TL;DR运用域和任务自适应预训练模型 ALBERT 在顾客联系文本上,通过利用未标注数据、进行半监督学习,将多分类问题解决为多任务学习,从而提高了客服意图分类模型的性能。
Abstract
In the area of customer support, understanding customers' intents is a
crucial step. Machine learning plays a vital role in this type of intent
classification. In reality, it is typical to collect confirmation from customer
support representatives (CSRs) regarding the intent prediction