Few-Shot Event Detection with Prototypical Amortized Conditional Random Field

TL;DR

This paper introduces a novel few-shot event detection model using a prototypical amortized CRF that models label dependencies and improves performance on benchmark datasets by addressing trigger discrepancy and data scarcity.

Contribution

The paper proposes a unified few-shot event detection model with PA-CRF that models label dependencies via prototypes and Gaussian transition scores, outperforming existing methods.

Findings

PA-CRF achieves state-of-the-art results on FewEvent dataset.

Unified model outperforms identify-then-classify approaches.

Gaussian transition scores improve robustness with limited data.

Abstract

Event detection tends to struggle when it needs to recognize novel event types with a few samples. The previous work attempts to solve this problem in the identify-then-classify manner but ignores the trigger discrepancy between event types, thus suffering from the error propagation. In this paper, we present a novel unified model which converts the task to a few-shot tagging problem with a double-part tagging scheme. To this end, we first propose the Prototypical Amortized Conditional Random Field (PA-CRF) to model the label dependency in the few-shot scenario, which approximates the transition scores between labels based on the label prototypes. Then Gaussian distribution is introduced for modeling of the transition scores to alleviate the uncertain estimation resulting from insufficient data. Experimental results show that the unified models work better than existing identify-then-classify models and our PA-CRF further achieves the best results on the benchmark dataset FewEvent. Our code and data are available at http://github.com/congxin95/PA-CRF.