Replacing the acquisition function in Bayesian optimization by a neural network

Replacing the acquisition function in Bayesian optimization by a neural network: How effectively do meta-learned acquisition functions in Bayesian optimization perform when optimizing for control variates of unknown functions, as compared to BO with standard acquisition functions

Ramezani, Shayan (TU Delft Electrical Engineering, Mathematics and Computer Science)

Spaan, M.T.J. (mentor)
de Vries, J.A. (mentor)
Lofi, C. (graduation committee)

Delft University of Technology

Computer Science and Engineering

CSE3000 Research Project

2023-06-28

Bayesian Optimization (BO) has demonstrated significant utility across numerous applications. However, due to it being designed as a universal optimizer, its performance can often be suboptimal in specialized environments. To overcome this issue, research has been conducted into the application of transfer learning for enhancing BO performance in these specialized contexts. This paper describes the research done into evaluating the MetaBO algorithm in some specific environments. MetaBO innovates by substituting the acquisition function component in BO with a neural network that serves as an acquisition function, trained via a reinforcement learning framework. Although the results indicate that the algorithm's performance is not optimal in the environments tested, these limitations are ascribed to elements of the implementation rather than the concept of the algorithm itself. Consequently, further research is necessary to refine the implementation process and fully exploit the potential of the MetaBO algorithm.

Bayesian Optimization
meta-learning
Reinforcement Learning
Acquisition Function

http://resolver.tudelft.nl/uuid:93f65530-308a-47ba-8b6f-5e47aa3715f3

Student theses

bachelor thesis

© 2023 Shayan Ramezani