Multiway Attention Networks for Modeling Sentence Pairs

Modeling sentence pairs plays the vital role forjudging the relationship between two sentences,such as paraphrase identification, natural languageinference, and answer sentence selection. Previouswork achieves very promising results using neuralnetworks with attention mechanism. In this paper,we propose the multiway attention networks whichemploy multiple attention functions to match sentence pairs under the matching-aggregation framework. Specifically, we design four attention functions to match words in corresponding sentences.Then, we aggregate the matching information fromeach function, and combine the information fromall functions to obtain the final representation. Experimental results demonstrate that the proposedmultiway attention networks improve the result onthe Quora Question Pairs, SNLI, MultiNLI, and answer sentence selection task on the SQuAD dataset.