LinkExplorer: predicting, explaining and exploring links in large biomedical knowledge graphs.

Numerous large knowledge graphs, such as DBpedia, Wikidata, Yago and Freebase, have been developed in the last decade, which contain millions of facts about various entities in the world. These knowledge graphs have proven to be incredibly useful for intelligent Web search, question understanding, in-context advertising, social media mining, and biomedicine. As some researchers have pointed out, a knowledge graph is not just a graph database, but it should have a layer of conceptual knowledge, which is usually represented as a set of first-order rules. However, it is challenging to automatically extract first-order rules from large knowledge graphs. In particular, traditional models are usually unable to handle rule learning in large knowledge graphs. This chapter aims to present state-of-the-art techniques and models on learning first-order rules using representation learning. By first recalling some basics of rule learning in knowledge graphs, we will introduce useful techniques of embedding-based rule learning through major models such as RLvLR and TyRuLe, which embed paths in knowledge graphs into latent spaces for candidate rule search. Then we evaluate the rule learning efficiency and the quality of automatically learned rules by applying them in link prediction. Before the chapter is concluded, we will also discuss some future research problems in the area.

Learning structured embeddings of knowledge graphs with generative adversarial framework

Knowledge Graphs (KGs) represent heterogeneous domain knowledge on the Web and within organizations. There exist shapes constraint languages to define validating shapes to ensure the quality of the data in KGs. Existing techniques to extract validating shapes often fail to extract complete shapes, are not scalable, and are prone to produce spurious shapes. To address these shortcomings, we propose the Quality Shapes Extraction (QSE) approach to extract validating shapes in very large graphs, for which we devise both an exact and an approximate solution. QSE provides information about the reliability of shape constraints by computing their confidence and support within a KG and in doing so allows to identify shapes that are most informative and less likely to be affected by incomplete or incorrect data. To the best of our knowledge, QSE is the first approach to extract a complete set of validating shapes from WikiData. Moreover, QSE provides a 12x reduction in extraction time compared to existing approaches, while managing to filter out up to 93% of the invalid and spurious shapes, resulting in a reduction of up to 2 orders of magnitude in the number of constraints presented to the user, e.g., from 11,916 to 809 on DBpedia.

A variety of knowledge graph embedding approaches have been developed. Most of them obtain embeddings by learning the structure of the knowledge graph within a link prediction setting. As a result, the embeddings reflect only the structure of a single knowledge graph, and embeddings for different knowledge graphs are not aligned, e.g., they cannot be used to find similar entities across knowledge graphs via nearest neighbor search. However, knowledge graph embedding applications such as entity disambiguation require a more global representation, i.e., a representation that is valid across multiple sources. We propose to learn universal knowledge graph embeddings from large-scale interlinked knowledge sources. To this end, we fuse large knowledge graphs based on the owl:sameAs relation such that every entity is represented by a unique identity. We instantiate our idea by computing universal embeddings based on DBpedia and Wikidata yielding embeddings for about 180 million entities, 15 thousand relations, and 1.2 billion triples. We believe our computed embeddings will support the emerging field of graph foundation models. Moreover, we develop a convenient API to provide embeddings as a service. Experiments on link prediction suggest that universal knowledge graph embeddings encode better semantics compared to embeddings computed on a single knowledge graph. For reproducibility purposes, we provide our source code and datasets open access.

Triple Pattern Fragments (TPFs) are a novel interface for accessing data in knowledge graphs on the web. So far, work on performance evaluation and optimization has focused mainly on SPARQL query execution over TPF servers. However, in order to devise querying techniques that efficiently access large knowledge graphs via TPFs, we need to identify and understand the variables that influence the performance of TPF servers on a fine-grained level. In this work, we assess the performance of TPFs by measuring the response time for different requests and analyze how the requests’ properties, as well as the TPF server configuration, may impact the performance. For this purpose, we developed the Triple Pattern Fragment Profiler to determine the performance of TPF server. The resource is openly available at https://doi.org/10.5281/zenodo.1211621. To this end, we conduct an empirical study over four large knowledge graphs in different server environments and configurations. As part of our analysis, we provide an extensive evaluation of the results and focus on the impact of the variables: triple pattern type, answer cardinality, page size, backend and the environment type on the response time. The results suggest that all variables impact on the measured response time and allow for deriving suggestions for TPF server configurations and query optimization.

Weakly supervised semantic instance segmentation with only image-level supervision, instead of relying on expensive pixel-wise masks or bounding box annotations, is an important problem to alleviate the data-hungry nature of deep learning. In this article, we tackle this challenging problem by aggregating the image-level information of all training images into a large knowledge graph and exploiting semantic relationships from this graph. Specifically, our effort starts with some generic segment-based object proposals (SOP) without category priors. We propose a multiple instance learning (MIL) framework, which can be trained in an end-to-end manner using training images with image-level labels. For each proposal, this MIL framework can simultaneously compute probability distributions and category-aware semantic features, with which we can formulate a large undirected graph. The category of background is also included in this graph to remove the massive noisy object proposals. An optimal multi-way cut of this graph can thus assign a reliable category label to each proposal. The denoised SOP with assigned category labels can be viewed as pseudo instance segmentation of training images, which are used to train fully supervised models. The proposed approach achieves state-of-the-art performance for both weakly supervised instance segmentation and semantic segmentation. The code is available at https://github.com/yun-liu/LIID.

The development of entirely new drugs for emerging diseases is challenging, time consuming and very costly. This underscores the importance of drug repurposing (DR), where treatments are found among clinically approved drugs for other diseases. DR gained more popularity with the emergence of the SARS-COV-2 virus. The current pandemic almost halted normal life in most parts of the world. Despite global research efforts on this task, no effective treatment has yet been developed. This research study is an effort to leverage deep learning techniques in network medicine to screen potential therapeutic drugs for COVID-19. We treat drug repurposing as a link prediction task on a large heterogeneous high-quality knowledge graph containing interactions between medical entities like drugs, diseases and genes from various resources. Using Graph Neural Network (GNN) based on GraphSAGE model, we were able to achieve AUC of 0.97. We highlighted the top 5 new drugs for repurposing for COVID-19 including Flurithromycin, Radotinib and Memantine, and predicted 7 drugs currently in clinical trials ranked among the top 100 drugs. The experimental results confirm that the model can effectively predict the drug-disease interaction for COVID-19.KeywordsGraph neural networkKnowledge graphDrug repurposingNetwork medicineCOVID-19 treatment

Inferring missing links in knowledge graphs (KG) has attracted a lot of attention from the research community. In this paper, we tackle a practical query answering task involving predicting the relation of a given entity pair. We frame this prediction problem as an inference problem in a probabilistic graphical model and aim at resolving it from a variational inference perspective. In order to model the relation between the query entity pair, we assume that there exists an underlying latent variable (paths connecting two nodes) in the KG, which carries the equivalent semantics of their relations. However, due to the intractability of connections in large KGs, we propose to use variation inference to maximize the evidence lower bound. More specifically, our framework (Diva) is composed of three modules, i.e. a posterior approximator, a prior (path finder), and a likelihood (path reasoner). By using variational inference, we are able to incorporate them closely into a unified architecture and jointly optimize them to perform KG reasoning. With active interactions among these sub-modules, Diva is better at handling noise and coping with more complex reasoning scenarios. In order to evaluate our method, we conduct the experiment of the link prediction task on multiple datasets and achieve state-of-the-art performances on both datasets.

This study aims to construct a knowledge syst em for Traditional Chinese Patent Medicine (TCPM) based on Large Language Model (LLM) and Knowledge Graph (K G) to address the complexity of TCPM knowledge and its <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$c$</tex> linical application challenges. It seeks to solve issues inhere nt in traditional TCPM knowledge utilization, such as infor mation silos, retrieval difficulties, and the lack of effective decision support. By integrating multi-source heterogeneous data, including the Pharmacopoeia of the People's Republic of China and various medical information platforms, this re search utilizes an LLM for efficient knowledge extraction a nd fusion to build the Traditional Chinese Patent Medicine Knowledge Graph (TCPM-KG). Experimental results show <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$t$</tex> hat the model achieved an F1-score of 0.9553 in the Name d Entity Recognition(NER) task, indicating that LLM can ef fectively process complex texts in the TCPM domain and si gnificantly enhance the automation of knowledge extraction. To date, TCPM-KG contains 1,607 TCPM instances and 77, 100 knowledge triplets. The advantages of this study lie in <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$i$</tex> ts integration of multi-source heterogeneous data to create a KG with broad coverage and fine granularity, and its prop osal of a “KG+LLM” methodology that spans the entire co nstruction process. This work successfully demonstrates the feasibility and efficiency of using an LLM as a core techno logy for building a TCPM-KG, providing a new technical <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p$</tex> ath and foundation for the modernization and intelligent de velopment of Traditional Chinese Medicine.

