A Conditional Cascade Model for Relational Triple Extraction
This work addresses a specific bottleneck in information extraction for researchers and practitioners, offering incremental improvements over existing methods.
The paper tackles the class imbalance problem in relational triple extraction by proposing a novel tagging-based model that reduces sample numbers and uses a confidence threshold loss, achieving state-of-the-art results on NYT and WebNLG datasets.
Tagging based methods are one of the mainstream methods in relational triple extraction. However, most of them suffer from the class imbalance issue greatly. Here we propose a novel tagging based model that addresses this issue from following two aspects. First, at the model level, we propose a three-step extraction framework that can reduce the total number of samples greatly, which implicitly decreases the severity of the mentioned issue. Second, at the intra-model level, we propose a confidence threshold based cross entropy loss that can directly neglect some samples in the major classes. We evaluate the proposed model on NYT and WebNLG. Extensive experiments show that it can address the mentioned issue effectively and achieves state-of-the-art results on both datasets. The source code of our model is available at: https://github.com/neukg/ConCasRTE.