Active Learning for Enumerating Local Minima Based on Gaussian Process Derivatives

DOI10.1162/NECO_A_01307MaRDI QIDQ3386417zbMATH OpenOpenAlexWikidataFDO

Authors Yu Inatsu, Daisuke Sugita, Kazuaki Toyoura, Ichiro Takeuchi

Publication date 4 January 2021

Published in Neural Computation (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1903.03279

Mathematics Subject Classification ID

Gaussian processes (60G15) Learning and adaptive systems in artificial intelligence (68T05) Stochastic programming (90C15)

Abstract: We study active learning (AL) based on Gaussian Processes (GPs) for efficiently enumerating all of the local minimum solutions of a black-box function. This problem is challenging due to the fact that local solutions are characterized by their zero gradient and positive-definite Hessian properties, but those derivatives cannot be directly observed. We propose a new AL method in which the input points are sequentially selected such that the confidence intervals of the GP derivatives are effectively updated for enumerating local minimum solutions. We theoretically analyze the proposed method and demonstrate its usefulness through numerical experiments.

Recommendations

• Minimax Bounds for Active Learning
• Minimax Bounds for Active Learning
• Discrimination between Gaussian process models: active learning and static constructions
• Minimax analysis of active learning
• Active learning Bayesian support vector regression model for global approximation
• Minimax robust active learning for approximately specified regression models
• Active learning with multiple localized regression models

Cites work

• Bayesian optimization in a billion dimensions via random embeddings
• Direct search methods: Then and now
• Gaussian processes for machine learning.
• scientific article; zbMATH DE number 6276166 (Why is no real title available?)
• Local maxima of Gaussian fields
• Posterior consistency of Gaussian process prior for nonparametric binary regression
• The correlated knowledge gradient for simulation optimization of continuous parameters using Gaussian process regression

Cited in

(4)

• Active learning for saddle point calculation
• An active learning approach for improving the performance of equilibrium based chemical simulations
• Active learning of Bayesian linear models with high-dimensional binary features by parameter confidence-region estimation
• Active learning for level set estimation under input uncertainty and its extensions

This page was built for publication: Active Learning for Enumerating Local Minima Based on Gaussian Process Derivatives

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3386417)