Title: A New Network Learning Model for Distributions
Despite the superior performance of deep neural networks, they are not efficient for regressing on function spaces. In particular, neural networks require many parameters to represent function inputs as each node encodes just a real value. We propose a novel idea: to encode an entire function in a single network node. We design a network that propagates functions in single nodes for the distribution regression task. Our proposed distribution regression network achieves higher accuracies while using fewer parameters and training data compared to traditional neural networks.
In this seminar, I will also spend 5 minutes to give an overview of other projects in our laboratory, in particular, projects related to cancer diagnosis.
|Date||March 25, 2019 (Mon) 11:00 - 12:00|