Search papers, labs, and topics across Lattice.
This paper introduces Decafs, a novel conditional generator that leverages Lie groups to disentangle latent embeddings in flow-based models, addressing the interpretability challenges posed by entangled generative factors. By employing an adversarial loss to align the disentangled latent space with the latent flow space, the method enables controlled and interpretable generation without increasing the dimensionality of the flow space. The results show that Decafs outperforms existing models like StyleGAN on benchmark datasets for both image and molecular generation tasks, highlighting its effectiveness in generative modeling.
Disentangling latent embeddings in flow-based models allows for controlled generation without dimensionality expansion, outperforming StyleGAN on key benchmarks.
Flow-based models have established state-of-the-art performance in generative modeling across domains, but are hard to interpret due to their complex latent embeddings. In particular, the entanglement of generative factors in the latent space hinders controlled generation. We circumvent this issue by appealing to a novel conditional generator based on Lie groups that disentangles an alternative latent space, which is aligned closely with the latent flow space using an adversarial loss. Our approach facilitates interpretable conditional generation while obviating the need to expand the dimensionality of the flow space (owing to its invertibility requirements). The proposed model demonstrates strong performance across conditional image (including, outperforming StyleGAN on MNIST, dSprites) and molecule (using standard QM9, ZINC and MOSES) generation tasks
