A Perspective on Machine Learning for Autonomous Experimentation
Document Type
Book
Role
Contributor
Publication
Methods and Applications of Autonomous Experimentation
Publisher
CRC Press
Standard Number
9781003359593
First Page
36
Last Page
46
Publication Date
2023
Abstract
This chapter introduces machine learning (ML) in the context of autonomous scientific experimentation. It outlines the types of roles that ML can play in supporting autonomous experimentation, and in doing so provide a language for discussing the applications. The chapter discusses specific applications of ML for autonomous experimentation and their successes. The most general definition of ML is as a method for creating functions (i.e., input-output relationships) not by explicit programming but instead by showing example data. “Function” here is used both in its mathematical sense, as in a relation that uniquely associates members of one set with members of another set, and in its computer programming sense, a sequence of computer instructions that perform a task. As famously noted by Breimann, the practice and epistemic philosophy of ML differs from the model-based parameter estimation traditionally practiced in statistics and the sciences.
Repository Citation
Schrier, J., & Norquist, A. J. (2023). A Perspective on Machine Learning for Autonomous Experimentation. In M. Noack & D. Ushizima (Eds.), Methods and Applications of Autonomous Experimentation (pp. 36–46). CRC Press. https://doi.org/10.1201/9781003359593-3
