Generative Adversarial Network PDF
Preview Generative Adversarial Network
Generative Adversarial Network (GAN) Restricted Boltzmann Machine: http://speech.ee.ntu.edu.tw/~tlkagk/courses/MLDS_2015_2/Lecture/RBM %20(v2).ecm.mp4/index.html Outlook: Gibbs Sampling: http://speech.ee.ntu.edu.tw/~tlkagk/courses/MLDS_2015_2/Lecture/MRF %20(v2).ecm.mp4/index.html NIPS 2016 Tutorial: Generative Adversarial Networks • Author: Ian Goodfellow • Paper: https://arxiv.org/abs/1701.00160 • Video: https://channel9.msdn.com/Events/Neural- Information-Processing-Systems- Conference/Neural-Information-Processing- Systems-Conference-NIPS-2016/Generative- Adversarial-Networks You can find tips for training GAN here: https://github.com/soumith/ganhacks Review http://www.rb139.com/index.ph p?s=/Lot/44547 Generation Writing Poems? Drawing? Review: Auto-encoder As close as possible NN c NN o d Encoder Decoder e Randomly generate c NN o Image ? d a vector as code Decoder e Review: Auto-encoder 2D NN code Decoder NN Decoder 1.5 0 -1.5 1.5 −1.5 0 NN Decoder Review: Auto-encoder -1.5 1.5 Auto-encoder NN NN input output Encoder Decoder code VAE Minimize m 1 reconstruction error m 2 m NN 𝑐 3 input 1 NN Encoder 𝜎 + 𝑐 output 1 exp 2 Decoder 𝑐 𝜎 3 2 𝜎 3 X 𝑐 = 𝑒𝑥𝑝 𝜎 × 𝑒 + 𝑚 𝑖 𝑖 𝑖 𝑖 𝑒 1 From a normal 𝑒 Minimize 2 distribution 𝑒 3 3 2 𝑒𝑥𝑝 𝜎 − 1 + 𝜎 + 𝑚 𝑖 𝑖 𝑖 Auto-Encoding Variational Bayes, 𝑖=1 https://arxiv.org/abs/1312.6114 Problems of VAE • It does not really try to simulate real images code NN As close as Output Decoder possible One pixel difference One pixel difference from the target from the target Realistic Fake The evolution of generation NN NN NN Generator Generator Generator v1 v2 v3 Discri- Discri- Discri- minator minator minator v1 v2 v3 Binary Classifier Real images:
Description:The list of books you might like
