ak0123 / scibertweb2

The goal of this project is to help people on the front-line fighting COVID-19 find the most relevant research papers to better understand key pain-points and apply these findings into their plans and operations in addressing the pandemic.

scibertweb2

Here are the lessons learned from this project:

https://cosmosinyou.medium.com/lessons-learned-from-deploying-a-pretrained-bert-model-as-a-web-site-88a9d967c751

#.. as well as Tutorial: Deploying a machine learning model to the web https://blog.cambridgespark.com/deploying-a-machine-learning-model-to-the-web-725688b851c7

Project Objective:

The goal of this project is to help people on the front-line fighting COVID-19 find the most relevant research papers to better understand key pain-points and apply these findings into their plans and operations in addressing the pandemic.

The main issue is that more papers are published than there are qualified people to refind and leverage the knowledge.

This project can help to simplify the search process and bring the most relevant search results to eye-level for doctors, virologists, and other contributors mitigating the pandemic.

This project was forked from Issac MG's SciBERT Embeddings Analysis https://www.kaggle.com/isaacmg/scibert-embeddings

important you follow setup instructions in link above before proceeding

as of May 2020, this code yields slug size of 1.2 G when web deploying to heroku via 'git push heroku master'

requires further research to get this to compliant slug size