Literature review can be a long process. In order to make it easier, we wanted to build a service that finds the seminal papers in any field of study.
What it does
What results should look like
It uses an NLP algorithm called Latent Dirichlet Allocation to generate a list of topics given a corpus of the paper abstract and title. It finds a recursive implementation to find papers with the highest numbers of citations, which indicates a higher quality.
How We built it
Using Algorithmia to creates relevant topics to the search paper and queried Google Scholar, which returns a list of relevant papers. We plan to recursively generate topics in order to find more frequently cited papers. After recursing enough times, we will take the most highly-cited papers and display them to the user.
Challenges We ran into
None of us knew anything about server-side programming. We spent a lot of time configuring our server and learning Flask.
Accomplishments that We're proud of
We put together a working product with very little prior experience in web programming.
What We learned
We learned a lot of new technologies, APIs, and improvising on the go.
What's next for intuition
Also, fine-tune the algorithm for better accuracy and user experience.