Attribute-based Regularization of Latent Spaces for Variational Auto-Encoders
This addresses the need for interpretable and controllable generative models in machine learning, though it is incremental as it builds on existing VAE frameworks.
The paper tackles the problem of selectively manipulating data attributes in deep generative models by introducing an attribute regularization loss for Variational Auto-Encoders (VAEs) to structure the latent space, resulting in disentangled and interpretable latent spaces that effectively manipulate attributes across image and symbolic music domains.
Selective manipulation of data attributes using deep generative models is an active area of research. In this paper, we present a novel method to structure the latent space of a Variational Auto-Encoder (VAE) to encode different continuous-valued attributes explicitly. This is accomplished by using an attribute regularization loss which enforces a monotonic relationship between the attribute values and the latent code of the dimension along which the attribute is to be encoded. Consequently, post-training, the model can be used to manipulate the attribute by simply changing the latent code of the corresponding regularized dimension. The results obtained from several quantitative and qualitative experiments show that the proposed method leads to disentangled and interpretable latent spaces that can be used to effectively manipulate a wide range of data attributes spanning image and symbolic music domains.