Investigating Capsule Networks with Dynamic Routing for Text Classification

EMNLP 2018 Wei Zhao • Jianbo Ye • Min Yang • Zeyang Lei • Suofei Zhang • Zhou Zhao

In this study, we explore capsule networks with dynamic routing for text classification. We propose three strategies to stabilize the dynamic routing process to alleviate the disturbance of some noise capsules which may contain "background" information or have not been successfully trained. A series of experiments are conducted with capsule networks on six text classification benchmarks.

Full paper

Evaluation


Task Dataset Model Metric name Metric value Global rank Compare
Text Classification AG News Capsule-B Error 7.4 # 6
Sentiment Analysis CR Capsule-B Accuracy 85.1 # 2
Sentiment Analysis MR Capsule-B Accuracy 82.3 # 2
Sentiment Analysis SST-2 Binary classification Capsule-B Accuracy 86.8 # 13
Subjectivity Analysis SUBJ Capsule-B Accuracy 93.8 # 5
Text Classification TREC-6 Capsule-B Error 7.2 # 9