A Style-Based Generator Architecture for Generative Adversarial Networks

• Category: Article
• Created: February 17, 2022 3:46 PM
• Status: Open
• URL: https://arxiv.org/pdf/1812.04948.pdf
• Updated: February 17, 2022 6:07 PM

Reference: https://towardsdatascience.com/explained-a-style-based-generator-architecture-for-gans-generating-and-tuning-realistic-6cb2be0f431

Highlights

1. The new generator architecture leads to an automatically learned, unsupervised separation of high-level attributes (eg. pose and identity when trained on human faces) and stochastic variation in the generated images (freckles, hair). and it enables intuitive, scale-specific control of the synthesis.
2. The new generator improves the state-of-the-art in terms of traditional distribution quality metrics, leads to demonstrably better interpolation properties, and also better disentangles the latent factors of variation.

Methods

Mapping Network

Given a latent code z in the input latent space \(z\), a non-linear mapping network \(f: z \rightarrow w\) first produces \(\mathbf{w} \in \mathcal{W}\). For simplicity, we set the dimensionality of both spaces to 512, and the mapping f is implemented using an 8-layer MLP.

StackGAN: Text to Photo-realistic Image Synthesis

with Stacked Generative Adversarial Networks

• Category: Article
• Created: February 16, 2022 1:54 PM
• Status: Open
• URL: https://arxiv.org/pdf/1612.03242.pdf
• Updated: February 16, 2022 6:11 PM

Highlights

1. In this paper, we propose Stacked Generative Adversarial Networks (StackGAN) to generate 256×256 photo-realistic images conditioned on text descriptions.
2. We decompose the hard problem into more manageable sub-problems through a sketch-refinement process.
3. we introduce a novel Conditioning Augmentation technique that encourages smoothness in the latent conditioning manifold.

Generative Adversarial Nets

Background

1. So far, the most striking successes in deep learning have involved discriminative models.
2. Deep generative models have had less of an impact, due to the difficulty of approximating many intractable probabilistic computations that arise in maximum likelihood estimation and related strategies, and due to difficulty of leveraging the benefits of piecewise linear units in the generative context.

Conditional Generative Adversarial Nets

• Category: Article
• Created: January 25, 2022 10:27 AM
• Status: Open
• URL: https://arxiv.org/pdf/1411.1784.pdf
• Updated: February 15, 2022 6:53 PM

Background

In this work we introduce the conditional version of generative adversarial nets, which can be constructed by simply feeding the data, \(y\), we wish to condition on to both the generator and discriminator.

Highlights

1. By conditioning the model on additional information it is possible to direct the data generation process.
2. Many interesting problems are more naturally thought of as a probabilistic one-to-many mapping. One way to address the problem is to use a conditional probabilistic generative model, the input is taken to be the conditioning variable and the one-to-many mapping is instantiated as a conditional predictive distribution.

Methods

Generative adversarial nets can be extended to a conditional model if both the generator and discriminator are conditioned on some extra information \(y\). \(y\) could be any kind of auxiliary information, such as class labels or data from other modalities. We can perform the conditioning by feeding \(y\) into the both the discriminator and generator as additional input layer.

Unsupervised representation learning with deep convolutional generative adversarial network

• Category: Article
• Created: January 24, 2022 5:31 PM
• Status: Open
• URL: https://arxiv.org/pdf/1511.06434.pdf
• Updated: February 15, 2022 6:49 PM

Background

1. In this work we hope to help bridge the gap between the success of CNNs for supervised learning and unsupervised learning.
2. We introduce a class of CNNs called deep convolutional generative adversarial networks (DCGANs), that have certain architectural constraints, and demonstrate that they are a strong candidate for unsupervised learning.

