Revisiting Catastrophic Forgetting in Continual Knowledge Graph Embedding

• URL: http://arxiv.org/abs/2604.19401v1
• Date: Tue, 21 Apr 2026 12:26:56 GMT
• Title: Revisiting Catastrophic Forgetting in Continual Knowledge Graph Embedding
• Abstract summary: Knowledge Graph Embeddings (KGEs) support a wide range of downstream tasks over Knowledge Graphs (KGs)<n>Current CKGE approaches address catastrophic forgetting primarily by limiting changes to existing embeddings.<n>When new entities are introduced, their embeddings can interfere with previously learned ones, causing the model to predict them in place of previously correct answers.
• Score: 2.422410293747519
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Knowledge Graph Embeddings (KGEs) support a wide range of downstream tasks over Knowledge Graphs (KGs). In practice, KGs evolve as new entities and facts are added, motivating Continual Knowledge Graph Embedding (CKGE) methods that update embeddings over time. Current CKGE approaches address catastrophic forgetting (i.e., the performance degradation on previously learned tasks) primarily by limiting changes to existing embeddings. However, we show that this view is incomplete. When new entities are introduced, their embeddings can interfere with previously learned ones, causing the model to predict them in place of previously correct answers. This phenomenon, which we call entity interference, has been largely overlooked and is not accounted for in current CKGE evaluation protocols. As a result, the assessment of catastrophic forgetting becomes misleading, and CKGE methods performance is systematically overestimated. To address this issue, we introduce a corrected CKGE evaluation protocol that accounts for entity interference. Through experiments on multiple benchmarks, we show that ignoring this effect can lead to performance overestimation of up to 25%, particularly in scenarios with significant entity growth. We further analyze how different CKGE methods and KGE models are affected by the different sources of forgetting, and introduce a catastrophic forgetting metric tailored to CKGE.

