Multivariate-Information Adversarial Ensemble for Scalable Joint Distribution Matching

07/08/2019
by   Ziliang Chen, et al.
6

A broad range of cross-m-domain generation researches boil down to matching a joint distribution by deep generative models (DGMs). Hitherto algorithms excel in pairwise domains while as m increases, remain struggling to scale themselves to fit a joint distribution. In this paper, we propose a domain-scalable DGM, i.e., MMI-ALI for m-domain joint distribution matching. As an m-domain ensemble model of ALIs dumoulin2016adversarially, MMI-ALI is adversarially trained with maximizing Multivariate Mutual Information (MMI) w.r.t. joint variables of each pair of domains and their shared feature. The negative MMIs are upper bounded by a series of feasible losses that provably lead to matching m-domain joint distributions. MMI-ALI linearly scales as m increases and thus, strikes a right balance between efficacy and scalability. We evaluate MMI-ALI in diverse challenging m-domain scenarios and verify its superiority.

READ FULL TEXT

page 6

page 7

page 8

research
06/08/2018

JointGAN: Multi-Domain Joint Distribution Learning with Generative Adversarial Nets

A new generative adversarial network is developed for joint distribution...
research
03/01/2020

Joint Wasserstein Distribution Matching

Joint distribution matching (JDM) problem, which aims to learn bidirecti...
research
07/07/2023

Distilling Universal and Joint Knowledge for Cross-Domain Model Compression on Time Series Data

For many real-world time series tasks, the computational complexity of p...
research
07/17/2020

Learning to Match Distributions for Domain Adaptation

When the training and test data are from different distributions, domain...
research
05/22/2019

Joint Information Preservation for Heterogeneous Domain Adaptation

Domain adaptation aims to assist the modeling tasks of the target domain...
research
06/29/2022

MaNi: Maximizing Mutual Information for Nuclei Cross-Domain Unsupervised Segmentation

In this work, we propose a mutual information (MI) based unsupervised do...
research
09/24/2019

dAUTOMAP: decomposing AUTOMAP to achieve scalability and enhance performance

AUTOMAP is a promising generalized reconstruction approach, however, it ...

Please sign up or login with your details

Forgot password? Click here to reset