Closed jacklxc closed 3 years ago
I'll double-check on this and get back to you soon.
If I understand correctly, you're asking how we trained the rationale selector in cases where there was no evidence to be found in a candidate document?
You can see the code to train the FEVER rationale selector here. For FEVER + Scifact, we just took the FEVER-trained model, and then trained further using the scifact rationale training script.
Concerning negative samples, our training loop for FEVER looked like this:
The "negative samples" are just the sentences in the evidence document d that aren't actually rationales. Indeed, most sentences in a given evidence document aren't relevant for verifying c. This provides all the negative samples we need.
Does this make sense?
I see. Thank you for your response. My proposed solution is to perform a paragraph-level prediction, so I got confused.
Makes sense. Yes, in that case maybe generate FEVER negatives with tf-idf retrieval or something?
I'll close this.
How did you perform fever_scifact training? The paper mentioned pre-training on FEVER then training on SciFact, but the FEVER dataset provided only has positive examples included. How is the rationale selection module pre-trained on FEVER? i.e. how did you train rationale+roberta_large + fever_scifact? Did you sample / retrieve negative FEVER examples? How did you do so?
Thank you.