Adaptive Deep Kernel Learning

TL;DR

This paper introduces a method for learning a family of kernels using deep neural networks, enabling better adaptation to diverse few-shot regression tasks, especially in drug discovery applications.

Contribution

It proposes a novel deep kernel learning approach that learns a kernel family for multiple tasks, improving task-specific kernel selection during inference.

Findings

Outperforms existing algorithms on real-world few-shot regression tasks.

Effective in identifying suitable kernels for each task.

Demonstrates advantages in drug discovery applications.

Abstract

Deep kernel learning provides an elegant and principled framework for combining the structural properties of deep learning algorithms with the flexibility of kernel methods. By means of a deep neural network, we learn a parametrized kernel operator that can be combined with a differentiable kernel algorithm during inference. While previous work within this framework has focused on learning a single kernel for large datasets, we learn a kernel family for a variety of few-shot regression tasks. Compared to single deep kernel learning, our algorithm enables the identification of the appropriate kernel for each task during inference. As such, it is well adapted for complex task distributions in a few-shot learning setting, which we demonstrate by comparing against existing state-of-the-art algorithms using real-world, few-shot regression tasks related to the field of drug discovery.