Python Natural Language Processing

Deep learning in NLP

The early era of NLP is based on the rule-based system, and for many applications, an early prototype is based on the rule-based system because we did not have huge amounts of data. Now, we are applying ML techniques to process natural language, using statistical and probability-based approaches where we are representing words in form of one-hot encoded format or co-occurrence matrix.

In this approach, we are getting mostly syntactic representations instead of semantic representations. When we are trying out lexical-based approaches such as bag of words, ngrams, and so on, we cannot differentiate certain context.

We hope that all these issues will be solved by DNN and DL because nowadays, we have huge amounts of data ...