Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Few-shot Open Relation Extraction with Gaussian Prototype and Adaptive Margin

Tianlin Guo, Lingling Zhang, Jiaxin Wang, Yuokuo Lei, Yifei Li, Haofen Wang, Jun Liu

TL;DR

A novel framework based on Gaussian prototype and adaptive margin named GPAM for FsRE with NOTA, which includes three modules, semi-factual representation, GMM-prototype metric learning and decision boundary learning, which shows that GPAM surpasses previous prototype methods and achieves state-of-the-art performance.

Abstract

Few-shot relation extraction with none-of-the-above (FsRE with NOTA) aims at predicting labels in few-shot scenarios with unknown classes. FsRE with NOTA is more challenging than the conventional few-shot relation extraction task, since the boundaries of unknown classes are complex and difficult to learn. Meta-learning based methods, especially prototype-based methods, are the mainstream solutions to this task. They obtain the classification boundary by learning the sample distribution of each class. However, their performance is limited because few-shot overfitting and NOTA boundary confusion lead to misclassification between known and unknown classes. To this end, we propose a novel framework based on Gaussian prototype and adaptive margin named GPAM for FsRE with NOTA, which includes three modules, semi-factual representation, GMM-prototype metric learning and decision boundary learning. The first two modules obtain better representations to solve the few-shot problem through debiased information enhancement and Gaussian space distance measurement. The third module learns more accurate classification boundaries and prototypes through adaptive margin and negative sampling. In the training procedure of GPAM, we use contrastive learning loss to comprehensively consider the effects of range and margin on the classification of known and unknown classes to ensure the model's stability and robustness. Sufficient experiments and ablations on the FewRel dataset show that GPAM surpasses previous prototype methods and achieves state-of-the-art performance.

Few-shot Open Relation Extraction with Gaussian Prototype and Adaptive Margin

TL;DR

A novel framework based on Gaussian prototype and adaptive margin named GPAM for FsRE with NOTA, which includes three modules, semi-factual representation, GMM-prototype metric learning and decision boundary learning, which shows that GPAM surpasses previous prototype methods and achieves state-of-the-art performance.

Abstract

Few-shot relation extraction with none-of-the-above (FsRE with NOTA) aims at predicting labels in few-shot scenarios with unknown classes. FsRE with NOTA is more challenging than the conventional few-shot relation extraction task, since the boundaries of unknown classes are complex and difficult to learn. Meta-learning based methods, especially prototype-based methods, are the mainstream solutions to this task. They obtain the classification boundary by learning the sample distribution of each class. However, their performance is limited because few-shot overfitting and NOTA boundary confusion lead to misclassification between known and unknown classes. To this end, we propose a novel framework based on Gaussian prototype and adaptive margin named GPAM for FsRE with NOTA, which includes three modules, semi-factual representation, GMM-prototype metric learning and decision boundary learning. The first two modules obtain better representations to solve the few-shot problem through debiased information enhancement and Gaussian space distance measurement. The third module learns more accurate classification boundaries and prototypes through adaptive margin and negative sampling. In the training procedure of GPAM, we use contrastive learning loss to comprehensively consider the effects of range and margin on the classification of known and unknown classes to ensure the model's stability and robustness. Sufficient experiments and ablations on the FewRel dataset show that GPAM surpasses previous prototype methods and achieves state-of-the-art performance.

Paper Structure

This paper contains 18 sections, 10 equations, 4 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Relation Extraction
  4. Few-shot Relation Extraction
  5. Few-shot Relation Extraction with NOTA
  6. Problem Formalization
  7. Methodology
  8. Semi-Factual Representation
  9. GMM-Prototype Metric Learning
  10. Decision Boundary Learning
  11. Experiment
  12. Settings
  13. Performance Comparison
  14. Ablation Study on Semi-Factual Representation
  15. Ablation Study on GMM-Prototype Metric Learning
  16. ...and 3 more sections

Figures (4)

  • Figure 1: Differences bewteen two tasks RE and FsRE with NOTA. (a) There is a large-scale annotated dataset and all classification results in the query set are within the given set of known classes. (b) There is a few-shot dataset and the query set contains some relations that are not among the known classes and should be classified as NOTA.
  • Figure 2: Overview of Our Model GPAM.
  • Figure 3: Results with increasing NOTA rate from 0 to 50%.
  • Figure 4: The performance variation curves for different negative sampling rates under four task settings.