Abstract Number:
1297
Submission Type:
Invited Paper Session
Participants:
Chih-Li Sung (1), Annie Sauer (2), Robert Gramacy (2), Jake Gardner (3), Eytan Bakshy (4), Rui Tuo (5)
Institutions:
(1) Michigan State University, N/A, (2) Virginia Tech, N/A, (3) University of Pennsylvania, N/A, (4) Facebook, N/A, (5) Texas A&M University, N/A
Chair:
Session Organizer:
Speaker(s):
Session Description:
This invited session delves into the cutting-edge advancements in active learning and Bayesian optimization, highlighting their transformative influence on diverse domains such as predictive modeling, experimental design, and optimization. These dynamic statistical techniques are instrumental in guiding data-driven decisions, particularly in scenarios with complex, high-dimensional, and structured data.
The session will concentrate on exploring the methodologies, applications, and theoretical underpinnings of active learning and Bayesian optimization. By bringing together experts in the field (from both academia and industry), it aims to foster an insightful discussion on the evolving landscape of uncertainty quantification and optimization.
The content includes
- "Bayesian Optimization with High-Dimensional Preference Information", presented by Eytan Bakshy (Meta)
- "Contour Location for Reliability in Airfoil Simulation Experiments using Deep Gaussian Processes", presented by Robert B. Gramacy (Virginia Tech)
- "Uncertainty Quantification for Bayesian Optimization", presented by Rui Tuo (TAMU)
- "Bayesian Optimization for High-Dimensional and Structured Problems", presented by Jake Gardner (U Penn)
This session holds broad appeal for statisticians, data scientists, researchers, and practitioners interested in efficiently learning a predictive model and/or strategically optimizing the outcome by sequentially collecting the data. The presentations will cover a spectrum of applications, making it relevant to a wide audience, from those working in engineering and optimization to those focusing on predictive modeling and experimental design. It provides a unique opportunity to gain knowledge from leading experts in these rapidly evolving fields.
In the era of complex data, machine/deep learning, and data-driven decision-making, active learning and Bayesian optimization have emerged as indispensable tools. As such, this session is exceptionally timely, offering insights into the latest techniques that can help inform policy decisions, optimize complex systems, and counteract the spread of misinformation.
Sponsors:
Section on Physical and Engineering Sciences 1
Technometrics 3
Uncertainty Quantification in Complex Systems Interest Group 2
Theme:
Statistics and Data Science: Informing Policy and Countering Misinformation
Yes
Applied
Yes
Estimated Audience Size
Small (<80)
I have read and understand that JSM participants must abide by the Participant Guidelines.
Yes
I understand and have communicated to my proposed speakers that JSM participants must register and pay the appropriate registration fee by June 1, 2024. The registration fee is nonrefundable.
I understand