Descriptive knowledge graph for explaining entity relationships

We propose DEER (Descriptive Knowledge Graph for Explaining Entity Relationships) – an open and informative form of modeling entity relationships. In DEER, relationships between entities are represented by free-text relation descriptions. For instance, the relationship between entities of machine learning and algorithm can be represented as “Machine learning explores the study and construction of algorithms that can learn from and make predictions on data.” To construct DEER, we propose a self-supervised learning method to extract relation descriptions with the analysis of dependency patterns and generate relation descriptions with a transformer-based relation description synthesizing model, where no human labeling is required. Experiments demonstrate that our system can extract and generate high-quality relation descriptions for explaining entity relationships. The results suggest that we can build an open and informative knowledge graph without human annotation. We also present a novel system that automates the extraction or generation of informative and descriptive sentences from biomedical corpus and builds a descriptive knowledge graph to facilitate efficient search for relational knowledge. In contrast to previous search engines or exploration systems that retrieve unconnected passages, our system organizes descriptive sentences into a graph, enabling researchers to explore relationships between entities. Our system also includes a relation synthesis model that generates concise descriptive sentences from retrieved sentences, reducing the need for human reading effort. With our system, researchers can quickly obtain a high-level overview of directly related entities to a query entity (e.g., diseases treated by a chemical) or indirect connections between two entities (e.g., candidate drugs for treating a disease). This information can guide literature surveys and facilitate the discovery of potential research topics. Our system also speeds up the literature curation and drug repurposing process. We demonstrate the effectiveness of our system on the CORD-19 dataset, but it can be deployed on any biomedical corpus without manual adaptation.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Zhu, Kerui

Contributors dc:contributor

• Chang, Kevin Chen-Chuan

• Descriptive Knowledge Graph
• Relation Description

Statement dc:rights

• Copyright 2023 Kerui Zhu

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/120112

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F120112