关键词: 抽取式机器阅读理解, 自然语言处理, 深度学习, 迁移学习, 注意力机制

Abstract:

Machine reading comprehension requires machines to understand natural language texts and answer related questions, which is the core technology in the field of natural language processing and one of the most challenging tasks in the field of natural language processing. Extractive machine reading comprehension is an important branch of machine reading comprehension task. Because it is more suitable for the actual situation and can reflect the understanding ability of the machine, it has become a research hotspot in the current academic and industrial circles. This paper makes a comprehensive review of extractive machine reading comprehension from four aspects, first of all, the paper introduces the task of machine reading comprehension and its development process. Secondly, it describes the task of extractive machine reading comprehension and its difficulties at present. Then, the main data sets and methods of the extractive machine reading comprehension task are summarized. Finally, the future development direction of extractive machine reading comprehension is discussed.

Key words: extractive machine reading comprehension, natural language processing, deep learning, transfer learning, attention mechanism