Towards optimisers that `Keep Learning'

Hart, Emma; Miguel, Ian; Stone, Christopher; Renau, Quentin

Abstract

We consider optimisation in the context of the need to apply an optimiser to a continual stream of instances from one or more domains, and consider how such a system might 'keep learning': by drawing on past experience to improve performance and learning how to both predict and react to instance and/or domain drift.

Citation

Hart, E., Miguel, I., Stone, C., & Renau, Q. (2023). Towards optimisers that `Keep Learning'. In GECCO '23 Companion: Proceedings of the Companion Conference on Genetic and Evolutionary Computation (1636-1638). https://doi.org/10.1145/3583133.3596344

• Conference Name: Companion Conference on Genetic and Evolutionary Computation
• Conference Location: Lisbon, Portugal
• Start Date: Jul 15, 2023
• End Date: Jul 19, 2023
• Online Publication Date: Jul 24, 2023
• Publication Date: 2023-07
• Deposit Date: Aug 1, 2023
• Publisher: Association for Computing Machinery (ACM)
• Pages: 1636-1638
• Book Title: GECCO '23 Companion: Proceedings of the Companion Conference on Genetic and Evolutionary Computation
• DOI: https://doi.org/10.1145/3583133.3596344
• Keywords: Optimisation, continual learning, transfer-learning