Focus Set Subontologies and Semantic Differences for Large ELH Ontologies

Ontologies have been widely used in a variety of fields to serve as sources for formally structured knowledge. The Systemized Nomenclature of Medicine - Clinical Terms (SNOMED CT) ontology, among others, is widely used in the (bio-)medical domain, as it provides a comprehensive multilingual vocabulary for capturing all aspects of electronic health records and clinical knowledge, resulting in a very large and complex ontology. Ontology extraction methods enable efficient use of such large ontologies by splitting them into interconnected smaller parts. This thesis presents a novel method for extracting focus set subontologies from ELH ontologies for a set of symbols selected by the user. The resulting subontologies satisfy the requirements sought by SNOMED CT users, including providing complete semantics for the description of input symbols while being concise and conforming to SNOMED CT modelling principles. The findings show that, in comparison to locality-based modularisation and uniform interpolation, the resulting subontologies satisfy the requirements of SNOMED CT in terms of size, encapsulating the entire semantics of the definitions solely for the set of input symbols, and retaining the original ontology's structure. This thesis also investigates the computation of semantic differences between extracted subontologies using a combination of our notion of subontologies and the uniform interpolation-based semantic difference method in order to constrain the development of such differences to a particular subdomain of the ontology identified by the user. The findings demonstrate that our method is capable of revealing differences in the meaning of focus concept definitions associated with a particular subdomain of the ontology, where some of these differences would not have been generated without this focused approach to tracking semantic differences between ontologies.

To access the thesis, click here.

Focus Set Subontology Extraction

Project Duration: Feb 2020 - Dec 2020

This project targeted the SNOMED CT medical ontology. As a continuation of our collaboration with SNOMED International on extracting focused extracts for their subsets, I devised a novel method for generating focus set subontologies in this research. The resulting subontologies are more beneficial and relevant to SNOMED International.

To access the implementation and the data of the project, click here.

The following publications arose from the project:

1. Ghadah Alghamdi, Renate A. Schmidt, Warren Del-Pinto, and Yongsheng Gao. Upwardly Abstracted Definition-Based Subontologies. In Proceedings of the 34th International Workshop on Description Logics (DL 2021) part of Bratislava Knowledge September (BAKS 2021), Bratislava, Slovakia, September 19th to 22nd, 2021, volume 2954 of CEUR Workshop Proceedings. CEUR-WS.org, 2021.
2. Ghadah Alghamdi, Renate A. Schmidt, Warren Del-Pinto, and Yongsheng Gao. Upwardly Abstracted Definition-Based Subontologies. In K-CAP’21: Knowledge Capture Conference, pages 209–216. ACM, 2021.

The work [1] includes our initial results of the work in [2].

Tracking Logical Difference between ELH Ontologies

Project Duration: Dec 2020 - Feb 2021

As a continuation of the project that started with Babylon Health, we collaborated with Nanjing University on the development of a tool for tracking semantic differences across versions of ELH ontologies. I conducted the experiments of generating focused semantic differences related to a standard subset of SNOMED CT, wrote the results and provided overall feedback on the work.

To access the implementation and the data of the project, click here.

The following publications arose from the project:

3. Zhao Liu, Chang Lu, Ghadah Alghamdi, Renate A. Schmidt, and Yizheng Zhao. Tracking Semantic Evolutionary Changes in Large-Scale Ontological Knowledge Bases. In CIKM’21: The 30th ACM International Conference on Information and Knowledge Management, Virtual Event, Queensland, Australia, November 1 - 5, 2021, pages 1130-1139, ACM, 2021.
4. Zhao Liu, Chang Lu, Ghadah Alghamdi, Renate A. Schmidt, and Yizheng Zhao. Tracking Semantic Evolutionary Changes in Large-Scale Ontological Knowledge Bases. In Proceedings of the 34th International Workshop on Description Logics (DL 2021) part of Bratislava Knowledge September (BAKS 2021), Bratislava, Slovakia, September 19th to 22nd, 2021, volume 2954 of CEUR Workshop Proceedings. CEUR-WS.org, 2021.

Ontology Modularity and Forgetting

Project Duration: Feb 2019 - Sep 2019

In this project, we worked closely with SNOMED International and proposed a workflow of computing smaller ontology extracts based on ontology modularity and forgetting from SNOMED CT. To extract such extracts, we used a core subset of the SNOMED CT medical ontology, which was part of a quality improvement project at SNOMED International:

To access the implementation and the data of the project, click here.

