BriefGPT.xyz
May, 2021
伪Siamese网络用于少样本意图生成
Pseudo Siamese Network for Few-shot Intent Generation
HTML
PDF
Congying Xia, Caiming Xiong, Philip Yu
TL;DR
本文提出了一种基于Pseudo Siamese Network的少样本意图检测方法，实现了在较少标注数据的情况下生成有效的标签，将句子的动作和对象分开，使用transformer-based variational autoencoder来实现解码。在两个真实数据集上的实验证明，该方法在少样本意图检测任务中达到了最先进水平。
Abstract
few-shot intent detection
is a challenging task due to the scare annotation problem. In this paper, we propose a
pseudo siamese network
(PSN) to generate labeled data for few-shot intents and alleviate this probl
→