gan   1237

« earlier    

ENHANCE! ‘Photo-Realistic’ Emojis and Emotes With Progressive Face Super-Resolution
Because this model is trained specifically to look for facial landmarks it will take any excuse to draw eyes and nostrils on a pixel. And I’m pretty surethe pepperoni on that pizza is made out of human lips…
yesterday by NightOwlCity
[1910.11626] Seeing What a GAN Cannot Generate
"Despite the success of Generative Adversarial Networks (GANs), mode collapse remains a serious issue during GAN training. To date, little work has focused on understanding and quantifying which modes have been dropped by a model. In this work, we visualize mode collapse at both the distribution level and the instance level. First, we deploy a semantic segmentation network to compare the distribution of segmented objects in the generated images with the target distribution in the training set. Differences in statistics reveal object classes that are omitted by a GAN. Second, given the identified omitted object classes, we visualize the GAN's omissions directly. In particular, we compare specific differences between individual photos and their approximate inversions by a GAN. To this end, we relax the problem of inversion and solve the tractable problem of inverting a GAN layer instead of the entire generator. Finally, we use this framework to analyze several recent GANs trained on multiple datasets and identify their typical failure cases."
15 days ago by arsyed
gallium nitride transistor manufacturer, specializing in bespoke semiconductor solutions (like SAR transmitters for sats)
GaN  gallium  nitride  transistor  semiconductor  manufacturer 
21 days ago by asteroza

« earlier    

related tags

2019  adversarial-learning  adversary  ai  animals  anime  application  architecture  armpl  art  audio  autoencoders  biggan  book  books  by:liorpachter  camouflage  chargers  cnn  cocish  coco  coevolution  competition  computervision  courses  dataanalysis  datacompression  datascience  dataset  deception  deep-learning  deep  deepdream  deepfake  deepfakes  deeplearning  design  detournement  differentialprivacy  dimension  dimensionality  dimensionalityreduction  disentanglement  dl  encoders  face  faces  facial-recognition  factor_models  fake  floorplan  free  freeculture  gallium  games  gans  garfield  gems  generative  generativeadversarialnetworks  generator  generators  gif  github  gpu  graphic  hallucination  hcomp  human-in-the-loop  humor  image  imageprocessing  images  infogan  inspiration  interpretability  keras  landscape  latentspace  learning  lowdimensionaldata  machine-learning  machine  machine_learning  machinelearning  magic  manufacturer  material  misinformation  ml  mocap  movie  music  netart  network  neural  neuralnets  neuralnetwork  neuralnetworks  neuralsynesthesia  nitride  nn  opensource  overview  painting  papers  pca  perception  photo  photograph  photography  photos  picture  pix2pix  pixelart  pose-tracking  principalcomponentanalysis  privacy  procedural  project  python  pytorch  raygan  remix  res  research  retronet  rotoscope  science  selfie  semiconductor  specialeffects  speech-synthesis  speech  stock  style  stylegan  styletransfer  survey  svm  synthetic  syntheticrealities  tensorflow  texture  texturesynthesis  transfer  transistor  trends2019  trends2020  truth  tts  tutorial  tutorials  tweetit  twitter  usb  usecase  v2020  vae  variationalautoencoders  vfx  video  visual  visualisation  wasserstein  wtf  yolo  your_favorite_deep_neural_network_sucks 

Copy this bookmark: