Hello everyone!

I'm working on some problems in Natural Language Processing. Now I've been working on Stance Detection. I have an idea but I'm not sure either it's useful or not.

I want to use RTE data (labels: "entailment", "contradiction", "neutral") for training Stance model. I think "entailment" and "contradiction" labels in RTE and "support" and "deny" labels in Stance are equivalent respectively.

I wonder if you think so. if not please give me some reasons for it.

Thanks alot in advance.

More Masoud Sadeghi's questions See All
Similar questions and discussions