SILO: Bayesian Preference Exploration: Making Optimization Accessible to Non-Experts

Optimization problems are everywhere — routing trucks, buying groceries, building a datacenter.  Yet optimization methodology is hard to use. It requires the user to write down their objective and constraints as mathematical functions.  In practice, the objective and constraints are unknown and must be tuned iteratively. An expert presents possible solutions to the decision-makers, they provide feedback, and the expert tunes the objective and constraints. The time and expertise required limits adoption of optimization.

We tackle this challenge with Bayesian optimization, preference learning, and LLMs. Our algorithm, Bayesian Optimization with Preference Exploration (BOPE), enables non-experts to effectively optimize complex systems. BOPE learns the user’s preferences through interaction: user feedback on algorithmically-chosen solutions or, enabled by an LLM, a natural language description of the user’s goals. BOPE is particularly valuable when solution candidates are expensive to evaluate. BOPE iteratively updates a Bayesian posterior distribution on the user’s utility function. It uses this posterior to intelligently select which solution pairs to compare or which solution candidates to evaluate. We describe the use of this method for product design at a major social media platform and ongoing work to schedule final exams at Cornell and support logistics decisions made by the US Marines.

Peter Frazier received a B.S. in Physics and Engineering/Applied Science from the California Institute of Technology in 2000, after which he spent several years in industry as a software engineer, working for two different start-up companies and for the Teradata division of NCR. In 2005, he entered graduate school in the Department of Operations Research & Financial Engineering at Princeton University, and received an M.A. in 2007 and a Ph.D. in 2009. He joined the faculty at Cornell in 2009 as an Assistant Professor in the School of Operations Research & Information Engineering, where he is now an Associate Professor. His research is in sequential decision-making under uncertainty, optimal methods for collecting information, and machine learning, focusing on applications in simulation, e-commerce, medicine and biology. He is the recipient of a CAREER Award from the National Science Foundation and a Young Investigator Award from the Air Force Office of Scientific Research. He is currently on leave at Uber, where he is a Staff Data Scientist and Data Science Manager. At Uber, he worked on UberPOOL from 2015-17, and on broader pricing efforts from 2016-17. He now leads a data science team focused on pricing.