personalized data driven analytical methods for a wide range of business analytics applications. To address unmet real-world needs in healthcare and operations engineering, he generates novel state-of-the-art analytics techniques to yield insights and new functionality. For the 2020 INFORMS Decision Analysis Society Best Paper Award, he was a finalist (2nd place). He won both the 2020 Katta G. Murty Best Paper Award on Optimization, and 2019 Richard Wilson Best Paper Award on Service Operations. He has received several other awards, including the 2017 IOE Bonder Fellowship in Applied Operations Research and the prestigious 2020 University of Michigan Rackham Predoctoral Fellowship. Esmaeil is a Ph.D. candidate in Operations Research at the Department of Industrial and Operations Engineering (IOE) at University of Michigan, working under the supervision of Prof. Mark Van Oyen and Prof. Cong Shi. He received his M.Sc. degrees in Statistics from University of Michigan, and in Industrial Engineering and Operations Research from Iowa State University.
Related Links
Explore Similar Events
-
Loading Similar Events...