Bayesian functional optimisation with shape prior
| Autor: | Pratibha Vellanki, Murray Height, Sunil Gupta, Svetha Venkatesh, Santu Rana, Alessandra Sutti, David Rubin de Celis Leal |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2018 |
| Předmět: |
FOS: Computer and information sciences
Mathematical optimization Computer Science - Machine Learning Basis (linear algebra) Computer science Statistics - Machine Learning Bayesian probability Control variable Machine Learning (stat.ML) Degree of a polynomial General Medicine Bernstein polynomial Machine Learning (cs.LG) |
| Zdroj: | AAAI Scopus-Elsevier |
| Popis: | Real world experiments are expensive, and thus it is important to reach a target in minimum number of experiments. Experimental processes often involve control variables that changes over time. Such problems can be formulated as a functional optimisation problem. We develop a novel Bayesian optimisation framework for such functional optimisation of expensive black-box processes. We represent the control function using Bernstein polynomial basis and optimise in the coefficient space. We derive the theory and practice required to dynamically adjust the order of the polynomial degree, and show how prior information about shape can be integrated. We demonstrate the effectiveness of our approach for short polymer fibre design and optimising learning rate schedules for deep networks. Submitted to AAAI 2019 |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|