High-Level Project Summary
We are the team, "The one after the X" and we have a complete searching system implemented using NLTK, Spacy, Next.js, and other modern technologies, as well as text summarisation and tag generation, to help users find the exact research papers they are looking for simply by entering a relevant word, phrase, or tag.Furthermore, we have built a powerful filtering engine that takes into consideration the images in the PDFs, allowing customers to narrow their search to just what they want. Along with that, we have built a system that allows users to collaborate, post their own work, and do peer evaluations, supporting research in the subject while respecting the users` privacy.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Problem Statement -
The challenge is to develop an AI application to improve the accessibility and discoverability of records in the NTRS. For example, You could use AI to read documents, generate text analytic data, and produce a list of topic keywords to help researchers find the documents they need. We need to think about what types of information future researchers will need to locate desired documents and what would be the best data to aid them in their search. A simple and user friendly web interface for users to seamlessly search the required research paper while leveraging the power of AI. A robust filter system allowing user to select from the plethora of documents, which ones they wish to take look.
Our Solution -
- A simple and user friendly web interface for users to seamlessly search the required research paper while leveraging the power of AI.
- Complete automated solution to promote collaborative work and peer reviews on research papers, allowing one to upload their own documents
- A robust filter system allowing the user to select from the plethora of documents, which ones they wish to take a look at.
USPs -
- Inclusion of images as a part of our search system to provide more elaborate search results.
- Increasing accessibility to all through implementation of text-to-speech and speech-to-text features
- Taking it one step forward, and helping researchers find more relevant documents through a bold recommendation system
Features:
- Summarization and Tag-generation using NLP
- Paper Publication
- Auto-Recommendation system
- Robust filters using keywords
- Caption generation and text extraction from images
Space Agency Data
https://ntrs.nasa.gov/api/openapi/citations/search
https://paperswithcode.com/
Current NTRS has millions of documents and it is very difficult for a researcher to search a particular document. And this inspired us to optimize the search algorithm and change the approach to get the required document.
We used the documents provided by the api to fetch the documents and research papers. With these documents we have trained our NLP model to generate the summary and to extract keywords.
Hackathon Journey
It was a great experience working together as team for 48 hours and brainstorming our minds with innovative ideas to solve the real life problem. We have come up with a solution to increase the efficiency of the current database and optimized the search algorithm using AI . We got great feedbacks from our mentors who helped us to refine our solution .
Although there were times when the mentoring session were preponed and it was quite challenging for us to manage things in a very short time. But even then we presented ourselves with the best efforts and felt really good about it.
References
https://pypi.org/
https://fastapi.tiangolo.com/
https://ntrs.nasa.gov/api/openapi/citations/search/
https://pypi.org/project/pytesseract/
https://spacy.io/
https://spacy.io/universe/project/spacy-pytextrank
https://pymupdf.readthedocs.io/en/latest/module.html
https://nextjs.org/
https://docs.aws.amazon.com/index.html?nc2=h_ql_doc_do
https://nodejs.org/en/docs/
https://getbootstrap.com/docs/5.2/getting-started/introduction/
https://www.nltk.org/
Tags
#NLP #AI #WEB #Scalable #NASA #NTRS #Accessibility #text-analysis #keyword-extraction