Scientific Paper Recommender

This project is a scientific paper recommender that leverages embeddings stored in a DynamoDB table. The embeddings are created using the OpenAI embeddings API. When a user enters a query in the search box, the application compares the query using cosine similarity to the 5 most similar embeddings stored in the database. The application assumes the user already has their own embeddings that they want to compare their query to.

In this example, over 60,000 arXiv papers are stored in DynamoDB. If you need access to this test data, please feel free to reach out.