This paper focuses on building an innovative industrial financial risk rapid discovery model based on large language model and knowledge graph inference. Firstly, this work deeply analyzes the diversity, concealment and rapid propagation of industrial financial risks. Secondly, it points out the limitations of traditional risk assessment methods when dealing with massive and heterogeneous financial data. Then, it designs an integrated strategy to integrate the powerful natural language processing and understanding ability of large language models. Combined with the advantages of knowledge graph in structured information representation, this model analyzes financial text data through a large language model. Hence, key information can be automatically extracted and a preliminary risk knowledge base can be built. Finally, an industrial financial network can be established to realize correlation mining and dynamic evolution analysis of risk factors. The experimental results show that the model can effectively capture risk signals in the industrial chain, and detect potential financial risks in advance. This study not only enriches the theoretical framework of industrial financial risk management, but also provides strong technical support for risk prevention and control in practice.

In recent years, Large Language Models (LLMs) have rapidly advanced in language understanding, reasoning, and generation, and are increasingly adopted as the “brain” of industrial intelligent systems. Nevertheless, in high‑risk and strongly regulated domains they still exhibit hallucination, weak domain grounding, limited interpretability, and privacy as well as security constraints. Knowledge graphs (KGs) encode domain entities, relations, rules, and events explicitly, providing controllable semantics and an explainable reasoning substrate. Retrieval‑augmented generation (RAG) injects external evidence into LLM prompting, while GraphRAG further introduces graph indexing and community‑level retrieval to preserve global structure and support multi‑hop reasoning. This review summarizes the evolution of LLMs, KG modeling and extraction, GraphRAG mechanisms, and a general fusion framework. Typical industrial applications are surveyed, and a coal mine flood emergency plan generation and evaluation approach is discussed to illustrate the practical value of graph‑grounded large models. KG‑enhanced retrieval also supports provenance tracking, allowing industrial users to audit the evidence behind model outputs.

A survey of large language model-augmented knowledge graphs for advanced complex product design

AIREG: Enhanced Educational Recommender System with Large Language Models and Knowledge Graphs

Scheduling satellite tasks requires intricate coordination of numerous interconnected activities, posing challenges for ground station operators who struggle to decipher autonomous decision-making systems without explanatory responses. This paper advocates for explainable artificial intelligence to address this gap, leveraging large language models and knowledge graphs to enhance the transparency of scheduling logic. The study explores how operators can retrieve relevant information through natural language queries by integrating language processing with knowledge graphs as core data structures. This approach allows for manual access to knowledge graphs, supplemented with textual explanations, and enables operators to explore different scheduling scenarios, thus improving system robustness and adaptability. Additionally, this paper analyzes how current knowledge graphs and natural language processing techniques are utilized and how they can enhance explainability in satellite scheduling. It investigates the modeling of satellite schedules and environmental data within the knowledge graph. The study introduces a novel approach for generating categorized queries, executable code for knowledge graph data extraction via language modeling, and providing explanatory answers. These are evaluated for correctness, validity, and linguistic quality, demonstrating the effectiveness of this approach.

We describe the integration of a Large Language Model (LLM) and a knowledge graph (KG) in an application of medical informatics. The LLM is BioBERT, a pre-trained biomedical language representation model for biomedical text mining. The task, prediction of Type 2 Diabetes (T2D) in the UK Biobank cohort, a population-based prospective cohort study aimed at improving the prevention, diagnosis and treatment of a wide range of serious diseases, including diabetes. Patient history is represented as date-stamped ICD-10 codes, a hierarchical classification of diseases (the KG). The connection between the LLM and the KG is precisely the shared medical vocabulary used in the ICD-10 descriptions and the resources used in fine tuning the LLMs. We explore the impact of the various levels of the KG hierarchy (subcategories, categories, blocks, and chapters) on the performance of classifiers developed for T2D.