Academic Journal
Are we Forgetting about Compositional Optimisers in Bayesian Optimisation?
| العنوان: | Are we Forgetting about Compositional Optimisers in Bayesian Optimisation? |
|---|---|
| المؤلفون: | Grosnit, Antoine, Cowen-Rivers, Alexander, Tutunov, Rasul, Griffiths, Ryan-Rhys, Wang, Jun, Bou-Ammar, Haitham |
| المصدر: | Journal of Machine Learning Research , 22 pp. 1-78. (2021) |
| بيانات النشر: | MICROTOME PUBL |
| سنة النشر: | 2021 |
| المجموعة: | University College London: UCL Discovery |
| مصطلحات موضوعية: | Science & Technology, Technology, Automation & Control Systems, Computer Science, Artificial Intelligence, Black Box Optimisation, Bayesian Optimisation, Compositional Optimisation, Acquisition Functions, Empirical Analysis, ANT COLONY OPTIMIZATION, EVOLUTION STRATEGIES, GLOBAL OPTIMIZATION, DERIVATIVE-FREE, NEWTON METHOD, ALGORITHM, SEARCH, CONVERGENCE |
| الوصف: | Bayesian optimisation presents a sample-efficient methodology for global optimisation. Within this framework, a crucial performance-determining subroutine is the maximisation of the acquisition function, a task complicated by the fact that acquisition functions tend to be non-convex and thus nontrivial to optimise. In this paper, we undertake a comprehensive empirical study of approaches to maximise the acquisition function. Additionally, by deriving novel, yet mathematically equivalent, compositional forms for popular acquisition functions, we recast the maximisation task as a compositional optimisation problem, allowing us to benefit from the extensive literature in this field. We highlight the empirical advantages of the compositional approach to acquisition function maximisation across 3958 individual experiments comprising synthetic optimisation tasks as well as tasks from Bayesmark. Given the generality of the acquisition function maximisation subroutine, we posit that the adoption of compositional optimisers has the potential to yield performance improvements across all domains in which Bayesian optimisation is currently being applied. An open-source implementation is made available at https://github.com/huawei-noah/noah-research/tree/CompBO/BO/HEBO/CompBO. |
| نوع الوثيقة: | article in journal/newspaper |
| وصف الملف: | application/pdf |
| اللغة: | English |
| Relation: | https://discovery.ucl.ac.uk/id/eprint/10154115/1/20-1422.pdf; https://discovery.ucl.ac.uk/id/eprint/10154115/ |
| الاتاحة: | https://discovery.ucl.ac.uk/id/eprint/10154115/1/20-1422.pdf https://discovery.ucl.ac.uk/id/eprint/10154115/ |
| Rights: | open |
| رقم الانضمام: | edsbas.BAB0D3D0 |
| قاعدة البيانات: | BASE |
| الوصف غير متاح. |