Answer triggering is the task of selecting the best-suited answer for a given question from a set of candidate answers if exists. In this paper, we present a hybrid deep learning model for answer triggering, which combines several dependency graph based alignment features, namely graph edit distance, graph-based similarity and dependency graph coverage, with dense vector embeddings from a Convolutional Neural Network (CNN). Our experiments on the WikiQA dataset show that such a combination can more accurately trigger a candidate answer compared to the previous state-of-the-art models. Comparative study on WikiQA dataset shows 5.86% absolute F-score improvement at the question level.
@article{arxiv.1808.01650,
title = {Combining Graph-based Dependency Features with Convolutional Neural Network for Answer Triggering},
author = {Deepak Gupta and Sarah Kohail and Pushpak Bhattacharyya},
journal= {arXiv preprint arXiv:1808.01650},
year = {2018}
}
Comments
19th International Conference on Computational Linguistics and Intelligent Text Processing (CICLing 2018)