Robust Graph Embedding with Noisy Link Weights
This work addresses robust graph embedding for data with noisy links, presenting an incremental improvement in handling contamination.
The paper tackles the problem of learning robust graph embeddings from noisy link weights by introducing a β-score to reduce contamination influence, achieving improved robustness in synthetic and real-world experiments.
We propose $β$-graph embedding for robustly learning feature vectors from data vectors and noisy link weights. A newly introduced empirical moment $β$-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 $β$-score. We conduct numerical experiments on synthetic and real-world datasets.