Glow: Generative Flow with Invertible 1x1 Convolutions
By Diederik P. Kingma, Prafulla Dhariwal. 2018
3,641 citations recorded by the source corpus.
Flow-based generative models (Dinh et al., 2014) are conceptually attractive due to tractability of the exact log-likelihood, tractability of exact latent-variable inference, and parallelizability of both training and synthesis. In this paper we propose Glow, a simple type of generative flow using an invertible 1x1 convolution. Using our method we demonstrate a significant improvement in log-likelihood on standard benchmarks. Perhaps most strikingly, we demonstrate that a generative model optimized towards the plain log-likelihood objective is capable of efficient realistic-looking synthesis and manipulation of large images. The code for our model is available at this https URL
Subfields: Computer Vision and Pattern Recognition, Statistical and Nonlinear Physics, Generative Models, Probabilistic & Bayesian Methods, Signal Processing
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Ming-Hsuan Yang
- Trevor Darrell
- Xiaogang Wang
- Jitendra Malik
- Andrew Zisserman
- Andrea Vedaldi
In Statistical and Nonlinear Physics
- Arash Vahdat
- Yang Song
- Stefano Ermon
- Karsten Kreis
- Diederik P. Kingma
- Jonathan Ho
- Yoshua Bengio
- Jun Zhu
In Generative Models
- Stefano Ermon
- Arash Vahdat
- Jiaming Song
- Karsten Kreis
- Jun-Yan Zhu
- Jan Kautz
- Diederik P. Kingma
- Yoshua Bengio
In Probabilistic & Bayesian Methods
- Stefano Ermon
- Yoshua Bengio
- Yang Song
- Max Welling
- Diederik P. Kingma
- Arash Vahdat
- Aaron C. Courville
- Sergey Levine
In Signal Processing
Citations in the corpus
- Denoising Diffusion Probabilistic Models (2020, background): “While diffusion models might resemble flows [9, 46, 10, 32, 5, 16, 23] and VAEs [33, 47, 37], diffusion models are designed so that q has no parameters and the top-level latent xT has nearly zero mutual information with the data x0.”
- High-Resolution Image Synthesis with Latent Diffusion Models (2021, background): “Variational autoencoders (VAE) [45] and flow-based models [18, 19] enable efficient synthesis of high resolution images [9, 43, 89], but sample quality is not on par with GANs.”
- Stochastic Normalizing Flows (2020, background): “NFs can be categorized into two families: (i) Coupling layers (Dinh et al., 2014; 2016; Kingma & Dhariwal, 2018; M ¨ uller et al., 2018), which are a subclass of autoregressive flows (Germain et al., 2015; Papamakarios et al., 2017; Huang et al., 2018; De Cao et al., 2019; Durkan et al., 2019), and…”
- Your Classifier is Secretly an Energy Based Model and You Should Treat it Like One (2019, background): “Nalisnick et al. (2018) showed that tractable deep generative models such as Kingma & Dhariwal (2018) and Salimans et al. (2017) can assign higher densities to OOD examples than in-distribution examples.”
- Flow Contrastive Estimation of Energy-Based Models (2019, methodology): “The gradient of the first term resembles MLE, which forces q α to cover the modes of data distribution, and tends to lead to an over-dispersed model, which is also pointed out in [32].”
- Generating Diverse High-Fidelity Images with VQ-VAE-2 (2019, methodology): “We distinguish two main types of generative models: likelihood based models, which include VAEs [16, 31], flow based [9, 30, 10, 17] and autoregressive models [20, 39]; and implicit generative ∗Equal contributions.”
- Learning Non-Convergent Non-Persistent Short-Run MCMC Toward Energy-Based Model (2019, methodology): “We also show that the learned short-run MCMC can be used as a generative model, such as a generator model [13, 24] or the flow model [8, 9, 25, 4, 14], with the Langevin dynamics serving as a noiseinjected residual network, with the initial image serving as the latent variables, and with the initial”
- BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modeling (2019, methodology): “The three main classes of powerful explicit density models are autoregressive models [26, 57], flow-based models [8, 9, 21, 16], and probabilistic latent variable models [24, 40, 33].”
- Flow++: Improving Flow-Based Generative Models with Variational Dequantization and Architecture Design (2019, methodology): “The key idea in the de-sign of a flow model is to form f by stacking individual simple invertible transformations (Dinh et al., 2014; 2016; Kingma & Dhariwal, 2018; Rezende & Mohamed, 2015; Kingma et al., 2016; Louizos & Welling, 2017).”
- A Style-Based Generator Architecture for Generative Adversarial Networks (2018, methodology): “Here we used a variation of the truncation trick [38, 4, 31] with ψ = 0.”
- Waveglow: A Flow-based Generative Network for Speech Synthesis (2018, methodology): “We take the flow-based approach of [1] and include the architectural insights of WaveNet.”
- FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Models (2018, methodology): “On MNIST we find that FFJORD can model the data as well as Glow and Real NVP by integrating a single flow defined by one neural network.”
- Generating High Fidelity Images with Subscale Pixel Networks and Multidimensional Upscaling (2018, methodology): “We demonstrate experimentally that our model is capable of high fidelity samples at high resolution, producing unconditional CelebA-HQ samples of quality better than the Glow model (Kingma & Dhariwal, 2018) and improving the MLE scores.”
- Classifier-Free Diffusion Guidance (2022, background): “Low temperature sampling in Glow [7] has a similar effect.”
- Projected GANs Converge Faster (2021, background): “[28] initialize their convolutional layers as a random rotation matrix as a good starting point for optimization.”