CODEWITHSUNDEEP
pythonnlpartificial-intelligencenltk
Macaroonman
Macaroonman

Reputation: 69

Python rephrasing/paraphrasing options

While I’m aware there have been some questions on this, I want to know all my options when it comes to rephrasing statements in python. I tried making synonym fetchers but I got phrases like:

The spry brown slyboots jumps over the work-shy hound

And other things no one would really say. Any specific nltk/api/lib recommendations for the best way to approach this? Thanks

Upvotes: 0

Views: 3245

Answers (2)

dheeraj
dheeraj

Reputation: 368

You can use the below libraries or methods to perform data augmentation or paraphrasing.

  1. NLPAug - (https://github.com/makcedward/nlpaug)
  2. Back Translation Method - (https://github.com/google-research/uda)
  3. T5 model and Google UDA implementation (https://github.com/dheerajiiitv/T5-paraphrase-generation)

Upvotes: 1

Esraa Ibrahim
Esraa Ibrahim

Reputation: 26

The result is meaningless because you don't put the word ambiguity in your mind , each word may have a lot of meanings like (bank ) can be a financial institution or the river side

so you have to solve that - you can solve that by a lot of algorithms like Lesk Algorithms that will solve your problem

also you can make paraphrasing using the translation to a pivot language then re-translate it to English again this will make some differences too

Upvotes: 0

Related Questions