BoTorch: A Framework for Efficient Monte-Carlo Bayesian Optimization

TL;DR

BoTorch is a flexible, efficient framework for Bayesian optimization that leverages Monte-Carlo methods, auto-differentiation, and variance reduction to improve sample efficiency and ease of implementation.

Contribution

It introduces a modular, PyTorch-based framework with novel theoretical convergence results and a new 'one-shot' Knowledge Gradient formulation for Bayesian optimization.

Findings

Demonstrates improved sample efficiency over existing libraries

Provides a novel 'one-shot' formulation of the Knowledge Gradient

Offers a flexible, hardware-accelerated implementation for probabilistic models.

Abstract

Bayesian optimization provides sample-efficient global optimization for a broad range of applications, including automatic machine learning, engineering, physics, and experimental design. We introduce BoTorch, a modern programming framework for Bayesian optimization that combines Monte-Carlo (MC) acquisition functions, a novel sample average approximation optimization approach, auto-differentiation, and variance reduction techniques. BoTorch's modular design facilitates flexible specification and optimization of probabilistic models written in PyTorch, simplifying implementation of new acquisition functions. Our approach is backed by novel theoretical convergence results and made practical by a distinctive algorithmic foundation that leverages fast predictive distributions, hardware acceleration, and deterministic optimization. We also propose a novel "one-shot" formulation of the Knowledge Gradient, enabled by a combination of our theoretical and software contributions. In experiments, we demonstrate the improved sample efficiency of BoTorch relative to other popular libraries.