On the Importance of the Kullback-Leibler Divergence Term in Variational Autoencoders for Text Generation
This work addresses a known bottleneck in VAEs for text generation, offering incremental improvements in understanding and controlling latent representations.
The authors tackled the problem of uninformative latent representations in Variational Autoencoders for text generation by imposing an explicit constraint on the Kullback-Leibler divergence term, which helped avoid posterior collapse and revealed a trade-off between information encoding and generative capacity.
Variational Autoencoders (VAEs) are known to suffer from learning uninformative latent representation of the input due to issues such as approximated posterior collapse, or entanglement of the latent space. We impose an explicit constraint on the Kullback-Leibler (KL) divergence term inside the VAE objective function. While the explicit constraint naturally avoids posterior collapse, we use it to further understand the significance of the KL term in controlling the information transmitted through the VAE channel. Within this framework, we explore different properties of the estimated posterior distribution, and highlight the trade-off between the amount of information encoded in a latent code during training, and the generative capacity of the model.