Performance Evaluation of Knowledge Graph Embedding Approaches under Non-adversarial Attacks

TL;DR

This paper evaluates how non-adversarial attacks, such as label and parameter perturbations, impact the robustness of state-of-the-art Knowledge Graph Embedding algorithms across multiple datasets.

Contribution

It fills a research gap by systematically analyzing the effects of non-adversarial attacks on KGE approaches, which was previously underexplored.

Findings

Label perturbation significantly degrades KGE performance

Parameter perturbation has a moderate impact

Graph perturbation has a low impact

Abstract

Knowledge Graph Embedding (KGE) transforms a discrete Knowledge Graph (KG) into a continuous vector space facilitating its use in various AI-driven applications like Semantic Search, Question Answering, or Recommenders. While KGE approaches are effective in these applications, most existing approaches assume that all information in the given KG is correct. This enables attackers to influence the output of these approaches, e.g., by perturbing the input. Consequently, the robustness of such KGE approaches has to be addressed. Recent work focused on adversarial attacks. However, non-adversarial attacks on all attack surfaces of these approaches have not been thoroughly examined. We close this gap by evaluating the impact of non-adversarial attacks on the performance of 5 state-of-the-art KGE algorithms on 5 datasets with respect to attacks on 3 attack surfaces-graph, parameter, and label perturbation. Our evaluation results suggest that label perturbation has a strong effect on the KGE performance, followed by parameter perturbation with a moderate and graph with a low effect.