Common machine reading comprehension models
Abstract
Many outstanding machine reading comprehension (MRC) models have emerged in recent years. These models contain various innovations in network architecture, modular design, and training methods. They achieve remarkable results in many MRC datasets and competitions. Thus it is important to analyze the design philosophy and implementation techniques behind these models to create better architectures. This chapter selects four outstanding MRC models and provides a detailed analysis of the features, architecture, and innovations of each model.
Keywords
Encoding layer; interaction layer; output layer; BiDAF; R-NET; FusionNet; ET-RR
5.1 Bidirectional attention flow model
The Bidirectional Attention ...
Get Machine Reading Comprehension now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
