Awards & Nominations
TheAISmart has received the following awards and nominations. Way to go!
Global Nominee
YOUR NEW SUMMRAZING INTERN
High-Level Project Summary
[note my team HAS 2 PEOPLE But one of my members have a problem in using this website]Using Transformer pipeline our team made a solution to accurately summarize and extract keywords out of it Using the abstract introduction and summarization of the paper we ran it through the pipelineSource code:https://colab.research.google.com/drive/1sM6LmZPdyKlwDHhgYxe2v0xkiSZlFmOM?usp=sharingDeployment:https://colab.research.google.com/drive/1Wg1Ovl-MzrLLfYukgq5ajVuJ3kjZ3tnm#scrollTo=B0Isy5v9xay7reports of summarization will be on github \\ and source code for streamlit:https://github.com/PluzNtp/NASA_WEB
Link to Final Project
Link to Project "Demo"
Detailed Project Description
Using Transformer pipeline our team made a solution to accurately summarize and extract keywords out of CMDS
by classifying the pdfs to two types:
1)PDF with multiple research papers
2)PDF with One research papers
we then extracted the text out of the pdf
then we cleaned it form any special character : (@\[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)|^rt|http.+?"
then we corrected it using a pretrained model we found
then
By extracting the abstract introduction and summarization of the paper we ran it through the pipeline and extracted the summarization
We also did keyword extraction using a pretrained model we found on hugging face.
for that we use the entire text of the research paper
Space Agency Data
We used CDMS from :https://ntrs.nasa.gov/search?center=CDMS
Hackathon Journey
I learned alot about nlp in general the reason i choosed this challange is because i was interested in machine learning in general and wanted to tackle something resembling it
References
https://docs.streamlit.io/
https://huggingface.co/docs/transformers/main_classes/pipelines
Tags
#software #nlp