Hybrid Tiled Convolutional Neural Networks for Text Sentiment Classification
This work addresses sentiment analysis for text data, but it is incremental as it modifies an existing method for a specific bottleneck.
The paper tackled the problem of adapting tiled convolutional neural networks from computer vision to NLP for sentiment classification, resulting in a hybrid tiled CNN that outperforms CNN and tiled CNN on IMDB and SemEval 2017 datasets.
The tiled convolutional neural network (tiled CNN) has been applied only to computer vision for learning invariances. We adjust its architecture to NLP to improve the extraction of the most salient features for sentiment analysis. Knowing that the major drawback of the tiled CNN in the NLP field is its inflexible filter structure, we propose a novel architecture called hybrid tiled CNN that applies a filter only on the words that appear in the similar contexts and on their neighbor words (a necessary step for preventing the loss of some n-grams). The experiments on the datasets of IMDB movie reviews and SemEval 2017 demonstrate the efficiency of the hybrid tiled CNN that performs better than both CNN and tiled CNN.