Distribution Regression Network

Connie Kou, Hwee Kuan Lee, Teck Khim Ng

Abstract: We introduce our Distribution Regression Network (DRN) which performs regression from input probability distributions to output probability distributions. Compared to existing methods, DRN learns with fewer model parameters and easily extends to multiple input and multiple output distributions. On synthetic and real-world datasets, DRN performs similarly or better than the state-of-the-art. Furthermore, DRN generalizes the conventional multilayer perceptron (MLP). In the framework of MLP, each node encodes a real number, whereas in DRN, each node encodes a probability distribution.

TL;DR: A learning network which generalizes the MLP framework to perform distribution-to-distribution regression

Keywords: distribution regression, supervised learning, regression analysis