July 2017
Beginner to intermediate
486 pages
13h 49m
English
NLG is an area where we are trying to teach machines how to generate NL in a sensible manner. This, in itself, is a challenging AI task. Deep learning has really helped us perform this kind of challenging task. Let me give you an example. If you are using Google's new inbox, then you may notice that when you reply to any mail, you will get three most relevant replies in the form of sentences for the given mail. Google used millions of e-mails and made an NLG model that was trained using deep learning to generate or predict the most relevant reply for any given mail. You can refer to Figure 9.8:

Read now
Unlock full access