A Stable Variational Autoencoder for Text Modelling
This addresses a key stability issue in VAE-based text generation, which is incremental as it builds on existing VAE-RNN architectures.
The paper tackles the problem of latent variable collapse in Variational Autoencoders (VAEs) for text modeling, introducing a holistic regularization VAE (HR-VAE) that achieves more stable training and generates text with significantly better quality.
Variational Autoencoder (VAE) is a powerful method for learning representations of high-dimensional data. However, VAEs can suffer from an issue known as latent variable collapse (or KL loss vanishing), where the posterior collapses to the prior and the model will ignore the latent codes in generative tasks. Such an issue is particularly prevalent when employing VAE-RNN architectures for text modelling (Bowman et al., 2016). In this paper, we present a simple architecture called holistic regularisation VAE (HR-VAE), which can effectively avoid latent variable collapse. Compared to existing VAE-RNN architectures, we show that our model can achieve much more stable training process and can generate text with significantly better quality.