English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

For Bayesian optimization (BO) on high-dimensional data with complexstructure, neural network-based kernels for Gaussian processes (GPs) have beenused to learn flexible surrogate functions by the high representation power ofdeep learning. However, existing methods train neural networks by maximizingthe marginal likelihood, which do not directly improve the BO performance. Inthis paper, we propose a meta-learning method for BO with neural network-basedkernels that minimizes the expected gap between the true optimum value and thebest value found by BO. We model a policy, which takes the current evaluateddata points as input and outputs the next data point to be evaluated, by aneural network, where neural network-based kernels, GPs, and mutualinformation-based acquisition functions are used as its layers. With our model,the neural network-based kernel is trained to be appropriate for theacquisition function by backpropagating the gap through the acquisitionfunction and GP. Our model is trained by a reinforcement learning frameworkfrom multiple tasks. Since the neural network is shared across different tasks,we can gather knowledge on BO from multiple training tasks, and use theknowledge for unseen test tasks. In experiments using three text documentdatasets, we demonstrate that the proposed method achieves better BOperformance than the existing methods.

End-to-End Learning of Deep Kernel Acquisition Functions for Bayesian Optimization