Text Similarity Search Application using Pretrained Language Models and Elasticsearch
In this notebook, I show how to use pretrained language models and Elasticsearch to build a simple text similarity search application on the COVID-19 Open Research Dataset Challenge (CORD-19). These are a collection of scholarly articles including studies around COVID-19. The main goal is to use embeddings for queries and documents and let a search engine return the most similar stored documents given the query.
Preview of the notebook below: