Robust Graph Convolutional Networks for Adversarial Resilience and Anomaly Detection
Bibliographic record
Abstract
Adversarial attacks and anomaly detection are closely related, both focusing on identifying irregularities that deviate from normal patterns across various data types, including graph-structured data. Adversarial attacks on graphs pose a significant threat to graph convolutional networks (GCNs), involving intentional manipulation of graph data to mislead GCNs into making incorrect predictions. Standard GCNs, while powerful, often exhibit vulnerabilities to adversarial attacks that can significantly degrade their performance in anomaly detection tasks. These networks also have inherent limitations, such as their inability to effectively consider higher-order neighbour information, restricting their capacity to capture the full context of a node within the graph. To address these challenges, this thesis introduces an iterative graph filtering framework, which builds upon the graph signal processing concept of iteratively solving graph filtering using the fixed-point iterative method. The proposed framework is designed to enhance resilience against adversarial attacks while improving anomaly detection capabilities. The thesis makes two main contributions: a flexible spectral modulation filter that selectively attenuates high-frequency components of graph signals; and a robust aggregation mechanism that efficiently captures information from higher-order node neighbors, expanding the networks receptive field without increasing computational complexity. Extensive experiments are conducted on benchmark datasets to evaluate the effectiveness of the proposed methods. The results demonstrate significant improvements in anomaly detection accuracy and adversarial robustness compared to strong baselines. This highlights the potential of the proposed framework for reliable graph-based downstream tasks, paving the way for robust GCNs that can handle the complexities and adversarial threats inherent in real-world applications.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".