Skip to main content

Hyper-parameter Optimisation by Restrained Stochastic Hill Climbing.

Stubbs, R., Rostami, S. and Wilson, K., 2019. Hyper-parameter Optimisation by Restrained Stochastic Hill Climbing. In: UKCI: 19th Annual UK Workshop on Computational Intelligence, 4-6 September 2019, Portsmouth, UK.

Full text available as:

[img]
Preview
PDF
UKCI2019_Paper__Rhys_.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial.

1MB

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

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.

Item Type:Conference or Workshop Item (Paper)
Uncontrolled Keywords:hyper-parameter optimisation; global optimisation; neuroevolution; artificial neural networks; random search; stochastic hill climbing
Group:Faculty of Science & Technology
ID Code:32508
Deposited By: Unnamed user with email symplectic@symplectic
Deposited On:10 Jul 2019 14:01
Last Modified:09 Sep 2019 09:34

Downloads

Downloads per month over past year

More statistics for this item...
Repository Staff Only -