Deep Kernels

Описание к видео Deep Kernels

Presenters: Sebastian Ober and Austin Tripp (University of Cambridge)
Abstract: Deep kernel learning methods try to combine the expressive power of neural networks with the uncertainty representation of Gaussian processes. This is achieved by learning a feature extractor to transform the input data before using a Gaussian process model. In this talk, we will describe what deep kernel learning is in depth, before discussing recent advances and insights.

Комментарии

Информация по комментариям в разработке

Похожие видео