Open research questions in Semantic Web and Ontologies

What the literature leaves open

• Future work will investigate lightweight architectures and efficiency optimization strategies to improve computational efficiency. Future work will explore low‑resource methods, including semi‑supervised learning, self‑supervised learning, and extraction approaches based on generative large language models, to reduce dependence on manual annotations.

  Joint Entity-Relation Extraction from Wheat Variety Certification Texts for Knowledge Graph Construction and Variety Relationship Analysis · 2026 · DOI

• Future research will pursue further breakthroughs in the following directions: (1) The framework will be extended from anomaly diagnosis and production cognition to more complex industrial scenarios, including multi-process collaborative optimization, digital twin-driven cyber–physical interaction, equipment lifecycle health assessment, intelligent scheduling, and autonomous production coordination.

  Semantic Translation and LLM-RAG Fusion of Multi-Source Heterogeneous Data for Production Cognition in Discrete Manufacturing · 2026 · DOI

• Future research will evaluate domain-oriented fine-tuning as a potential route to improving answer accuracy and reliability for complex service-performance queries. Several limitations should be noted. First, although the evaluation combines automated metrics and expert assessment, part of the scoring process still relies on LLM-assisted judgment, which may introduce evaluator bias.

  Automating Heterogeneous Creep-Data Management for Aerospace Superalloy Fasteners: A Hybrid KG-RAG Framework with Iterative Prompting and Hallucination-Mitigated QA · 2026 · DOI

• Addressing the identified limitations is essential for maximizing its potential. The broader implications of the framework highlight its significant value to the geospatial data community, impacting data integration, research, policy-making, and various application domains.

  Use of Semantic Web Technologies to Enhance the Integration and Interoperability of Environmental Geospatial Data: A Framework Based on Ontology-Based Data Access · 2025 · DOI

• Future work should focus on enhancing initial code generation accuracy to prevent error propagation.

  AutoGEEval: A Multimodal and Automated Evaluation Framework for Geospatial Code Generation on GEE with Large Language Models · 2025 · DOI

• Despite these contributions, several avenues for future work remain: Ontology Enrichment and Alignment: This involves extending MedinaOnto to include intangible heritage and urban evolution data, while deepening alignment with other linked open data sources.

  Building an Ontology for the Tangible Cultural Heritage of the Medina of Fez · 2025 · DOI

• The paper identifies a gap in the current state of digital twins in cultural heritage, highlighting the need for intelligent and sentient digital twins that can reason over complex cultural data. It recognizes the limitations of standard Named Entity Recognition (NER) systems in handling historical documents, multilingual corpora, or domain-specific ontologies.

  Knowledge Graphs and Artificial Intelligence for the Implementation of Cognitive Heritage Digital Twins · 2025 · DOI

• Such an algebra would need to account for: bounded distinguishability, structural continuity, functional continuity, causal continuity, semantic continuity, spatiotemporal continuity, institutional continuity, domain-specific admissibility rules. This is one of the most important future tasks because the Continuity Identity Principle depends on it. Continuity Metrics Relatedly, TST requires formal continuity metrics.

  Things in States Through Time: A Constructive Ontological Framework for Predicate Logic and Derived Semantic Structure · 2026 · DOI

• For 2PO, future research could investigate an improvement of Iterative Improvement (II) and Simulated Annealing (SA), such that 2PO explores the search space in a more optimal manner.

  Metaheuristics for Ontology-Based Information Extraction Rule Learning · 2026 · DOI

• Future work should investigate alternative embedding strategies or heuristics that maintain semantic grounding while reducing computational footprint.

  LLMs for industrial databases: an agro-food production plant use case · 2026 · DOI

• The integration of AAS, OPC UA, and KGs for semantic interoperability is a complex challenge. The development of more comprehensive frameworks for integrating AAS, OPC UA, and KGs is needed.

  Semantic Interoperability in Industry 4.0: A Systematic Mapping Study on Integrating Knowledge Graphs with AAS and OPC UA · 2026 · DOI

• Large-scale industrial deployment remains limited. While technical feasibility has been established, industrial deployment and best practices have not yet matured. The dominance of validation research over evaluation research indicates a need for more evaluation studies.

  Semantic Interoperability in Industry 4.0: A Systematic Mapping Study on Integrating Knowledge Graphs with AAS and OPC UA · 2026 · DOI

• The study does not provide a comprehensive comparison of all possible knowledge extraction methods. The study is limited to the field of microcontroller-based device design. The study does not provide a detailed analysis of the computational cost of the methods.

  Application of a Hybrid Approach in the Synthesis of a Knowledge Extraction Module of an Intelligent Assistant for a Microcontroller Technical Specialist · 2026 · DOI

• The need for semantic search rather than simple keyword matching. The lack of methodological support for knowledge extraction for intelligent assistants in the field of microcontroller-based device design.

  Application of a Hybrid Approach in the Synthesis of a Knowledge Extraction Module of an Intelligent Assistant for a Microcontroller Technical Specialist · 2026 · DOI

• The lack of a paradigmatic transformation of area studies in the digital age. The lack of construction of semantic infrastructure through high-quality data annotation. The lack of development of a dual-engine model integrating United Nations normative corpora and Open Source INTelligence data annotation.

  From “Textual Interpretation” to “Cognitive Infrastructure”: The Data-Driven Transformation and Paradigm Reconstruction in Country and Regional Studies for the Digital Age — With a Strategic Value Analysis of Multilingual Open Source INTelligence (OSINT) Data Annotation · 2026 · DOI

• The framework will be extended from anomaly diagnosis and production cognition to more complex industrial scenarios, including multi-process collaborative optimization, digital twin-driven cyber-physical interaction, equipment lifecycle health assessment, intelligent scheduling, and autonomous production coordination.

