Demystifying MMD GANs

Bińkowski, Mikołaj; Sutherland, Danica J.; Arbel, Michael; Gretton, Arthur

doi:10.48550/arxiv.1801.01401

Cited by 138 publications

(208 citation statements)

References 0 publications

Supporting

Mentioning

189

Contrasting

Order By: Relevance

“…Metrics. We evaluate our methods on FID [15], KID [3] for image quality. To assess the overfitting of the methods, we further evaluate them on latent recovery [54], density and coverage [39] metrics.…”

Section: Results Of Slim-fastgansmentioning

confidence: 99%

Learning Inception Attention for Image Synthesis and Image Recognition

Shen¹,

Wu²

2021

Preprint

View full text Add to dashboard Cite

Image synthesis and image recognition have witnessed remarkable progress, but often at the expense of computationally expensive training and inference. Learning lightweight yet expressive deep model has emerged as an important and interesting direction. Inspired by the well-known split-transform-aggregate design heuristic in the Inception building block, this paper proposes a Skip-Layer Inception Module (SLIM) that facilitates efficient learning of image synthesis models, and a same-layer variant (dubbed as SLIM too) as a stronger alternative to the well-known ResNeXts for image recognition. The proposed method is built on the idea of developing spatially-adaptive feature modulation/attention in a network. Specifically, SLIM is done from an input source feature map to a target one, either skip-layer or same-layer. In SLIM, the input feature map is first split into a number of groups (e.g., 4). Each group is then transformed to a latent style vector (via channel-wise attention) and a latent spatial mask (via spatial attention). The learned latent masks and latent style vectors are aggregated to modulate the target feature map. For generative learning, SLIM is built on a recently proposed lightweight Generative Adversarial Networks (i.e., FastGANs) which present a skip-layer excitation (SLE) module. For few-shot image synthesis tasks, the proposed SLIM achieves better performance than the SLE work and other related methods. For one-shot image synthesis tasks, it shows stronger capability of preserving image structures than prior arts such as the SinGANs. For image classification tasks, the proposed SLIM is used as a drop-in replacement for convolution layers in ResNets (resulting the ResNeXt-like models) and achieves better accuracy in the ImageNet-1000 dataset, with significantly smaller model complexity.

show abstract

Section: Results Of Slim-fastgansmentioning

confidence: 99%

Learning Inception Attention for Image Synthesis and Image Recognition

Shen¹,

Wu²

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…For ViT discriminator, they use embedding feature loss calculated using positional and patch features from the transformer encoder layers by successfully inserting the real and synthesized FA images. Their quantitative results in terms of Frechet inception Distance [320] and Kernel Inception Distance [321] demonstrate the superiority of their approach over baseline methods on diabetic retinopathy dataset provided by Hajeb et al [322].…”

Section: Semi-supervised Methodsmentioning

confidence: 92%

Transformers in Medical Imaging: A Survey

Shamshad¹,

Khan²,

Zamir³

et al. 2022

Preprint

View full text Add to dashboard Cite

Following unprecedented success on the natural language tasks, Transformers have been successfully applied to several computer vision problems, achieving state-of-the-art results and prompting researchers to reconsider the supremacy of convolutional neural networks (CNNs) as de facto operators. Capitalizing on these advances in computer vision, the medical imaging field has also witnessed growing interest for Transformers that can capture global context compared to CNNs with local receptive fields. Inspired from this transition, in this survey, we attempt to provide a comprehensive review of the applications of Transformers in medical imaging covering various aspects, ranging from recently proposed architectural designs to unsolved issues. Specifically, we survey the use of Transformers in medical image segmentation, detection, classification, reconstruction, synthesis, registration, clinical report generation, and other tasks. In particular, for each of these applications, we develop taxonomy, identify application-specific challenges as well as provide insights to solve them, and highlight recent trends. Further, we provide a critical discussion of the field's current state as a whole, including the identification of key challenges, open problems, and outlining promising future directions. We hope this survey will ignite further interest in the community and provide researchers with an up-to-date reference regarding applications of Transformer models in medical imaging. Finally, to cope with the rapid development in this field, we intend to regularly update the relevant latest papers and their open-source implementations at https://github.com/fahadshamshad/awesome-transformers-in-medical-imaging.

show abstract

“…Evaluating generative models in any domain is a notoriously difficult task (Theis et al, 2016). However, researchers have found success through the use of sample-based evaluation metrics that estimate the distance ρ between real and generated distributions P r and P g by drawing random samples (Heusel et al, 2017;You et al, 2018;Bińkowski et al, 2018). That is, they compute ρ(S g , S r ) ≈ ρ(P g , P r ), with S r = {x r 1 , .…”

Section: Background and Related Workmentioning

confidence: 99%

“…MMD (Gretton et al, 2006) (Equation 1) can also be used to measure the dissimilarity between graph embedding distributions. The original Kernel Inception Distance (KID) (Bińkowski et al, 2018) proposed a polynomial kernel with MMD, k(x i , x j ) = ( 1 d x i x j + 1) 3 , where d is the embedding dimension. The linear kernel k(x i , x j ) = x i • x j is another parameter-free kernel used with MMD to evaluate generative models (O'Bray et al, 2021).…”

Section: Neural-network-based Metricsmentioning

confidence: 99%

“…In image generation evaluation, classifiers trained on ImageNet (Deng et al, 2009) are frequently used to extract image embeddings (Bińkowski et al, 2018;Heusel et al, 2017;Kynkäänniemi et al, 2019;Xu et al, 2018;Naeem et al, 2020). While classifiers such as Inception v3 (Szegedy et al, 2016) are consistently used, recent works have investigated the use of randomly-initialized CNNs with no further training (hereafter referred to as a random network) in generative model evaluation.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On Evaluation Metrics for Graph Generative Models

Thompson¹,

Knyazev²,

Ghalebi³

et al. 2022

Preprint

View full text Add to dashboard Cite

In image generation, generative models can be evaluated naturally by visually inspecting model outputs. However, this is not always the case for graph generative models (GGMs), making their evaluation challenging. Currently, the standard process for evaluating GGMs suffers from three critical limitations: i) it does not produce a single score which makes model selection challenging, ii) in many cases it fails to consider underlying edge and node features, and iii) it is prohibitively slow to perform. In this work, we mitigate these issues by searching for scalar, domain-agnostic, and scalable metrics for evaluating and ranking GGMs.To this end, we study existing GGM metrics and neural-network-based metrics emerging from generative models of images that use embeddings extracted from a task-specific network. Motivated by the power of certain Graph Neural Networks (GNNs) to extract meaningful graph representations without any training, we introduce several metrics based on the features extracted by an untrained random GNN. We design experiments to thoroughly test metrics on their ability to measure the diversity and fidelity of generated graphs, as well as their sample and computational efficiency. Depending on the quantity of samples, we recommend one of two random-GNN-based metrics that we show to be more expressive than pre-existing metrics. While we focus on applying these metrics to GGM evaluation, in practice this enables the ability to easily compute the dissimilarity between any two sets of graphs regardless of domain. Our code is released at: https://github.com/uoguelph-mlrg/GGM-metrics.

show abstract

Demystifying MMD GANs

Cited by 138 publications

References 0 publications

Learning Inception Attention for Image Synthesis and Image Recognition

Learning Inception Attention for Image Synthesis and Image Recognition

Transformers in Medical Imaging: A Survey

On Evaluation Metrics for Graph Generative Models

Contact Info

Product

Resources

About