A Survey of Knowledge Graph Reasoning on Graph Types: Static, Dynamic, and Multi-Modal

IEEE Transactions on Pattern Analysis and Machine Intelligence, Journal Year: 2024, Volume and Issue: 46(12), P. 9456 - 9478

Published: June 28, 2024

Knowledge graph reasoning (KGR), aiming to deduce new facts from existing based on mined logic rules underlying knowledge graphs (KGs), has become a fast-growing research direction. It been proven significantly benefit the usage of KGs in many AI applications, such as question answering, recommendation systems, and etc. According types, KGR models can be roughly divided into three categories, i.e., static models, temporal multi-modal models. Early works this domain mainly focus KGR, recent try leverage information, which are more practical closer real-world. However, no survey papers open-source repositories comprehensively summarize discuss important To fill gap, we conduct first for tracing then KGs. Concretely, reviewed bi-level taxonomy, top-level (graph types) base-level (techniques scenarios). Besides, performances, well datasets, summarized presented. Moreover, point out challenges potential opportunities enlighten readers.

Language: Английский

---

Making Large Language Models Perform Better in Knowledge Graph Completion

Published: Oct. 26, 2024

Language: Английский

---

Learning to Denoise Biomedical Knowledge Graph for Robust Molecular Interaction Prediction

IEEE Transactions on Knowledge and Data Engineering, Journal Year: 2024, Volume and Issue: 36(12), P. 8682 - 8694

Published: Oct. 4, 2024

Molecular interaction prediction plays a crucial role in forecasting unknown interactions between molecules, such as drug-target (DTI) and drug-drug (DDI), which are essential the field of drug discovery therapeutics.Although previous methods have yielded promising results by leveraging rich semantics topological structure biomedical knowledge graphs (KGs), they primarily focused on enhancing predictive performance without addressing presence inevitable noise inconsistent semantics.This limitation has hindered advancement KGbased methods.To address this limitation, we propose BioKDN (Biomedical Knowledge Graph Denoising Network) for robust molecular prediction.BioKDN refines reliable local subgraphs denoising noisy links learnable manner, providing general module extracting task-relevant interactions.To enhance reliability refined structure, maintains consistent smoothing relations around target interaction.By maximizing mutual information smoothed relations, emphasizes informative to enable precise predictions.Experimental real-world datasets show that surpasses state-of-theart models DTI DDI tasks, confirming effectiveness robustness unreliable within contaminated KGs.

Language: Английский

---

Relational message passing with mutual information maximization for inductive link prediction

The Journal of Supercomputing, Journal Year: 2025, Volume and Issue: 81(1)

Published: Jan. 1, 2025

Language: Английский

---

Self-Supervision Improves Diffusion Models for Tabular Data Imputation

Published: Oct. 20, 2024

Incomplete tabular datasets are ubiquitous in many applications for a number of reasons such as human error data collection or privacy considerations.One would expect natural solution this is to utilize powerful generative models diffusion models, which have demonstrated great potential across image and continuous domains.However, vanilla often exhibit sensitivity initialized noises.This, along with the sparsity inherent domain, poses challenges thereby impacting robustness these imputation.In work, we propose an advanced model named Self-supervised imputation Diffusion Model (SimpDM brevity), specifically tailored tasks.To mitigate noise, introduce self-supervised alignment mechanism that aims regularize model, ensuring consistent stable predictions.Furthermore, carefully devised state-dependent augmentation strategy within SimpDM, enhancing when dealing limited data.Extensive experiments demonstrate SimpDM matches outperforms state-of-the-art methods various scenarios.

Language: Английский

---

GraphLearner: Graph Node Clustering with Fully Learnable Augmentation

Published: Oct. 26, 2024

Language: Английский

---

Chongqing GZOO: Black-box Node Injection Attack on Graph Neural Networks via Zeroth-order Optimization

IEEE Transactions on Knowledge and Data Engineering, Journal Year: 2024, Volume and Issue: 37(1), P. 319 - 333

Published: Oct. 21, 2024

Language: Английский

---

Zero-shot Knowledge Graph Question Generation via Multi-agent LLMs and Small Models Synthesis

Published: Oct. 20, 2024

Language: Английский

---

Causal Subgraph Learning for Generalizable Inductive Relation Prediction

Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Journal Year: 2024, Volume and Issue: unknown, P. 1610 - 1620

Published: Aug. 24, 2024

Inductive relation reasoning in knowledge graphs aims at predicting missing triplets involving unseen entities and/or relations. While subgraph-based methods that reason about the local structure surrounding a candidate triplet have shown promise, they often fall short accurately modeling causal dependence between triplet's subgraph and its ground-truth label. This limitation typically results susceptibility to spurious correlations caused by confounders, adversely affecting generalization capabilities. Herein, we introduce novel front-door adjustment-based approach designed learn relationship subgraphs their labels, specifically for inductive prediction. We conceptualize semantic information of as mediator employ graph data augmentation mechanism create augmented subgraphs. Furthermore, integrate fusion module decoder within adjustment framework, enabling estimation mediator's combination with also reparameterization trick model enhance robustness. Extensive experiments on widely recognized benchmark datasets demonstrate proposed method's superiority prediction, particularly tasks Additionally, reconstructed our offer valuable insights into model's decision-making process, enhancing transparency interpretability.

Language: Английский

---

Feature enhancement based on hierarchical reconstruction framework for inductive prediction on sparse graphs

Information Processing & Management, Journal Year: 2024, Volume and Issue: 62(1), P. 103894 - 103894

Published: Sept. 19, 2024

Language: Английский

---