This notebook provides textual similarity between given two paragraphs. Google universal sentence encoder is used to create embeddings for these words.
This can be used to compare any two types of sentences and find similarity between them, simply replace dataset file with you file and made changes in noebook accordingly.
This notebook was created in Google colab, so couple fo starting lines can be ignored. Results are pretty convincing, thanks to Google's pre-trained model.
Because these techniques does not find similarity on the basis of text, but find similarity on the basis of semantics. For more info on this: Search Word2Vec and Google universal Sentence encoder.