Related papers

• Beyond Homophily: Towards Generalized Graph Reconstruction Attack and Defense [80.28330585803133]
  Graph neural networks (GNNs) are widely deployed on computation data, yet they can leak sensitive or proprietary information about the training graph adjacency.<n>This work studies graph reconstruction attacks (GRA), a form of model inversion that reconstructs the training adjacency from a trained GNN.<n>We propose MC-GRA (+), which reconstructs the adjacency by optimizing a surrogate adjacency whose GNN-induced representations align with those of the target model.<n>On the defense side, we propose MC-GPB (+), which suppresses adjacency-dependent information throughout the representation chain
  arXiv  Detail & Related papers  (2026-06-06T09:29:27Z)
• Temporal Motif-aware Graph Test-time Adaptation for OOD Blockchain Anomaly Detection [55.18577006122169]
  We propose a novel framework termed textbfTEmporal textbfMotif-aware textbfGraph textbfTest-textbfTime textbfAdaptation (textbfTEMG-TTA)<n>First, we comprehensively capture the 3-node temporal motif distribution of each active address using an efficient computational mechanism.<n>Second, we design a simple yet effective test-time adaptation strategy to facilitate the sharing of common patterns between training and testing graphs.
  arXiv  Detail & Related papers  (2026-05-28T07:43:20Z)
• Learning to Evolve: Bayesian-Guided Continual Knowledge Graph Embedding [20.479556500981044]
  Key challenge facing continual knowledge graph embedding (CKGE) is that the model is prone to "catastrophic forgetting"<n>In order to effectively alleviate this problem, we propose a new CKGE model BAKE.
  arXiv  Detail & Related papers  (2025-08-04T13:46:33Z)
• SINDER: Repairing the Singular Defects of DINOv2 [61.98878352956125]
  Vision Transformer models trained on large-scale datasets often exhibit artifacts in the patch token they extract. We propose a novel fine-tuning smooth regularization that rectifies structural deficiencies using only a small dataset.
  arXiv  Detail & Related papers  (2024-07-23T20:34:23Z)
• Performance Evaluation of Knowledge Graph Embedding Approaches under Non-adversarial Attacks [1.6986898305640263]
  We evaluate the impact of non-adversarial attacks on the performance of 5 state-of-the-art Knowledge Graph Embedding (KGE) algorithms. Label perturbation has a strong effect on the KGE performance, followed by parameter perturbation with a moderate and graph with a low effect.
  arXiv  Detail & Related papers  (2024-07-09T13:42:14Z)
• Counterfactual Reasoning with Knowledge Graph Embeddings [3.6311338398148534]
  Knowledge graph embeddings (KGEs) were originally developed to infer true but missing facts in incomplete knowledge repositories. In this paper, we link knowledge graph completion and counterfactual reasoning via our new task CFKGR.
  arXiv  Detail & Related papers  (2024-03-11T17:21:39Z)
• CausE: Towards Causal Knowledge Graph Embedding [13.016173217017597]
  Knowledge graph embedding (KGE) focuses on representing the entities and relations of a knowledge graph (KG) into the continuous vector spaces. We build the new paradigm of KGE in the context of causality and embedding disentanglement. We propose a Causality-enhanced knowledge graph Embedding (CausE) framework.
  arXiv  Detail & Related papers  (2023-07-21T14:25:39Z)
• GIF: A General Graph Unlearning Strategy via Influence Function [63.52038638220563]
  Graph Influence Function (GIF) is a model-agnostic unlearning method that can efficiently and accurately estimate parameter changes in response to a $epsilon$-mass perturbation in deleted data. We conduct extensive experiments on four representative GNN models and three benchmark datasets to justify GIF's superiority in terms of unlearning efficacy, model utility, and unlearning efficiency.
  arXiv  Detail & Related papers  (2023-04-06T03:02:54Z)
• Studying How to Efficiently and Effectively Guide Models with Explanations [52.498055901649025]
  'Model guidance' is the idea of regularizing the models' explanations to ensure that they are "right for the right reasons" We conduct an in-depth evaluation across various loss functions, attribution methods, models, and 'guidance depths' on the PASCAL VOC 2007 and MS COCO 2014 datasets. Specifically, we guide the models via bounding box annotations, which are much cheaper to obtain than the commonly used segmentation masks.
  arXiv  Detail & Related papers  (2023-03-21T15:34:50Z)
• GraphCoCo: Graph Complementary Contrastive Learning [65.89743197355722]
  Graph Contrastive Learning (GCL) has shown promising performance in graph representation learning (GRL) without the supervision of manual annotations. This paper proposes an effective graph complementary contrastive learning approach named GraphCoCo to tackle the above issue.
  arXiv  Detail & Related papers  (2022-03-24T02:58:36Z)
• Rethinking Graph Convolutional Networks in Knowledge Graph Completion [83.25075514036183]
  Graph convolutional networks (GCNs) have been increasingly popular in knowledge graph completion (KGC) In this paper, we build upon representative GCN-based KGC models and introduce variants to find which factor of GCNs is critical in KGC. We propose a simple yet effective framework named LTE-KGE, which equips existing KGE models with linearly transformed entity embeddings.
  arXiv  Detail & Related papers  (2022-02-08T11:36:18Z)
• Evaluating Prediction-Time Batch Normalization for Robustness under Covariate Shift [81.74795324629712]
  We call prediction-time batch normalization, which significantly improves model accuracy and calibration under covariate shift. We show that prediction-time batch normalization provides complementary benefits to existing state-of-the-art approaches for improving robustness. The method has mixed results when used alongside pre-training, and does not seem to perform as well under more natural types of dataset shift.
  arXiv  Detail & Related papers  (2020-06-19T05:08:43Z)
• Evaluating the Calibration of Knowledge Graph Embeddings for Trustworthy Link Prediction [16.553439385317876]
  We investigate the trustworthiness of predictions made by knowledge graph embedding (KGE) models. We first conduct an evaluation under the standard closed-world assumption (CWA), in which predicted triples not already in the knowledge graph are considered false. We then introduce the more realistic but challenging open-world assumption (OWA), in which unobserved predictions are not considered true or false until ground-truth labels are obtained.
  arXiv  Detail & Related papers  (2020-04-02T17:46:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.