Robust Graph Representation Learning for Local Corruption Recovery

TL;DR

This paper introduces a novel graph learning scheme that detects and recovers from local attribute corruptions, enhancing the robustness of graph embeddings for prediction tasks without assuming corruption distributions.

Contribution

It proposes a graph autoencoder-based method that identifies corrupted features and recovers robust embeddings using sparsity regularization and framelet domain sparsity, improving resilience to anomalies.

Findings

Effective recovery of robust graph representations from poisoned data

Outperforms existing methods in anomaly detection and prediction accuracy

Demonstrates robustness against black-box poisoning attacks

Abstract

The performance of graph representation learning is affected by the quality of graph input. While existing research usually pursues a globally smoothed graph embedding, we believe the rarely observed anomalies are as well harmful to an accurate prediction. This work establishes a graph learning scheme that automatically detects (locally) corrupted feature attributes and recovers robust embedding for prediction tasks. The detection operation leverages a graph autoencoder, which does not make any assumptions about the distribution of the local corruptions. It pinpoints the positions of the anomalous node attributes in an unbiased mask matrix, where robust estimations are recovered with sparsity promoting regularizer. The optimizer approaches a new embedding that is sparse in the framelet domain and conditionally close to input observations. Extensive experiments are provided to validate our proposed model can recover a robust graph representation from black-box poisoning and achieve excellent performance.