Regret Bounds and Experimental Design for Estimate-then-Optimize

10/27/2022
by   Samuel Tan, et al.
0

In practical applications, data is used to make decisions in two steps: estimation and optimization. First, a machine learning model estimates parameters for a structural model relating decisions to outcomes. Second, a decision is chosen to optimize the structural model's predicted outcome as if its parameters were correctly estimated. Due to its flexibility and simple implementation, this “estimate-then-optimize” approach is often used for data-driven decision-making. Errors in the estimation step can lead estimate-then-optimize to sub-optimal decisions that result in regret, i.e., a difference in value between the decision made and the best decision available with knowledge of the structural model's parameters. We provide a novel bound on this regret for smooth and unconstrained optimization problems. Using this bound, in settings where estimated parameters are linear transformations of sub-Gaussian random vectors, we provide a general procedure for experimental design to minimize the regret resulting from estimate-then-optimize. We demonstrate our approach on simple examples and a pandemic control application.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
04/13/2023

Estimate-Then-Optimize Versus Integrated-Estimation-Optimization: A Stochastic Dominance Perspective

In data-driven stochastic optimization, model parameters of the underlyi...
research
11/25/2022

A Note on Model-Free Reinforcement Learning with the Decision-Estimation Coefficient

We consider the problem of interactive decision making, encompassing str...
research
12/27/2021

The Statistical Complexity of Interactive Decision Making

A fundamental challenge in interactive learning and decision making, ran...
research
02/29/2020

Decision Trees for Decision-Making under the Predict-then-Optimize Framework

We consider the use of decision trees for decision-making problems under...
research
10/16/2019

Scaling up Psychology via Scientific Regret Minimization: A Case Study in Moral Decision-Making

Do large datasets provide value to psychologists? Without a systematic m...
research
03/12/2023

Branch Learn with Post-hoc Correction for Predict+Optimize with Unknown Parameters in Constraints

Combining machine learning and constrained optimization, Predict+Optimiz...
research
12/03/2021

Neural Pseudo-Label Optimism for the Bank Loan Problem

We study a class of classification problems best exemplified by the bank...

Please sign up or login with your details

Forgot password? Click here to reset