Deconvolutional Latent-Variable Model for Text Sequence Matching
This work addresses text matching challenges for natural language processing applications, offering an incremental improvement over existing methods.
The paper tackles the problem of text sequence matching by introducing a latent-variable model that uses deconvolutional networks as decoders to improve semantic information and generalization, resulting in stronger predictive performance than LSTM-based decoders with fewer parameters and faster training, and it significantly outperforms baselines in semi-supervised settings.
A latent-variable model is introduced for text matching, inferring sentence representations by jointly optimizing generative and discriminative objectives. To alleviate typical optimization challenges in latent-variable models for text, we employ deconvolutional networks as the sequence decoder (generator), providing learned latent codes with more semantic information and better generalization. Our model, trained in an unsupervised manner, yields stronger empirical predictive performance than a decoder based on Long Short-Term Memory (LSTM), with less parameters and considerably faster training. Further, we apply it to text sequence-matching problems. The proposed model significantly outperforms several strong sentence-encoding baselines, especially in the semi-supervised setting.