Skip to content

README.md

Latest commit

 

History

History
35 lines (26 loc) · 2.01 KB

README.md

File metadata and controls

35 lines (26 loc) · 2.01 KB

In this Repo two problem statements are solved -

  1. Fake News Detection through a classifier using BERT. Colab Link(In case if it doesn load via Github) - https://colab.research.google.com/github/singularity014/BERT_FakeNews_Detection_Challenge/blob/master/Detect_fake_news.ipynb

  2. A Contextual Text Similarity approach for putting similar narrative news articles in similar cluster group. Colab Link(In case if it doesn load via GitHub) - https://colab.research.google.com/github/singularity014/BERT_FakeNews_Detection_Challenge/blob/master/SImilarity_News.ipynb

Environment - Step by step procedure is present in Google Colab's of Problem 1 and Problem 2 in "Detect_fake_news.ipynb" & "SImilarity_News.ipynb" respectively.

Following Model Architectures have been used in achieving the goals -

  • BERT based Classifier Model,
  • BERT based text summarizer Model,
  • USE based feature extracter.

Hence, for leveraging the computational GPU benefits of cloud frameworks, Google Colab has been used.

The link to open the Google Colab is already present in the github hyperlink indicated as 'Open in Colab' (Blue indication, at the very top of both the codes.. Please feel free to open and running the code.

The code is well commented and organised in Step by Step approach, to make it easier to follow. For any doubts, please feel free to reach out at - techprafull123@gmail.com