US Patent Application 18347406. Optimization of Parameter Values for Machine-Learned Models simplified abstract
Contents
Optimization of Parameter Values for Machine-Learned Models
Organization Name
Inventor(s)
Daniel Reuben Golovin of Pittsburgh PA (US)
Benjamin Solnik of Pittsburgh PA (US)
Subhodeep Moitra of Pittsburgh PA (US)
David W. Sculley, Ii of Cambridge MA (US)
Gregory Peter Kochanski of Pittsburgh PA (US)
Optimization of Parameter Values for Machine-Learned Models - A simplified explanation of the abstract
- This abstract for appeared for US patent application number 18347406 Titled 'Optimization of Parameter Values for Machine-Learned Models'
Simplified Explanation
The abstract describes a computing system and methods for optimizing adjustable parameters of a system. It introduces a parameter optimization system that uses black-box optimization techniques to suggest new parameter values for evaluation. This iterative process aims to improve the overall performance of the system, as measured by an objective function. The abstract also mentions a new optimization technique called "Gradientless Descent," which is faster than random search while maintaining its positive qualities.
Original Abstract Submitted
The present disclosure provides computing systems and associated methods for optimizing one or more adjustable parameters (e.g. operating parameters) of a system. In particular, the present disclosure provides a parameter optimization system that can perform one or more black-box optimization techniques to iteratively suggest new sets of parameter values for evaluation. The iterative suggestion and evaluation process can serve to optimize or otherwise improve the overall performance of the system, as evaluated by an objective function that evaluates one or more metrics. The present disclosure also provides a novel black-box optimization technique known as “Gradientless Descent” that is more clever and faster than random search yet retains most of random search's favorable qualities.
