Dynamic Feature Generation Network for Answer Selection
This work addresses feature extraction for answer selection, an incremental improvement in natural language processing.
The paper tackles the problem of feature representation for answer selection by introducing a Dynamic Feature Generation Network (DFGN) that generates and filters sentence-level features, resulting in significant outperformance over state-of-the-art baselines on multiple datasets.
Extracting appropriate features to represent a corpus is an important task for textual mining. Previous attention based work usually enhance feature at the lexical level, which lacks the exploration of feature augmentation at the sentence level. In this paper, we exploit a Dynamic Feature Generation Network (DFGN) to solve this problem. Specifically, DFGN generates features based on a variety of attention mechanisms and attaches features to sentence representation. Then a thresholder is designed to filter the mined features automatically. DFGN extracts the most significant characteristics from datasets to keep its practicability and robustness. Experimental results on multiple well-known answer selection datasets show that our proposed approach significantly outperforms state-of-the-art baselines. We give a detailed analysis of the experiments to illustrate why DFGN provides excellent retrieval and interpretative ability.