Learning to Warm-Start Bayesian Hyperparameter Optimization

Abstract

Hyperparameter optimization aims to find the optimal hyperparameter configuration of a machine learning model, which provides the best performance on a validation dataset. Manual search usually leads to get stuck in a local hyperparameter configuration, and heavily depends on human intuition and experience. A simple alternative of manual search is random/grid search on a space of hyperparameters, which still undergoes extensive evaluations of validation errors in order to find its best configuration. Bayesian optimization that is a global optimization method for black-box functions is now popular for hyperparameter optimization, since it greatly reduces the number of validation error evaluations required, compared to random/grid search. Bayesian optimization generally finds the best hyperparameter configuration from random initialization without any prior knowledge. This motivates us to let Bayesian optimization start from the configurations that were successful on similar datasets, which are able to remarkably minimize the number of evaluations. In this paper, we propose deep metric learning to learn meta-features over datasets such that the similarity over them is effectively measured by Euclidean distance between their associated meta-features. To this end, we introduce a Siamese network composed of deep feature and meta-feature extractors, where deep feature extractor provides a semantic representation of each instance in a dataset and meta-feature extractor aggregates a set of deep features to encode a single representation over a dataset. Then, our learned meta-features are used to select a few datasets similar to the new dataset, so that hyperparameters in similar datasets are adopted as initializations to warm-start Bayesian hyperparameter optimization.

14 pages, a preliminary version was presented at NIPS 2017 workshop on Bayesian optimization

References in corpus (7)

• Practical Bayesian Optimization of Machine Learning Algorithms
• A Tutorial on Bayesian Optimization of Expensive Cost Functions, with Application to Active User Modeling and Hierarchical Reinforcement Learning
• Theoretical Models of Learning to Learn
• Matching Networks for One Shot Learning
• Learning to Learn without Gradient Descent by Gradient Descent
• Zero-Shot Learning -- A Comprehensive Evaluation of the Good, the Bad and the Ugly
• Towards a Neural Statistician

Cited by in corpus (5)

• Hyperparameter Learning via Distributional Transfer
• MxML: Mixture of Meta-Learners for Few-Shot Classification
• HyperSTAR: Task-Aware Hyperparameters for Deep Networks
• Siamese Meta-Learning and Algorithm Selection with 'Algorithm-Performance Personas' [Proposal]
• On Hyper-parameter Tuning for Stochastic Optimization Algorithms