Generative Adversarial Networks for Graph Data Imputation from Signed Observations

Abstract

We study the problem of missing data imputation for graph signals from signed one-bit quantized observations. More precisely, we consider that the true graph data is drawn from a distribution of signals that are smooth or bandlimited on a known graph. However, instead of observing these signals, we observe a signed version of them and only at a subset of the nodes on the graph. Our goal is to estimate the true underlying graph signals from our observations. To achieve this, we propose a generative adversarial network (GAN) where the key is to incorporate graph-aware losses in the associated minimax optimization problem. We illustrate the benefits of the proposed method via numerical experiments on hand-written digits from the MNIST dataset.

Graph Deep Learning: State of the Art and Challenges

“…Data imputation methods vary widely, from simple statistical replacement, e.g., mean replacement, k nearest neighbors [153], to matrix completion [154], [155] and generative adversarial networks [156], [157]. General imputation methods can be applied to graph models, however, since the increased attention the problem has also been targeted at graph deep learning.…”