Demystifying Inter-Class Disentanglement
This work addresses the challenge of learning disentangled representations for AI applications, offering a novel approach that improves upon current methods in supervised settings.
The paper tackles the problem of disentangling class and content factors in data by introducing LORD, a method based on latent optimization with asymmetric noise regularization, which outperforms existing adversarial and non-adversarial methods in disentanglement performance.
Learning to disentangle the hidden factors of variations within a set of observations is a key task for artificial intelligence. We present a unified formulation for class and content disentanglement and use it to illustrate the limitations of current methods. We therefore introduce LORD, a novel method based on Latent Optimization for Representation Disentanglement. We find that latent optimization, along with an asymmetric noise regularization, is superior to amortized inference for achieving disentangled representations. In extensive experiments, our method is shown to achieve better disentanglement performance than both adversarial and non-adversarial methods that use the same level of supervision. We further introduce a clustering-based approach for extending our method for settings that exhibit in-class variation with promising results on the task of domain translation.