[Paper Review] CEVO: Comprehensive EVent Ontology Enhancing Cognitive Annotation

Saeedeh Shekarpour, Faisal Alshargi+3 more|arXiv (Cornell University)|Jan 19, 2017

Natural Language Processing Techniques11 references3 citations

TL;DR

This paper introduces CEVO, a comprehensive event ontology derived from Beth Levin’s conceptual hierarchy of English verbs, to enhance cognitive annotation of relations in text and ontologies. By mapping verbs to abstract semantic and syntactic classes—such as 'Communication' or 'Amalgamate'—CEVO enables consistent relation annotation, alignment across diverse vocabularies, and linking of textual relations to ontological properties, demonstrated through three use cases with real-world data.

ABSTRACT

While the general analysis of named entities has received substantial research attention on unstructured as well as structured data, the analysis of relations among named entities has received limited focus. In fact, a review of the literature revealed a deficiency in research on the abstract conceptualization required to organize relations. We believe that such an abstract conceptualization can benefit various communities and applications such as natural language processing, information extraction, machine learning, and ontology engineering. In this paper, we present Comprehensive EVent Ontology (CEVO), built on Levin's conceptual hierarchy of English verbs that categorizes verbs with shared meaning, and syntactic behavior. We present the fundamental concepts and requirements for this ontology. Furthermore, we present three use cases employing the CEVO ontology on annotation tasks: (i) annotating relations in plain text, (ii) annotating ontological properties, and (iii) linking textual relations to ontological properties. These use-cases demonstrate the benefits of using CEVO for annotation: (i) annotating English verbs from an abstract conceptualization, (ii) playing the role of an upper ontology for organizing ontological properties, and (iii) facilitating the annotation of text relations using any underlying vocabulary. This resource is available at https://shekarpour.github.io/cevo.io/ using https://w3id.org/cevo namespace.

Motivation & Objective

• Address the lack of abstract conceptualization for organizing relations in unstructured and structured data.
• Overcome deficiencies in relation extraction, contextual equivalencing of relations, and heterogeneity across ontologies.
• Provide a unified, psychologically principled framework for annotating relations based on semantic and syntactic verb classes.
• Enable interoperability between textual relations and ontological properties through a shared abstract representation.
• Facilitate cognitive annotation in NLP, information extraction, and ontology engineering by leveraging Levin’s verb hierarchy.

Proposed method

• Construct CEVO by mapping over 3,000 English verbs into 230+ semantic-syntactic classes based on Beth Levin’s conceptual hierarchy of verbs.
• Use the NIF (Named Individual Format) vocabulary to annotate textual mentions of verbs with their positions and syntactic roles.
• Apply the Web Annotation Data Model (WADM) to link textual relations to ontological properties via shared CEVO event classes.
• Assign CEVO event classes (e.g., cevo:Communication, cevo:Amalgamate) to verb instances in text and ontologies to enable semantic alignment.
• Use SPARQL queries to link textual relations (e.g., 'marry') to corresponding ontological properties (e.g., dbp:spouse) through shared CEVO annotations.
• Leverage the CEVO namespace (https://w3id.org/cevo/) to ensure persistent, resolvable URIs for event classes and their instances.

Experimental results

Research questions

• RQ1 How can an abstract conceptualization of relations improve the consistency and cognitive plausibility of relation annotation in text and ontologies?
• RQ2 To what extent can Levin’s verb classification serve as a foundation for a reusable, upper-level event ontology?
• RQ3 Can CEVO enable contextual equivalencing of relations across diverse vocabularies and ontologies?
• RQ4 How effectively can CEVO support the linking of textual relations to ontological properties in real-world data?
• RQ5 What is the impact of using CEVO on relation extraction and interoperability in knowledge graph applications?

Key findings

• CEVO successfully maps over 3,000 English verbs into 230+ semantically coherent event classes based on shared meaning and syntactic behavior.
• Verbs like 'announce' and 'say' are semantically equated under the CEVO class cevo:Communication despite lacking lexical synonymy, demonstrating cognitive consistency.
• The use of WADM enables direct linking of textual relations (e.g., 'marry') to ontological properties (e.g., dbp:spouse) via shared CEVO event classes.
• SPARQL queries can resolve the link between textual relations and ontological properties using CEVO annotations, as shown with the verb 'marry' and the property dbp:spouse.
• CEVO enables consistent annotation of relations across diverse data sources and ontologies, reducing ambiguity and improving interoperability.
• The ontology is publicly available at https://shekarpour.github.io/cevo.io/ with the namespace https://w3id.org/cevo/, supporting long-term reuse and integration.