The Hessian Penalty: A Weak Prior for Unsupervised Disentanglement
Existing disentanglement methods for deep generative models rely on hand-picked priors and complex encoder-based architectures. In this paper, we propose the Hessian Penalty, a simple regularization term that encourages the Hessian of a generative model with respect to its input to be diagonal. We introduce a model-agnostic, unbiased stochastic approximation of this term based on Hutchinson's estimator to compute it efficiently during training. Our method can be applied to a wide range of deep generators with just a few lines of code. We show that training with the Hessian Penalty often causes axis-aligned disentanglement to emerge in latent space when applied to ProGAN on several datasets. Additionally, we use our regularization term to identify interpretable directions in BigGAN's latent space in an unsupervised fashion. Finally, we provide empirical evidence that the Hessian Penalty encourages substantial shrinkage when applied to over-parameterized latent spaces.
Code (1)
Tasks
DisentanglementMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Orthogonal Jacobian Regularization for Unsupervised Disentanglement in Image Generation
Unsupervised disentanglement learning is a crucial issue for understanding and exploiting deep generative models. Recently, SeFa tries to find latent disentangled directions by performing SVD on the first projection of a…
DisentanglementImage GenerationWeakly Supervised Disentanglement by Pairwise Similarities
Recently, researches related to unsupervised disentanglement learning with deep generative models have gained substantial popularity. However, without introducing supervision, there is no guarantee that the factors of in…
DisentanglementWeakly Supervised Disentanglement with Guarantees
Learning disentangled representations that correspond to factors of variation in real-world data is critical to interpretable and human-controllable machine learning. Recently, concerns about the viability of learning di…
DisentanglementSW-VAE: Weakly Supervised Learn Disentangled Representation Via Latent Factor Swapping
Representation disentanglement is an important goal of representation learning that benefits various downstream tasks. To achieve this goal, many unsupervised learning representation disentanglement approaches have been …
DisentanglementRepresentation LearningA Causal Ordering Prior for Unsupervised Representation Learning
Unsupervised representation learning with variational inference relies heavily on independence assumptions over latent variables. Causal representation learning (CRL), however, argues that factors of variation in a datas…
Causal DiscoverycounterfactualRepresentation LearningVariational Inference