Automated Machine Learning on Big Data using Stochastic Algorithm Tuning

TL;DR

This paper presents a scalable stochastic Bayesian optimisation method for automating machine learning hyper-parameter tuning on big data, enabling efficient and effective model training without extensive manual effort.

Contribution

It introduces a novel stochastic, sparse Bayesian optimisation strategy using Nystrom approximation to scale to large evaluations on big data, outperforming existing methods.

Findings

Nystrom approximation provides the best scaling and performance.

The proposed method significantly improves hyper-parameter tuning efficiency.

Demonstrated superior results on real big data time series prediction tasks.

Abstract

We introduce a means of automating machine learning (ML) for big data tasks, by performing scalable stochastic Bayesian optimisation of ML algorithm parameters and hyper-parameters. More often than not, the critical tuning of ML algorithm parameters has relied on domain expertise from experts, along with laborious hand-tuning, brute search or lengthy sampling runs. Against this background, Bayesian optimisation is finding increasing use in automating parameter tuning, making ML algorithms accessible even to non-experts. However, the state of the art in Bayesian optimisation is incapable of scaling to the large number of evaluations of algorithm performance required to fit realistic models to complex, big data. We here describe a stochastic, sparse, Bayesian optimisation strategy to solve this problem, using many thousands of noisy evaluations of algorithm performance on subsets of data in order to effectively train algorithms for big data. We provide a comprehensive benchmarking of possible sparsification strategies for Bayesian optimisation, concluding that a Nystrom approximation offers the best scaling and performance for real tasks. Our proposed algorithm demonstrates substantial improvement over the state of the art in tuning the parameters of a Gaussian Process time series prediction task on real, big data.