Group Sparse CNNs for Question Classification with Answer Sets
This work addresses the problem of question classification for applications like QA systems by integrating answer data, representing an incremental advance over traditional methods.
The paper tackled question classification by incorporating answer set information into question modeling, achieving significant performance improvements over strong baselines on four datasets.
Question classification is an important task with wide applications. However, traditional techniques treat questions as general sentences, ignoring the corresponding answer data. In order to consider answer information into question modeling, we first introduce novel group sparse autoencoders which refine question representation by utilizing group information in the answer set. We then propose novel group sparse CNNs which naturally learn question representation with respect to their answers by implanting group sparse autoencoders into traditional CNNs. The proposed model significantly outperform strong baselines on four datasets.