QA-with-Tensorflow
QA-with-Tensorflow copied to clipboard
Language Independency
I want to design a QA system for a closed domain. Problem is that I have to work on a transliterated data. Is it possible to design a QA system with transliterated data following this project?
The one major difference between QA in the English language and QA in other languages (or using transliterated words, etc) is that you're not likely to have a full and proper word embedding. You should be able to use something in the same style as this project if you obtain an embedding (by finding one already created for the vocabulary you're using, or by creating one yourself using one of the many word vectorization tools available).
Hi Steven,
Can we add more layers to the model? And what does exactly mean by Close Domain?