Papers with relational patterns
Continual Few-shot Relation Learning via Embedding Space Regularization and Data Augmentation (2022.acl-long)
Copied to clipboard
| Challenge: | Existing continual relation learning methods rely on labeled training data for learning new tasks, which can be expensive and time-consuming. |
| Approach: | They propose a method that embeds space regularization and data augmentation to learn relational patterns with very few labeled data while avoiding catastrophic forgetting of previous task knowledge. |
| Outcome: | The proposed method outperforms existing state-of-the-art methods in CFRL task settings. |