|
gptkbp:instanceOf
|
gptkb:model
|
|
gptkbp:application
|
gptkb:fashion_designer
data privacy
drug discovery
speech synthesis
medical imaging
music generation
semantic segmentation
image editing
domain adaptation
art generation
3D object generation
face synthesis
handwriting generation
inpainting
video prediction
|
|
gptkbp:consistsOf
|
generator
discriminator
|
|
gptkbp:discriminatorGoal
|
distinguish real from fake data
|
|
gptkbp:fullName
|
gptkb:convolutional_neural_network
|
|
gptkbp:generatorGoal
|
produce realistic data
|
|
https://www.w3.org/2000/01/rdf-schema#label
|
GAN
|
|
gptkbp:introducedIn
|
2014
|
|
gptkbp:inventedBy
|
gptkb:Ian_Goodfellow
|
|
gptkbp:limitation
|
require large datasets
mode collapse
sensitive to hyperparameters
training instability
difficulty in evaluation
vanishing gradients
|
|
gptkbp:notablePublication
|
gptkb:Generative_Adversarial_Nets_(2014)
|
|
gptkbp:notableVariant
|
gptkb:BigGAN
gptkb:Conditional_GAN
gptkb:DCGAN
gptkb:InfoGAN
gptkb:Progressive_GAN
gptkb:SRGAN
gptkb:StackGAN
gptkb:StyleGAN
gptkb:Wasserstein_GAN
gptkb:CycleGAN
gptkb:Pix2Pix
gptkb:LSGAN
|
|
gptkbp:objective
|
minimax game
|
|
gptkbp:openSource
|
gptkb:TensorFlow
gptkb:Keras
gptkb:MXNet
gptkb:PyTorch
|
|
gptkbp:platform
|
two neural networks
|
|
gptkbp:relatedTo
|
gptkb:machine_learning
deep learning
unsupervised learning
|
|
gptkbp:trainer
|
adversarial training
|
|
gptkbp:usedFor
|
image generation
data augmentation
anomaly detection
image-to-image translation
style transfer
super-resolution
text-to-image synthesis
video generation
|
|
gptkbp:bfsParent
|
gptkb:UNEP_Global_Adaptation_Network
gptkb:convolutional_neural_network
gptkb:CycleGAN
|
|
gptkbp:bfsLayer
|
5
|