  Semantic Translation and LLM-RAG Fusion of Multi-Source Heterogeneous Data for Production Cognition in Discrete Manufacturing · 2026 · DOI

• The current study is conducted on a dataset constructed from a single steel smelting enterprise, and the generalizability of the method to other metallurgical equipment types or broader industrial domains requires further investigation. The information extraction task is currently limited to a small number of open-source LLMs due to constraints on computer hardware and considerations regarding data security.

  Large Language Model with Integrated Ontology and Inference Chain Constraints for Generative Information Extraction from Metallurgical Lifting Equipment Failure Reports · 2026 · DOI

• The findings of this article demonstrate that open-access models such as DeepSeek-R1- Distill-Qwen-1.5B, DeepSeek-R1-Distill-Qwen-7B, and GPT-2-Medium, when adequately fine-tuned, can deliver high-quality, context-aware recommendations that match or even surpass the performance of state-of-the-art commercial models like GPT-4.

  Smart Building Recommendations with LLMs: A Semantic Comparison Approach · 2025 · DOI

• While our proposed system demonstrates satisfactory performance in translating NL queries into Cypher, several enhancements can further improve its accuracy, efficiency, and adaptability across domains and languages.

  Domain- and Language-Adaptable Natural Language Interface for Property Graphs · 2025 · DOI

• The information extraction task is currently limited to a small number of open-source LLMs due to constraints on computer hardware and considerations regarding data security.

  Large Language Model with Integrated Ontology and Inference Chain Constraints for Generative Information Extraction from Metallurgical Lifting Equipment Failure Reports · 2026 · DOI

• Unlike previous efforts, which often focus on end-to-end pipelines or syntactic correctness, BLINKG defines a set of explicit, traceable tasks and provides reusable resources including gold standards, metrics, and diverse scenarios to enable systematic, comparable evaluation across a wide range of conditions.

  BLINKG: A Benchmark for LLM-Integrated Knowledge Graph Generation · 2026

• High computational demands, - Potential biases in training data, - Data contamination and ethical concerns, - Limited to specific LLMs…

  Fine-Tuning Large Language Models for Ontology Engineering: A Comparative Analysis of GPT-4 and Mistral · 2025 · DOI

• Continuous advancements in model architectures and training methodologies, - Exploration of LLMs' automation capabilities in complex tasks, - Investigation of LLMs' performance with different datasets…

  Fine-Tuning Large Language Models for Ontology Engineering: A Comparative Analysis of GPT-4 and Mistral · 2025 · DOI

• Elegant, restrained lines and well-controlled proportions support its use as a representative candidate. Same as above.

  A multi-representation knowledge graph approach for preserving and reasoning over Ming-style furniture · 2026 · DOI

• Investigating lightweight architectures and efficiency optimization strategies to improve computational efficiency - Exploring low-resource methods, including semi-supervised learning, self-supervised learning, and extraction approaches based on generative large language models - Including more comprehensive experiments to validate the model's…

  Joint Entity-Relation Extraction from Wheat Variety Certification Texts for Knowledge Graph Construction and Variety Relationship Analysis · 2026 · DOI

Most-cited papers in Semantic Web and Ontologies

• WordNet · Communications of the ACM · 1995 · 8,868 citations
• Critically Ill Patients With 2009 Influenza A(H1N1) Infection in Canada · JAMA · 2009 · 1,020 citations
• CYC · Communications of the ACM · 1995 · 945 citations
• Mapping knowledge structure by keyword co-occurrence: a first look at journal papers in Technology Foresight · Scientometrics · 2010 · 787 citations
• YAGO2: A spatially and temporally enhanced knowledge base from Wikipedia · Artificial Intelligence · 2012 · 762 citations
• Evaluating ontological decisions with OntoClean · Communications of the ACM · 2002 · 395 citations
• What IS-A Is and Isn't: An Analysis of Taxonomic Links in Semantic Networks · Computer · 1983 · 365 citations
• METEOR-S WSDI: A Scalable P2P Infrastructure of Registries for Semantic Publication and Discovery of Web Services · Information Technology and Management · 2005 · 279 citations
• Cobots in knowledge work · Journal of Business Research · 2020 · 275 citations
• Learning to construct knowledge bases from the World Wide Web · Artificial Intelligence · 2000 · 259 citations

Most recent work

• From manual to semi-automatic semantic annotation: about ontology-based text annotation tools · 2026
• VECTAETOS™ Canonical Ontology License v2.0 · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Semantic Fidelity Canonical Lexicon · Open MIND · 2026
• Predictive learning enables compositional representations · bioRxiv · 2026
• SPXI (Semantic Packet for eXchange & Indexing): A Formal Specification — EA-SPXI-01 · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Domain-specific SQL generation with LLMs: A hybrid framework combining knowledge graphs and retrieval-augmentation · Advanced Engineering Informatics · 2026
• A Material Passport Ontology for a circular economy · Computers in Industry · 2026
• Ontology-enhanced RAG for a personalised and sustainable food advisory system · Data & Knowledge Engineering · 2026
• Knowledge graphs generation from cultural heritage texts: combining LLMs and ontological engineering for scholarly debates · Journal of Documentation · 2026
• Automated Inference of Graph Transformation Rules · Fundamenta Informaticae · 2026

Related topics in Computer Science