Skip to content

DiscoverPath, a KG-based re- trieval system designed for biomedical research. This system aims to assist biomedical researchers in dynamically refining their queries and effectively retrieving articles. Accepted by CIKM'23 as Demo Paper.

Notifications You must be signed in to change notification settings

ynchuang/DiscoverPath

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

20 Commits
 
 
 
 
 
 
 
 

Repository files navigation

DiscoverPath: A Knowledge Refinement and Retrieval System for Interdisciplinarity on Biomedical Research

DiscoverPath is a Knowledge Graphs (KGs) based paper search engine for biomedical research. It addresses scenarios where users encounter challenges in pinpointing suitable search queries, particularly in interdisciplinary fields where researchers from diverse backgrounds tend to employ diverse terminologies to describe similar research. In contrast to conventional keyword-based paper search engines like Google Scholar and PubMed, DiscoverPath visualizes the papers and queries that are related to the given query with KGs, facilitating users to develop a progressive “path” toward discovering the most relevant papers. DiscoverPath is developed by Data Lab at Rice University.

📣 We have released our paper and demo video to help readers better understand the DiscoverPath system.

How DiscoverPath works?

The interdisciplinarity information can be discovered by the knowledge graphs presented in DiscoverPath system. The hidden knowledge appears in high-order path of knowledge graphs. Users are allowed to leverage any Cypher Query to find more relationships, such as shortest path of two terms, in the knowledge graphs.

Cite This Work

If you find this project useful, you can cite this work by:

Chuang Yu-Neng, et al. "DiscoverPath: A Knowledge Refinement and Retrieval System for Interdisciplinarity on Biomedical Research"

@article{chuang2023discoverpath,
  title={DiscoverPath: A Knowledge Refinement and Retrieval System for Interdisciplinarity on Biomedical Research},
  author={Chuang, Yu-Neng and Wang, Guanchu and Chang, Chia-Yuan and Lai, Kwei-Herng and Zha, Daochen and Tang, Ruixiang and Yang, Fan and Reyes, Alfredo Costilla and Zhou, Kaixiong and Jiang, Xiaoqian and others},
  journal={arXiv preprint arXiv:2309.01808},
  year={2023}
}

Utilization and Demonstration of DiscoverPath

Interface

The overall GUI interface of the DiscoverPath system, which consists of a frontend interaction, a backend platform, and a graph database.

Utilization

A using pipeline overview of DiscoverPath based on a client-server architecture. DiscoverPath follows a pipeline to gradually refine the knowledge graphs that meet the requirements of users. Users first give their initial queries to get the initial knowledge graphs, and then utilize the recommended queries to refine the initial one before visualizing the Interdisciplinary knowledge.

Demonstration

DiscoverPath system shows the results of interdisciplinary knowledge exploration. We here select serveral results in Alzheimer’s disease, which is shown as follows:

Developed Environment

Install packages

-  OpenJDK 1.8
-  Python 3.7
-  neo4j 3.5

Start neo4j server

<NEO4J_HOME>/bin/neo4j console
<NEO4J_HOME>/bin/neo4j start
./bin/neo4j-admin set-initial-password <Your_Password>

Initiaize Neo4j Data

Given a knowledge graph triplets input:

Paper_A Relation_1 Term_A
Paper_A Relation_1 Term_B
Paper_A Relation_1 Term_D
Paper_B Relation_1 Term_B
Paper_B Relation_1 Term_C
Paper_A Related_paper Paper_B 

Then initiaize

python clean.py
python neo2example.py

About

DiscoverPath, a KG-based re- trieval system designed for biomedical research. This system aims to assist biomedical researchers in dynamically refining their queries and effectively retrieving articles. Accepted by CIKM'23 as Demo Paper.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published