Bayesian optimisation for automated machine learning
<p>In this thesis, we develop a rich family of efficient and performant Bayesian optimisation (BO) methods to tackle various AutoML tasks. We first introduce a fast information-theoretic BO method, FITBO, that overcomes the computation bottleneck of information-theoretic acquisition functions...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| Subjects: |