A common Natural Language Processing (NLP) task is to determine semantic similarity between two sentences. Has the question of agreement/disagreement between two sentences been covered in NLP or other literature? I tried searching on Google Scholar but didn't get any relevant results.
-
$\begingroup$ What do you call semantic similarity ? Are "The gender of Steve is male" and "Steve is known not to be a girl" similar ? $\endgroup$– user16034Mar 18 at 9:47
-
$\begingroup$ How would you define it? $\endgroup$– Hamman SamuelMar 18 at 10:29
-
$\begingroup$ I am asking you. This is your post. $\endgroup$– user16034Mar 18 at 11:13
-
$\begingroup$ Please refrain from any further comments, I don't appreciate your tone, and you are not being helpful at all $\endgroup$– Hamman SamuelMar 18 at 15:01
-
$\begingroup$ You still did not answer my question. It is a pity that you don't want to clarify your post. $\endgroup$– user16034Mar 18 at 15:51
1 Answer
I would propose to do some researching in the field of Stance Classification. Given a target claim or argument we can classify whether a number of sentences are in favor, against or neither of that claim. So an idea is to extract the topic of those sentences classify whether they agree or disagree with it and depending that classification you can determine if those sentence agree or disagree.
Here are some papers you can look at:
https://paperswithcode.com/task/stance-classification
https://arxiv.org/abs/1907.00181
-
1$\begingroup$ Thanks, after your answer, I went through a few papers on stance detection and feel this one is captures the topic fairly well: arxiv.org/abs/2104.07467 What do you think? $\endgroup$ Mar 18 at 10:37
-
$\begingroup$ It is good to start. I have edited my answer including some papers I found interesting. Also on the first link you can find some papers as well as the implemented code. I hope those will help. $\endgroup$– AnoroahMar 18 at 19:45