The following publications arose from the project:

5. Jieying Chen, Ghadah Alghamdi, Renate A. Schmidt, Dirk Walther, and Yongsheng Gao. Modularity Meets Forgetting: A Case Study with the SNOMED CT Ontology. In Proceedings of the 32nd International Workshop on Description Logics, volume 2373. CEUR-WS.org, 2019.
6. Jieying Chen, Ghadah Alghamdi, Renate A. Schmidt, Dirk Walther, and Yongsheng Gao. Ontology Extraction for Large Ontologies via Modularity and Forgetting. In Proceedings of the 10th International Conference on Knowledge Capture, K-CAP’19, pages 45–52. ACM, 2019.

Logical Difference Case Study Project

Project Duration: Feb 2018 - Aug 2018

I was in a project that involved collaborating with Babylon Health to a case study that sought to find differences that might exist between several SNOMED CT ontology versions; international (core) and country extensions including the candian and australian extensions. For Babylon Health, detecting these differences was an important issue to make certain that integrating new information from SNOMED CT to their knowledge base is done in a safe manner. I was responsible for analysing the resulting logical differences between SNOMED CT core versions and the country extensions.

The project resulted in the following publications:

7. Giorgos Stoilos, David Geleta, Szymon Wartak, Sheldon Hall, Mohammad Khodadadi, Yizheng Zhao, Ghadah Alghamdi, and Renate A. Schmidt. Methods and Metrics for Knowledge Base Engineering and Integration. In Proceedings of the 9th Workshop on Ontology Design and Patterns (WOP 2018) co-located with 17th International Semantic Web Conference (ISWC 2018), Monterey, USA, October 9th, 2018, volume 2195 of CEUR Workshop Proceedings, pages 72--86. CEUR-WS.org, 2018.
8. Yizheng Zhao, Ghadah Alghamdi, Renate A. Schmidt, Hao Feng, Giorgos Stoilos, Damir Juric, and Mohammad Khodadadi. Tracking Logical Difference in Large-Scale Ontologies: A Forgetting-Based Approach. In The Thirty-Third AAAI Conference on Artificial Intelligence, AAAI 2019, The Thirty- First Innovative Applications of Artificial Intelligence Conference, IAAI 2019, The Ninth AAAI Symposium on Educational Advances in Artificial Intelligence, EAAI 2019, Honolulu, Hawaii, USA, January 27 - February 1, 2019, pages 3116–3124. AAAI Press, 2019.
9. Yizheng Zhao, Ghadah Alghamdi, Renate A. Schmidt, Hao Feng, Giorgos Stoilos, Damir Juric, and Mohammad Khodadadi. Tracking Logical Difference in Industrial-Scale Ontologies. In Proceedings of the 32nd International Workshop on Description Logics, Oslo, Norway, June 18-21, 2019, volume 2373 of CEUR Workshop Proceedings. CEUR-WS.org, 2019.

Web-based Ontology Browser with Advanced Analysis Features

Abstract:   Ontologies in information science are representation of certain domain concepts. They are used in knowledge representation systems that include areas such as medicine or engineering. These ontologies can be too large in order to fit all of an application’s vocabularies. Thus, there has been an increasing interest in producing smaller modular views of ontologies that might otherwise be too big to explore and analyse. These approaches include modularisation based on partitioning methods and module extraction, which assists analysis, inspection and use for various purposes by producing modules from larger ontologies. However, the resulting modules are likely to be semantically weak, with conceptual redundancies. LETHE is an implementation of uniform interpolation, logical differences and TBox abduction, supporting expressive description logics, which are languages that express concepts in a structured way. LETHE provides restricted views of ontologies based on saturation-based reasoning that helps to preserve the logical entailments of the smaller ontologies. This has many applications, including ontology analysis, information hiding and ontology reuse. The aim in this project is to develop a web ontology analyser with advanced analysis features based on the LETHE tool. The features include visualisation capabilities beneficial for users other than computer scientists or ontology developers, helping these users to gain deep understanding of ontologies and their characteristics. Ontology developers can exploit the analyses features to focus on selected details of an ontology view, which can be useful for debugging purposes. The principal aim is to deliver an ontology analyser that supports the functionalities of LETHE, exploiting visualisation features provided by VOWL. Development of the tool involves the use of the OWL API, making it possible to achieve LETHE functionalities within the ontology analyser.

To download the thesis, click here.