关键词: 深度学习, 预训练模型, 自然语言处理, 机器阅读理解

Abstract:

In recent years, deep learning technology has been advancing. With the application and development of pre-training model in natural language processing, machine reading comprehension is no longer simply based on the combination of network structure and word embedding. The development of pre-training language model has led to advances in machine reading comprehension that has surpassed human performance in some datasets. This paper briefly introduces the concepts of machine reading comprehension and pre-training language model, summarizes the current research progress of machine reading comprehension based on the pre-training model, analyzes the performance of the current pre-training model on the relevant data set, summarizes the existing problems and looks forward to the future.

Key words: deep learning, pre-training model, natural language processing, machine reading comprehension