Few-Shot Relation Classification Dataset
The FewRel (Few-Shot Relation Classification Dataset) contains 100 relations and 70,000 instances from Wikipedia. The dataset is divided into three subsets: training set (64 relations), validation set (16 relations) and test set (20 relations).
Source: Neural Snowball for Few-Shot Relation Learning
Image Source: https://www.aclweb.org/anthology/D18-1514.pdf
Variants: FewRel
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Relation Classification | DeepStruct multi-task | DeepStruct: Pretraining of Language Models … | 2022-05-21 |
| Relation Classification | DeepStruct multi-task w/ finetune | DeepStruct: Pretraining of Language Models … | 2022-05-21 |
| Relation Classification | Deepstruct zero-shot | DeepStruct: Pretraining of Language Models … | 2022-05-21 |
| Relation Classification | DeepEx (zero-shot top-10) | Zero-Shot Information Extraction as a … | 2021-09-23 |
| Relation Classification | DeepEx (zero-shot top-1) | Zero-Shot Information Extraction as a … | 2021-09-23 |
| Relation Extraction | RoCORE | A Relation-Oriented Clustering Method for … | 2021-09-15 |
| Relation Extraction | ERNIE | ERNIE: Enhanced Language Representation with … | 2019-05-17 |
Recent papers with results on this dataset: