High-Level Project Summary
Our search engine uses advanced Artificial Intelligence systems to rank search results in a relevancy level manner. Scout out tries to continue the sentences you're trying to write and helps find familiar queries (search texts) to the one you wrote to make sure we understand your search intent properly. We have 3 methods to represent document data once you see document links, either a summary of the text, sub-topics, or most frequent words, in addition to a filtering system for specific results.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
The project details were already mentioned before, however, for further details, you could check our video to see how we sketched the user interface. The project is used to improve search results for a certain NASA database that includes around 380k scientific research papers and articles, what we aimed to do is to rank these topics based on the search requirements, we also created further features to ease the operation of filtering documents and knowing general details about them. The project uses different frameworks to work properly and with high efficiency, a combination of ElasticSearch Engine in addition to Google's S-BERT, to get ranking based on sentence meanings and embeddings based on the requested query by the user. We take the data from the NASA Server to use it as training data to feed to the AI system. and finally we use Flask as our backend, the project is purely developed using Python programming language.
Space Agency Data
All data we need for the application we got from NASA NTRS OpenApI
Hackathon Journey
It was an incredible experience, we socialized, we had fun with the people, and we were supposed to sleep there, however, we didn't because we worked on the project throughout the night, we learned much as we chose a challenge then we changed it due to the complexity of data (as it requires a specialist in physics). Therefore, we had a backup project to work on, which was this current one, we're all software engineers and AI engineers, therefore, it was easier for us to implement NLP on this project. our challenges were the change of project, and the fact that we weren't able to sleep. we managed to continue through the day though. We'd like to thank everyone responsible for this amazing experience, especially Dr. Mona Dmaidi and N-Gate.
References
Tags
#software ,#AI , #NLP