Poisoning Knowledge Graph Embeddings via Relation Inference Patterns

Abstract

We study the problem of generating data poisoning attacks against Knowledge Graph Embedding (KGE) models for the task of link prediction in knowledge graphs. To poison KGE models, we propose to exploit their inductive abilities which are captured through the relationship patterns like symmetry, inversion and composition in the knowledge graph. Specifically, to degrade the model's prediction confidence on target facts, we propose to improve the model's prediction confidence on a set of decoy facts. Thus, we craft adversarial additions that can improve the model's prediction confidence on decoy facts through different inference patterns. Our experiments demonstrate that the proposed poisoning attacks outperform state-of-art baselines on four KGE models for two publicly available datasets. We also find that the symmetry pattern based attacks generalize across all model-dataset combinations which indicates the sensitivity of KGE models to this pattern.

Adversarial Attacks on Knowledge Graph Embeddings via Instance Attribution Methods

“…Additionally, as in Bhardwaj (2020); Bhardwaj et al (2021), we call for future proposals to defend against the security vulnerabilities of KGE models. Some promising directions might be to use adversarial training techniques or train ensembles of models over subsets of training data to prevent the model predictions being influenced by a few triples only.…”

“…We also study data poisoning attacks against KGE models in Bhardwaj et al (2021). Here, we exploit the inductive abilities of KGE models to select adversarial additions that improve the predictive performance of the model on a set of decoy triples; which in turn degrades the performance on target triples.…”

One Versus All Binary Tree Method to Classify Misbehaviors in Imbalanced VeReMi Dataset

“…Similar problems are seen in the VeReMi Extension database, which has been widely utilized to learn ML classifiers to classify misbehaving nodes in VANETs. There are several approaches to solving this issue: changing the machine learning approach, adding an inaccurate categorization fee, and data sampling [20]. In the context of VANETs, the researchers [21] attempt to oversample the minority category by employing the Synthetic Minority Oversampling Technique (SMOTE), an approach to data augmentation for the less category [22] [23] that duplicates instances from the less category, in order to avoid the issue of imbalanced datasets.…”