question about Algorithm 1 on step 3 and 4.

declare-lab / RelationPrompt

This repository implements our ACL Findings 2022 research paper RelationPrompt: Leveraging Prompts to Generate Synthetic Data for Zero-Shot Relation Triplet Extraction. The goal of Zero-Shot Relation Triplet Extraction (ZeroRTE) is to extract relation triplets of the format (head entity, tail entity, relation), despite not having annotated data for the test relation labels.

MIT License

121 stars 16 forks source link

question about Algorithm 1 on step 3 and 4. #20

Open IoET-y opened 10 months ago

IoET-y commented 10 months ago

Hi, sorry to bother you. I was confused about the unseen label in the test set used for synthetic Data generation in step 3 and consequently leveraged the synthetic Data to train the extractor in step 4. In this case, does it mean the model can see the label in the test data? Thanks!

IoET-y commented 10 months ago

in other word, kinda test data leak