Hyper-parameter Optimisation by Restrained Stochastic Hill Climbing

Authors: Stubbs, R., Wilson, K. and Rostami, S.

Journal: Advances in Intelligent Systems and Computing

Volume: 1043

Pages: 189-200

eISSN: 2194-5365

ISBN: 9783030299323

ISSN: 2194-5357

DOI: 10.1007/978-3-030-29933-0_16

Abstract:

Machine learning practitioners often refer to hyper-parameter optimisation (HPO) as an art form and a skill that requires intuition and experience; Neuroevolution (NE) typically employs a combination of manual and evolutionary approaches for HPO. This paper explores the integration of a stochastic hill climbing approach for HPO within a NE algorithm. We empirically show that HPO by restrained stochastic hill climbing (HORSHC) is more effective than manual and pure evolutionary HPO. Empirical evidence is derived from a comparison of: (1) a NE algorithm that solely optimises hyper-parameters through evolution and (2) a number of derived algorithms with random search optimisation integration for optimising the hyper-parameters of a Neural Network. Through statistical analysis of the experimental results it has been revealed that random initialisation of hyper-parameters does not significantly affect the final performance of the Neural Networks evolved. However, HORSHC, a novel optimisation approach proposed in this paper has been proven to significantly out-perform the NE control algorithm. HORSHC presents itself as a solution that is computationally comparable in terms of both time and complexity as well as outperforming the control algorithm.

https://eprints.bournemouth.ac.uk/32508/

Source: Scopus

Hyper-parameter Optimisation by Restrained Stochastic Hill Climbing

Authors: Stubbs, R., Wilson, K. and Rostami, S.

Journal: ADVANCES IN COMPUTATIONAL INTELLIGENCE SYSTEMS (UKCI 2019)

Volume: 1043

Pages: 189-200

eISSN: 2194-5365

ISBN: 978-3-030-29932-3

ISSN: 2194-5357

DOI: 10.1007/978-3-030-29933-0_16

https://eprints.bournemouth.ac.uk/32508/

Source: Web of Science (Lite)

Hyper-parameter Optimisation by Restrained Stochastic Hill Climbing

Authors: Stubbs, R. and Rostami, S.

Conference: UK Workshop on Computational Intelligence (UKCI)

Dates: 4-6 September 2019

https://eprints.bournemouth.ac.uk/32508/

Source: Manual

Hyper-parameter Optimisation by Restrained Stochastic Hill Climbing

Authors: Stubbs, R., Rostami, S. and Wilson, K.

Conference: UKCI: 19th Annual UK Workshop on Computational Intelligence

Abstract:

Abstract. Machine learning practitioners often refer to hyper-parameter optimisation (HPO) as an art form and a skill that requires intuition and experience; Neuroevolution (NE) typically employs a combination of manual and evolutionary approaches for HPO. This paper explores the integration of a stochastic hill climbing approach for HPO within a NE algorithm. We empirically show that HPO by restrained stochastic hill climbing (HORSHC) is more effective than manual and pure evolutionary HPO. Empirical evidence is derived from a comparison of: (1) a NE algorithm that solely optimises hyper-parameters through evolution and (2) a number of derived algorithms with random search optimisation integration for optimising the hyper-parameters of a Neural Network. Through statistical analysis of the experimental results it has been revealed that random initialisation of hyper-parameters does not significantly affect the final performance of the Neural Networks evolved. However, HORSHC, a novel optimisation approach proposed in this paper has been proven to significantly out-perform the NE control algorithm. HORSHC presents itself as a solution that is computationally comparable in terms of both time and complexity as well as outperforming the control algorithm.

https://eprints.bournemouth.ac.uk/32508/

https://www.ukci2019.port.ac.uk/

Source: BURO EPrints