Augmented Normalizing Flows: Bridging the Gap Between Generative Flows and Latent Variable Models

By Chin-Wei Huang, Laurent Dinh, Aaron C. Courville. 2020

99 citations recorded by the source corpus.

In this work, we propose a new family of generative flows on an augmented data space, with an aim to improve expressivity without drastically increasing the computational cost of sampling and evaluation of a lower bound on the likelihood. Theoretically, we prove the proposed flow can approximate a Hamiltonian ODE as a universal transport map. Empirically, we demonstrate state-of-the-art performance on standard benchmarks of flow-based generative modeling.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

1. Jan Kautz
2. Zhisheng Xiao
3. Yang Long
4. Sam Bond-Taylor
5. Frank No'e
6. Adam Leach
7. Ragavan Thurairatnam
8. Rayhane Mama

In Computer Graphics and Computer-Aided Design

1. Zhisheng Xiao
2. Frank No'e
3. Adam Leach
4. Ragavan Thurairatnam
5. Rayhane Mama
6. Louay Hazami
7. Chris G. Willcocks
8. Jan Kautz

In Computer Vision and Pattern Recognition

1. Arash Vahdat
2. Karsten Kreis
3. Jan Kautz
4. Zhisheng Xiao
5. Luc Van Gool
6. Chris G. Willcocks
7. Ragavan Thurairatnam
8. Rayhane Mama

In Generative Models

1. Arash Vahdat
2. Karsten Kreis
3. Jan Kautz
4. Zhisheng Xiao
5. Frank No'e
6. Ragavan Thurairatnam
7. Rayhane Mama
8. Louay Hazami

Citations in the corpus

• Stochastic Normalizing Flows (2020, background): “Other proposed solutions are real-and-discrete mixtures of flows [7] or augmentation of the bases space [8, 18] at the cost of losing asymptotically unbiased sampling.”
• Score-based Generative Modeling in Latent Space (2021, background): “However, previous works found that augmenting neural ODEs [16, 17] and more generally generative models [18, 19, 20, 21] with latent variables improves their expressivity.”
• Densely connected normalizing flows (2021, background): “Recent works propose to widen the flow by increasing the input dimensionality [24, 26].”
• Exponential Tilting of Generative Models: Improving Sample Quality by Training and Sampling from Latent Energy (2020, background): “Much effort has been devoted to improving the sample quality of non-adversarial, likelihood based generative models [26, 3, 11, 6] by modifying the training objectives.”
• Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models (2021, background): “Similar to augmented normalizing flows [94], this can be solved by providing additional dimensions for the flow to traverse [47].”
• Efficient-VDVAE: Less is more (2022, methodology): “In table 4, we test the effect of this removal on FFHQ 256× 256 and CelebAHQ 256× 256 which have been consistently quantized to 5-bits in prior work [16, 23, 47, 48, 38, 37, 32, 39, 34].”