High-Level Project Summary
There are a lot of data without keywords made by NASA. Our job is simply using NLP and machine learning to make keywords to the data to make it easier to use and ease the researcher's job.
Link to Final Project
Link to Project "Demo"
Detailed Project Description
To improve the accessibility of such documents, an application is required to: -
1) Read : the application will need to have the capability of interpreting the words and sentences in the document. -
2) Analyze : afterwards performing certain analysis steps to further understand each document and be able to breaking it's topics down for the next step. -
3) Produce : proper topic keywords and a summary will then be outputted for each document using an nlp Machine Learning algorithm -
4) Store : finally the output will be organized into a database to ease the process of querying and integration with the current NTRS server.
Space Agency Data
https://ntrs.nasa.gov/api
NTRS documents are usually tagged with Subject Category a summary and a list of topic keywords to ease the search for scientists using the platform, however a portion of those documents have been added by means of scanning and OCR. The problem is that those documents only contain a Subject Category but lack topic keywords and a summary.
Hackathon Journey
It's an excellent experience for students and non-students to learn and live the experience of working under pressure and in a limited time. Although there were difficulties the challenge that's what made it a good experience. We definitely learned how to work together as a team and manage time among the team members, and of course, we gained experience in our profession. We as a team love AI as a natural passion and have a curiosity for astronomy so merging those two up was an excellent choice for us. We created a web/app to show the topic points and a quick summary for every document by typing the document id in an input. Our way of resolving the setbacks and challenges was by managing our time correctly among the members. we would like to thank all the sponsors for supporting this amazing event and helping us have a better experience and the AUC of course for hosting the event and providing us with our needs, and we would like to especially thank all the volunteers for their hard work the organizing was managed perfectly.
References
https://ntrs.nasa.gov/
https://www.kaggle.com/code/donkeys/summarizing-topic-models-with-transformers/notebook
https://www.slidescarnival.com/joan-free-presentation-template/11687
https://towardsdataschttps://towardsdatascience.com/topic-modeling-with-bert-779f7db187e6ience.com/topic-modeling-with-bert-779f7db187e6
https://radimrehurek.com/gensim/parsing/preprocessing.html
Tags
#AI