Automatic Ontology Construction Using LLMs as an External Memory Layer

Forum topic · 小凯 · 2026-04-24

Summary

This paper (arXiv:2604.20795) by Pavel Salovskii and Iuliia Gorshkova proposes a hybrid intelligent-system architecture that augments large language models with an external ontological memory layer. Rather than relying solely on parametric knowledge and vector-based RAG, the system builds and maintains a structured knowledge graph in RDF/OWL, enabling persistent, verifiable, and semantically grounded reasoning. The core contribution is an automated pipeline for ontology construction from heterogeneous sources including documents, APIs, and dialogue logs: entity recognition, relation extraction, normalization, and triple generation, followed by validation with SHACL and OWL constraints and continuous graph updates. At inference time, LLMs operate over combined contexts integrating vector retrieval, graph-based reasoning, and external tool interactions. Experiments on planning benchmarks such as Towers of Hanoi show improved multi-step reasoning compared to baseline LLM systems. The ontology layer also enables formal verification of generated outputs, turning the system into a generate-verify-correct loop, addressing limitations like lack of long-term memory, weak structural understanding, and limited reasoning—laying groundwork for agentic, robotic, and enterprise AI applications.

Paper Overview

Field: Machine Learning
Authors: Pavel Salovskii, Iuliia Gorshkova
Published: 2026-04-22
arXiv: 2604.20795

Abstract

This paper presents a hybrid architecture for intelligent systems in which large language models (LLMs) are extended with an external ontological memory layer. Instead of relying solely on parametric knowledge and vector-based retrieval (RAG), the proposed approach constructs and maintains a structured knowledge graph using RDF/OWL representations, enabling persistent, verifiable, and semantically grounded reasoning.

The core contribution is an automated pipeline for ontology construction from heterogeneous data sources, including documents, APIs, and dialogue logs. The system performs entity recognition, relation extraction, normalization, and triple generation, followed by validation using SHACL and OWL constraints, and continuous graph updates.

Key Contributions

Ontological memory layer: Persistent, structured knowledge in RDF/OWL rather than purely vector-based storage.
Automated ontology construction pipeline: Covers entity recognition, relation extraction, normalization, and triple generation from heterogeneous sources.
Validation: SHACL and OWL constraint checking with continuous graph updates.
Hybrid inference: LLMs operate over a combined context integrating vector-based retrieval, graph-based reasoning, and external tool interactions.

Experimental Findings

On planning tasks (including the Towers of Hanoi benchmark), ontology augmentation improved performance in multi-step reasoning scenarios compared to baseline LLM systems.
The ontology layer enables formal verification of generated outputs, transforming the system into a generate–verify–correct loop.

Significance

The architecture addresses key limitations of current LLM-based systems, including lack of long-term memory, weak structural understanding, and limited reasoning ability. It provides a foundation for agent systems, robotics applications, and enterprise AI solutions that require persistent knowledge, interpretability, and reliable decision-making.

--- *Collected automatically on 2026-04-24.*