Research Paper Recommendation System using Transformer Model

Searching for relevant material on the internet has become a big difficulty with the growing number of internet users and the large volume of data available on the internet. Traditional search engines return a lot of irrelevant results in response to a user’s search query, wasting a lot of time and effort because the context and semantics of the user’s request are not examined. This emphasizes the requirement for an embedded approach in the search engine, allowing semantic web search. Keyword based search engines fetch online pages by comparing the tokens in the user query to the tokens in the web documents. This strategy has several flaws. To overcome these obstacles, the suggested system employs semantic search using a Sentence transformer model. The system that has been proposed can speed up data retrieval and improve data searching to provide relevant outcomes to the user. The proposed model uses the arXiv dataset which contains an archive of research papers and using BERT and FAISS, it gives relevant research paper titles as the output to the query. For semantic retrieval of data, the model behaves as a vector-based search engine. This vector-based model will save time and give efficient results to the user’s query.