How to generate a dataset for question answering from books (like Facebook's bAbI)?

Question

I wanted to train a chatbot for answering questions from books. I am trying to use Dynamic Memory Networks to do so.

How can I generate a data set, as Facebook did in the case of bAbI tasks, so that it can tackle a variety of questions on the data set?

score 2 · Answer 1 · edited Oct 03 '18 at 19:23

2

If you are talking about "generating" in the sense of generative models , it is pretty tough. since we are still far beyond understanding the actual structure of question-answering.

And even state of the art methods for question answering are also not able to score well on datasets like babi , mostly 16 out of 20 tasks can be solved.

edited Oct 03 '18 at 19:23

quintumnia

1,183
1
10
34

answered Jun 05 '18 at 15:42

thecomplexitytheorist

766
3
19

Thanks a lot for answering @thecomplexitytheorist. I wanted to create BaBI like tasks from a custom dataset - like a book. So it was not exactly like fully generating from scratch. – Shubham Bhardwaj Nov 02 '18 at 08:15

score 1 · Accepted Answer · answered Nov 02 '18 at 08:25

This repository maintained by Facebook AI Research talks about how they went about generating QA from stories.

In essence, they try to simulate how a reader reads a story. They also keep track about the knowledge the reader is assimilating when reading. Then they frame a question based on the knowledge assimilated, in order to asses if the reader can perform logical inference given what knowledge is present with him.

How to generate a dataset for question answering from books (like Facebook's bAbI)?

2 Answers2