Methods

Architecture guidelines for stable Deep Convolutional GANs

1. Replace any pooling layers with strided convolutions (discriminator) and fractional-strided convolutions (generator).
2. Use batchnorm in both the generator and the discriminator.
3. Remove fully connected hidden layers for deeper architectures.
4. Use ReLU activation in generator for all layers except for the output, which uses Tanh.
5. Use LeakyReLU activation in the discriminator for all layers.

Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks

• Category: Article
• Created: January 24, 2022 6:36 PM
• Status: Open
• URL: https://arxiv.org/pdf/1506.05751.pdf
• Updated: January 25, 2022 10:21 AM

Background

Building a model capable of producing high quality samples of natural images.

Highlights

1. Our approach uses a cascade of convolutional networks within a Laplacian pyramid framework to generate images in a coarse-to-fine fashion.
2. At each level of the pyramid, a separate generative convnet model is trained using the Generative Adversarial Nets (GAN) approach.

Adversarial Autoencoders

• Category: Article
• Created: January 25, 2022 4:01 PM
• Status: Open
• URL: https://arxiv.org/abs/1511.05644
• Updated: January 29, 2022 11:39 AM

Highlights

1. We propose a general approach, called an adversarial autoencoder (AAE) that can turn an autoencoder into a generative model.
2. An autoencoder is trained with dual objectives – a traditional reconstruction error criterion, and an adversarial training criterion.

Methods

The adversarial autoencoder is an autoencoder that is regularized by matching the aggregated posterior, \(q(z)\), to an arbitrary prior, \(p(z)\). In order to do so, an adversarial network is attached on top of the hidden code vector of the autoencoder as illustrated in Figure 1. It is the adversarial network that guides \(q(z)\) to match \(p(z)\). The autoencoder, meanwhile, attempts to minimize the reconstruction error. The generator of the adversarial network is also the encoder of the autoencoder \(q(z|x)\). The encoder ensures the aggregated posterior distribution can fool the discriminative adversarial network into thinking that the hidden code \(q(z)\) comes from the true prior distribution \(p(z)\).

DRAW: A Recurrent Neural Network For Image Generation

• Category: Article
• Created: February 8, 2022 3:44 PM
• Status: Open
• URL: https://arxiv.org/pdf/1502.04623.pdf
• Updated: February 15, 2022 6:35 PM

Highlights

1. DRAW networks combine a novel spatial attention mechanism that mimics the foveation of the human eye, with a sequential variational auto-encoding framework that allows for the iterative construction of complex images.
2. Where DRAW differs from its siblings is that, rather than generating images in a single pass, it iteratively constructs scenes through an accumulation of modifications emitted by the decoder, each of which is observed by the encoder.

InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets

• Category: Article
• Created: February 8, 2022 7:26 PM
• Status: Open
• URL: https://arxiv.org/pdf/1606.03657.pdf
• Updated: February 15, 2022 6:12 PM

Highlights

1. Adversarial Network that is able to learn disentangled representations in a completely unsupervised manner.
2. In this paper, we present a simple modification to the generative adversarial network objective that encourages it to learn interpretable and meaningful representations.

Image-to-Image Translation with Conditional Adversarial Networks

• Category: Article
• Created: February 12, 2022 2:39 PM
• Status: Open
• URL: https://arxiv.org/pdf/1611.07004.pdf
• Updated: February 15, 2022 5:15 PM

Highlights

1. We investigate conditional adversarial networks as a general-purpose solution to image-to-image translation problems.
2. As a community, we no longer hand-engineer our mapping functions, and this work suggests we can achieve reasonable results without hand-engineering our loss functions either.

Intuition

1. If we take a naive approach and ask the CNN to minimize the Euclidean distance between predicted and ground truth pix- els, it will tend to produce blurry results.
2. It would be highly desirable if we could instead specify only a high-level goal, like “make the output indistinguishable from reality”, and then automatically learn a loss function appropriate for satisfying this goal.

Methods

Our generator we use a U-Net-based architecture, and for our discriminator we use a convolutional PatchGAN classifier, which only penalizes structure at the scale of image patches.