The KGES Team:
- Amar Viswanathan - kannaa@rpi.edu
- Sabbir Rashid - rashis2@rpi.edu
- Ian Gross - grossi2@rpi.edu
Project Summary
Welcome! You have found the Knowledge Graph Evaluation System, or KGES. Our goal is to to detect and evaluate inconsistencies or potential incorrect labels from large scale Heterogeneous Knowledge Graphs that is constructed by outputs from IE toolkits. We use a supporting Ontology to achieve this. As a proof of concept we evaluate inconsistencies on a Biomedical Knowledge Graph. We constructed this graph using the ODIN toolkit's output.
Our ontology incorporates a number of pre-existing linked ontologies, including (but not limited to) GO, CLO, and UBERON. These ontologies are related to the Biomedical Domain and are used as supporting knowledge for PubMed documents.
We expect the KGES to fill a gap in current work in the area of evaluating large scale Knowledge Graphs. Evaluating a Knowledge Extraction system is generally done through the lens of precision, recall and the extended precision metrics. Since the output of these systems are rarely in any other form than XML documents, it is quite difficult for developers and users to figure out the semantic inadequacies of the information extraction system. Our system uses a supporting ontology to help detect mis-classified labels. This supporting ontology defines a variety inconsistencies to aid potential Knowledge Graph users.
Use and Development
This project is part of the Ontology Engineering class at RPI. The course web page can be found here. The professor overseeing the work is Prof. Deborah McGuinness.
This ontology is available for download and use under the MIT License restrictions. We highly encourage collaboration and feedback on the ontology, and we will always be looking for researchers to apply the pubmed recommendation system to their own information extraction system. Please see our getting involved page for more information. Refer to our GitHub page for the source code.