COntExt: Towards Context-Aware Ontology Extension from Operational Metrics

Noosaga Placements

• Knowledge Representation Reasoningsubfield95%
  The paper focuses on ontology extension, which is a core topic in knowledge representation and reasoning. It deals with formal ontologies, class hierarchies, properties, and relationships, and proposes methods to extend them.
  COntExt, a framework for context-aware ontology extension that takes structured metric definitions as input and suggests how referenced concepts and properties should be integrated into an existing ontologyThe framework defines the extension problem as three sub-tasks: parent class prediction, relation type prediction, and data property assignment.
• Ontologiesframework95%
  The paper is fundamentally about ontologies and their extension. It operates on formal ontologies and aims to extend them, so it is situated within the Ontologies framework.
  COntExt, a framework for context-aware ontology extension that takes structured metric definitions as input and suggests how referenced concepts and properties should be integrated into an existing ontologyThe framework defines the extension problem as three sub-tasks: parent class prediction, relation type prediction, and data property assignment.
• Machine Learningsubfield70%
  The paper evaluates various machine learning algorithms for the three sub-tasks, including BERT fine-tuning, TaxoExpan, TransE, and MLM prompting. The focus is on applying and comparing these algorithms for ontology extension.
  We evaluate different algorithms for each task.BERT Fine-tuning (Liu et al., 2020). A BERT model is fine-tuned as a binary classifier on (child, parent) concept pairs from the ontology hierarchy.TaxoExpan (Shen et al., 2020). A position-enhanced Graph Attention Network encodes local ego-graphs around each candidate parent position.
• Deep Learningframework70%
  The paper employs deep learning models such as BERT, TaxoExpan (GNN), and BertConvE, which are deep learning architectures.
  BERT Fine-tuning (Liu et al., 2020). A BERT model is fine-tuned as a binary classifier on (child, parent) concept pairs from the ontology hierarchy.TaxoExpan (Shen et al., 2020). A position-enhanced Graph Attention Network encodes local ego-graphs around each candidate parent position.BertConvE (Liu et al., 2022). A small BERT model is trained from scratch on random walks extracted from the ontology’s graph structure.
• Natural Language Processingsubfield60%
  The framework uses language models and text-based context enrichment, such as BERT and SentenceTransformer, to incorporate metric descriptions and ontology annotations. This involves natural language processing techniques.
  To evaluate context enrichment, we also implement unsupervised heuristics on top of pre-trained language models, such as BERT and SentenceTransformerPretrained MLM prompting. For each class pair (ci , c j ), the framework constructs sentences in both directions (“ci [MASK] c j ” and “c j [MASK] ci ”) and scores each candidate relation by its probability at the masked position using distilbert-base-uncased pretrained MLM.
• Statistical Learningframework60%
  The paper uses supervised machine learning methods for the sub-tasks, which fall under statistical learning. For example, BERT fine-tuning and TransE are statistical learning approaches.
  BERT Fine-tuning (Liu et al., 2020). A BERT model is fine-tuned as a binary classifier on (child, parent) concept pairs from the ontology hierarchy.TransE (Bordes et al., 2013). A translating-embedding model learns entity and relation vectors from the ontology’s existing triples such that h + r ≈ t for valid triples (h, r,t).
• Instance-Based Learningframework50%
  The ChildAgg heuristic uses sentence embeddings and cosine similarity to find similar classes, which is an instance-based approach.
  Child aggregation of SentenceTransformer embedding (ChildAgg). All ontology class names are embedded into dense vectors using a sentencetransformer (Reimers and Gurevych, 2019) model – we use all-MiniLM-L6-v2 model. For each class, we replace its embedding with the mean of its own and its children’s embeddings. For a candidate class, the framework computes cosine similarity against all existing class embeddings and returns the top-k most similar classes as parent suggestions.

Abstract

Organizations increasingly define operational metrics in structured, machine-readable formats to monitor systems, processes, and compliance. These metric definitions implicitly encode domain knowledge, such as referencing concepts, properties, and relationships, that often extends what is captured in formal ontologies. Yet the connection between operational metric catalogues and ontological knowledge remains manual, ad-hoc, and labor-intensive. We present COntExt, a framework for context-aware ontology extension that takes structured metric definitions as input and suggests how referenced concepts and properties should be integrated into an existing ontology, utilizing the context of these metrics. The framework defines the extension problem as three sub-tasks: parent class prediction, relation type prediction, and data property assignment. Across four cybersecurity ontologies, we evaluate different algorithms for each task. Our results show that metric-derived context improves the suggestions over ontology-context baselines for relation type prediction and data property assignment. Our work demonstrates that operational metric catalogues are a practical and underexploited source for ontology extension. This work enables organizations to maintain their ontologies at a significantly lower cost than manual engineering.

Paper Context