deepMINE - Natural Language Processing based Automatic Literature Mining and Research Summarization for Early-Stage Comprehension in Pandemic Situations specifically for COVID-19

Bhrugesh Joshi; Vishvajit Bakarola; Parth Shah; Ramar Krishnamurthy

Our primary contribution majorly focuses on the design and development of the system that can able to automatically mine the literature, provided the keyword(s) as an input. The system is able to generate the brief summary of the research article, that user interested with. Our ultimate objective is to provide a quick and easy literature mining service so that the researchers working for fighting current pandemic situation of COVID-19 can get early-stage comprehension. Currently we have used the COVID-19 Open Research Dataset (CORD-19) , publicly made available by Allen Institute of AI on 20th March 2020. Our system deepMINE is providing mining from 29,315 research articles with keywords by scanning nearly 1,46,115,136 English words available in literature dataset in not greater than 1.5 seconds. From experiments performed on the live deepMINE system we can observer that the system is helpful for the primary mining of the huge amount literature available in various openly available datasets. However, in many of the research articles of the dataset the affiliation and/or location has not been provided. We are in the process of cleaning and improving the dataset with automatic and manual processes.