A Hybrid Convolutional Variational Autoencoder for Text Generation
arXiv:1702.02390
Abstract
In this paper we explore the effect of architectural choices on learning a Variational Autoencoder (VAE) for text generation. In contrast to the previously introduced VAE model for text where both the encoder and decoder are RNNs, we propose a novel hybrid architecture that blends fully feed-forward convolutional and deconvolutional components with a recurrent language model. Our architecture exhibits several attractive properties such as faster run time and convergence, ability to better handle long sequences and, more importantly, it helps to avoid some of the major difficulties posed by training VAE models on textual data.
References in corpus (8)
- Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
- Neural Architecture Search with Reinforcement Learning
- Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
- Learning Deconvolution Network for Semantic Segmentation
- Layer Normalization
- Neural Machine Translation in Linear Time
- Show and Tell: A Neural Image Caption Generator
- Techniques for Learning Binary Stochastic Feedforward Neural Networks
Cited by in corpus (7)
- Deconvolutional Paragraph Representation Learning
- Topic-Guided Variational Autoencoders for Text Generation
- Image Captioning at Will: A Versatile Scheme for Effectively Injecting Sentiments into Image Descriptions
- Z-Forcing: Training Stochastic Recurrent Networks
- UVA: A Universal Variational Framework for Continuous Age Analysis
- Variational Auto-Encoder Based Approximate Bayesian Computation Uncertian Inverse Method for Sheet Metal Forming Problem
- Deep Generative Models: Deterministic Prediction with an Application in Inverse Rendering