Improving GAN Training via Feature Space Shrinkage
- URL: http://arxiv.org/abs/2303.01559v2
- Date: Sat, 8 Apr 2023 11:36:38 GMT
- Title: Improving GAN Training via Feature Space Shrinkage
- Authors: Haozhe Liu, Wentian Zhang, Bing Li, Haoqian Wu, Nanjun He, Yawen
Huang, Yuexiang Li, Bernard Ghanem, Yefeng Zheng
- Abstract summary: We propose AdaptiveMix, which shrinks regions of training data in the image representation space of the discriminator.
Considering it is intractable to directly bound feature space, we propose to construct hard samples and narrow down the feature distance between hard and easy samples.
The evaluation results demonstrate that our AdaptiveMix can facilitate the training of GANs and effectively improve the image quality of generated samples.
- Score: 69.98365478398593
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Due to the outstanding capability for data generation, Generative Adversarial
Networks (GANs) have attracted considerable attention in unsupervised learning.
However, training GANs is difficult, since the training distribution is dynamic
for the discriminator, leading to unstable image representation. In this paper,
we address the problem of training GANs from a novel perspective, \emph{i.e.,}
robust image classification. Motivated by studies on robust image
representation, we propose a simple yet effective module, namely AdaptiveMix,
for GANs, which shrinks the regions of training data in the image
representation space of the discriminator. Considering it is intractable to
directly bound feature space, we propose to construct hard samples and narrow
down the feature distance between hard and easy samples. The hard samples are
constructed by mixing a pair of training images. We evaluate the effectiveness
of our AdaptiveMix with widely-used and state-of-the-art GAN architectures. The
evaluation results demonstrate that our AdaptiveMix can facilitate the training
of GANs and effectively improve the image quality of generated samples. We also
show that our AdaptiveMix can be further applied to image classification and
Out-Of-Distribution (OOD) detection tasks, by equipping it with
state-of-the-art methods. Extensive experiments on seven publicly available
datasets show that our method effectively boosts the performance of baselines.
The code is publicly available at
https://github.com/WentianZhang-ML/AdaptiveMix.
Related papers
- Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think [72.48325960659822]
One main bottleneck in training large-scale diffusion models for generation lies in effectively learning these representations.
We study this by introducing a straightforward regularization called REPresentation Alignment (REPA), which aligns the projections of noisy input hidden states in denoising networks with clean image representations obtained from external, pretrained visual encoders.
The results are striking: our simple strategy yields significant improvements in both training efficiency and generation quality when applied to popular diffusion and flow-based transformers, such as DiTs and SiTs.
arXiv Detail & Related papers (2024-10-09T14:34:53Z) - E$^{2}$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation [69.72194342962615]
We introduce and address a novel research direction: can the process of distilling GANs from diffusion models be made significantly more efficient?
First, we construct a base GAN model with generalized features, adaptable to different concepts through fine-tuning, eliminating the need for training from scratch.
Second, we identify crucial layers within the base GAN model and employ Low-Rank Adaptation (LoRA) with a simple yet effective rank search process, rather than fine-tuning the entire base model.
Third, we investigate the minimal amount of data necessary for fine-tuning, further reducing the overall training time.
arXiv Detail & Related papers (2024-01-11T18:59:14Z) - The Journey, Not the Destination: How Data Guides Diffusion Models [75.19694584942623]
Diffusion models trained on large datasets can synthesize photo-realistic images of remarkable quality and diversity.
We propose a framework that: (i) provides a formal notion of data attribution in the context of diffusion models, and (ii) allows us to counterfactually validate such attributions.
arXiv Detail & Related papers (2023-12-11T08:39:43Z) - Few-shot Image Generation via Masked Discrimination [20.998032566820907]
Few-shot image generation aims to generate images of high quality and great diversity with limited data.
It is difficult for modern GANs to avoid overfitting when trained on only a few images.
This work presents a novel approach to realize few-shot GAN adaptation via masked discrimination.
arXiv Detail & Related papers (2022-10-27T06:02:22Z) - ReMix: Towards Image-to-Image Translation with Limited Data [154.71724970593036]
We propose a data augmentation method (ReMix) to tackle this issue.
We interpolate training samples at the feature level and propose a novel content loss based on the perceptual relations among samples.
The proposed approach effectively reduces the ambiguity of generation and renders content-preserving results.
arXiv Detail & Related papers (2021-03-31T06:24:10Z) - Ultra-Data-Efficient GAN Training: Drawing A Lottery Ticket First, Then
Training It Toughly [114.81028176850404]
Training generative adversarial networks (GANs) with limited data generally results in deteriorated performance and collapsed models.
We decompose the data-hungry GAN training into two sequential sub-problems.
Such a coordinated framework enables us to focus on lower-complexity and more data-efficient sub-problems.
arXiv Detail & Related papers (2021-02-28T05:20:29Z) - Unlabeled Data Guided Semi-supervised Histopathology Image Segmentation [34.45302976822067]
Semi-supervised learning (SSL) based on generative methods has been proven to be effective in utilizing diverse image characteristics.
We propose a new data guided generative method for histopathology image segmentation by leveraging the unlabeled data distributions.
Our method is evaluated on glands and nuclei datasets.
arXiv Detail & Related papers (2020-12-17T02:54:19Z) - Multiclass non-Adversarial Image Synthesis, with Application to
Classification from Very Small Sample [6.243995448840211]
We present a novel non-adversarial generative method - Clustered Optimization of LAtent space (COLA)
In the full data regime, our method is capable of generating diverse multi-class images with no supervision.
In the small-data regime, where only a small sample of labeled images is available for training with no access to additional unlabeled data, our results surpass state-of-the-art GAN models trained on the same amount of data.
arXiv Detail & Related papers (2020-11-25T18:47:27Z) - Training End-to-end Single Image Generators without GANs [27.393821783237186]
AugurOne is a novel approach for training single image generative models.
Our approach trains an upscaling neural network using non-affine augmentations of the (single) input image.
A compact latent space is jointly learned allowing for controlled image synthesis.
arXiv Detail & Related papers (2020-04-07T17:58:03Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.