基于生成元学习的零样本关系三元组抽取（英文版）

英文标题：Generative Meta-Learning for Zero-Shot Relation Triplet Extraction中文摘要：本研究提出了一种新颖的生成元学习框架，旨在提高生成模型的泛化能力，该框架在零样本关系三元组提取任务中达到了新的最先进性能。英文摘要：The zero-shot relation triplet extraction (ZeroRTE) task aims to extractrelation triplets from a piece of text with unseen relation types. The seminalwork adopts the pre-trained generative model to generate synthetic samples fornew relations. However, current generative models lack the optimization processof model generalization on different tasks during t