Guest Editorial: Special issue on “Neuro-Symbolic Intelligence: large Language Model enabled Knowledge Engineering”

World Wide Web (2025) 28:14 https://doi.org/10.1007/s11280-024-01327-7 EDITORIAL Guest Editorial: Special issue on “Neuro-Symbolic Intelligence: large Language Model enabled Knowledge Engineering” Haofen Wang1 · Arijit Khan2 · Jun Liu3 · Michael Witbrock4 © The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024 Neuro-symbolic methods aim to integrate cutting-edge neural techniques (e.g., large language models) with symbolic approaches (e.g., knowledge engineering), offering a ‘best of both worlds’ solution. These methods have garnered increasing attention in recent years. In this special issue, the authors present state-of-the-art research and application studies. The special issue received 24 submissions, of which seven articles were ultimately accepted. Each paper underwent a thorough peer-review and revision process. Zhu et al. presented a comprehensive evaluation of Large Language Models (LLMs) for Knowledge Graph (KG) construction and reasoning across eight datasets and four tasks, finding that LLMs, especially GPT-4, excel more in reasoning than in few-shot information extraction, leading to the proposal of AutoKG, a multi-agent-based approach for KG construction and reasoning, and the creation of the VINE dataset for Virtual Knowledge Extraction. Yang et al. proposed a paradigm for the autonomous integration of data and knowledge in industrial big data, which incorporates symbolic business logic and domain knowledge into a data-driven neural model, and based on this paradigm, introduced a method for predicting operation completion time in automated container terminals. Chen et al. proposed a vision-language model with multi-granular knowledge fusion, which can integrate diverse medical knowledge (such as medical entities, definitions, and auxiliary information) to enhance performance in medical imaging tasks, demonstrating its effectiveness in reducing diagnostic errors across various applications. Zhong et al. proposed knowledge-enhanced and prompt-aware large language models to improve short-text expansion by constructing a multi-dimensional knowledge graph from domain-specific text, mining prompts across these dimensions, and matching triplets to generate expanded short-text. Haofen Wang 1 Tongji University, Shanghai, China 2 Aalborg University, Aalborg, Denmark 3 Xi’an Jiaotong University, Xi’an, China 4 University of Auckland, Auckland, New Zealand 13 14 Page 2 of 2 World Wide Web (2025) 28:14 Wu et al. introduced a new triple confidence measurement method that combines heterogeneous evidence, including explicit concept paths, neighbor concept subgraphs, and embeddings from large language models, KG embedding models, contrastive learning, and graph convolutional networks, demonstrating its superiority in error detection and link prediction across real-world datasets. Yang et al. clarified the concept of knowledge graph reliability, encompassing correctness and uncertainty, analyzing corresponding research tasks, summarize studies based on knowledge representation learning techniques, and conclude with an analysis of benchmarks and future research directions. Mao proposed a fast similarity method for ad-hoc queries with user-given meta-paths on heterogeneous information networks (HGNNs), demonstrating its effectiveness and efficiency in outperforming stateof-the-art HGNNs and path-based approaches in tasks like link prediction, node classification, and clustering. Finally, we would like to thank all the authors who contributed their submissions to this special issue. We also would like to thank the Editors in Chief, Prof. Yanchun Zhang, for accepting to have this special issue done, all the referees for their conscientious work in the review process, as well as the staff of the journal, which assisted the production. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. 13  (...truncated)