Academic Novelty

Scientific progress is crucial for technological advancement and hence economic growth. There is a growing concern that the pace of scientific innovation, and researchers are incentivised to focus on quantity over quality in their publications. To assess this concern, we need a way to measure the novelty of new research. This project develops a method to measure the novelty of research abstracts using natural language processing (NLP) techniques.

Author

Ricky Wang

How It Works

Using data from OpenAlex, I encode the abstracts of research papers into high-dimensional vectors (embeddings). By comparing these embeddings, I can assess how novel a new research abstract is relative to existing literature.

See Details for more information on the methodology and technical implementation.