BERT and Broad Learning for Textual Emotion Classification

TEC is an important task in NLP. Most existing methods for TEC are based on DL. However, there are some weaknesses in DL, such as long training time and difficult convergence. Due to a flat network model of BL, which has the advantages of simple structure, incremental modeling and short training time, in this chapter, we explore to combine BERT and BL, named BBL, for TEC task. In addition, we conduct the relevant experiments on two public data sets to verify the performance of the proposed BBL.