Robust Graph Embedding with Noisy Link Weights

TL;DR

This paper introduces a robust graph embedding method called $eta$-graph embedding that effectively handles noisy link weights by using a novel empirical moment $eta$-score and an efficient stochastic algorithm, demonstrated through experiments.

Contribution

The paper presents a new $eta$-graph embedding technique that improves robustness to noise in link weights and provides a scalable stochastic optimization algorithm.

Findings

Effective in reducing influence of contaminated data

Computationally efficient with stochastic algorithm

Validated on synthetic and real-world datasets

Abstract

We propose $\beta$-graph embedding for robustly learning feature vectors from data vectors and noisy link weights. A newly introduced empirical moment $\beta$-score reduces the influence of contamination and robustly measures the difference between the underlying correct expected weights of links and the specified generative model. The proposed method is computationally tractable; we employ a minibatch-based efficient stochastic algorithm and prove that this algorithm locally minimizes the empirical moment $\beta$-score. We conduct numerical experiments on synthetic and real-world datasets.