The existing unsupervised GML solution for Aspect-Term Sentiment Analysis (ATSA) is limited by inaccurate and insufficient knowledge conveyance. However, the Supervised GML (S-GML) can effectively exploit labeled training data to improve knowledge conveyance. The S-GML has the following characteristics:
-
It leverages binary polarity relations between instances, the most direct way of knowledge conveyance, to enable supervised gradual learning.
-
Besides the explicit polarity relations indicated by discourse structures, it also separately supervises a polarity classification DNN and a binary siamese network to extract implicit polarity relations.