Showing 1–1 of 1 results for author: Mane, S S
-
Biomedical Multi-hop Question Answering Using Knowledge Graph Embeddings and Language Models
Authors:
Dattaraj J. Rao,
Shraddha S. Mane,
Mukta A. Paliwal
Abstract:
Biomedical knowledge graphs (KG) are heterogenous networks consisting of biological entities as nodes and relations between them as edges. These entities and relations are extracted from millions of research papers and unified in a single resource. The goal of biomedical multi-hop question-answering over knowledge graph (KGQA) is to help biologist and scientist to get valuable insights by asking q…
▽ More
Biomedical knowledge graphs (KG) are heterogenous networks consisting of biological entities as nodes and relations between them as edges. These entities and relations are extracted from millions of research papers and unified in a single resource. The goal of biomedical multi-hop question-answering over knowledge graph (KGQA) is to help biologist and scientist to get valuable insights by asking questions in natural language. Relevant answers can be found by first understanding the question and then querying the KG for right set of nodes and relationships to arrive at an answer. To model the question, language models such as RoBERTa and BioBERT are used to understand context from natural language question. One of the challenges in KGQA is missing links in the KG. Knowledge graph embeddings (KGE) help to overcome this problem by encoding nodes and edges in a dense and more efficient way. In this paper, we use a publicly available KG called Hetionet which is an integrative network of biomedical knowledge assembled from 29 different databases of genes, compounds, diseases, and more. We have enriched this KG dataset by creating a multi-hop biomedical question-answering dataset in natural language for testing the biomedical multi-hop question-answering system and this dataset will be made available to the research community. The major contribution of this research is an integrated system that combines language models with KG embeddings to give highly relevant answers to free-form questions asked by biologists in an intuitive interface. Biomedical multi-hop question-answering system is tested on this data and results are highly encouraging.
△ Less
Submitted 10 November, 2022;
originally announced November 2022.
