Download to read offline
![G
D
Latent Variables
(add noise)
Fake
Training Data
Real
MLP
MLP
Relu
B
A
b b=[1,1,1,1,1,1,1,1,1,1] N=10
B=[0,0,0,0,0,1,1,1,1,1]
P(B=b) = 0.5
H(B)=-SumP(B=b).logP(B=b)
H(B)=-0.5.log(0.5)
H(B)=-0.15
H(A|B=b)=-0.8.log(0.8)
H(A|B=b)=-0.04
B=[0,0,0,0,0,1,1,1,1,1]
A=[0,0,0,0,0,0,1,1,1,1]
P(A|B=b) = 4/5= 0.8 (subsample)](https://image.slidesharecdn.com/ganssummerschool14layout-190223151115/75/GANs-Deep-Learning-Summer-School-55-2048.jpg)
Uploaded byRubens Zimbres, PhD
GANs Deep Learning Summer School
This document discusses generative adversarial networks (GANs) and provides several summaries: 1. GANs use two neural networks, a generator and discriminator, that compete in a game theoretic framework to generate new data instances that match the training data distribution. 2. Training GANs involves training the generator to generate more realistic samples to fool the discriminator while training the discriminator to better distinguish real and generated samples. 3. Several strategies for training GANs are discussed, including varying the update rates of the generator and discriminator, using cooperative or random update strategies, and applying penalties for noisy samples.
More Related Content
![[GAN by Hung-yi Lee]Part 1: General introduction of GAN](https://cdn.slidesharecdn.com/ss_thumbnails/part1-180809095233-thumbnail.jpg?width=640&height=640&fit=bounds)
![[GAN by Hung-yi Lee]Part 2: The application of GAN to speech and text processing](https://cdn.slidesharecdn.com/ss_thumbnails/part2v2-180809095331-thumbnail.jpg?width=640&height=640&fit=bounds)
![[한국어] Safe Multi-Agent Reinforcement Learning for Autonomous Driving](https://cdn.slidesharecdn.com/ss_thumbnails/safemultiagentrlforautonomousedriving4-170508165059-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to GANs Deep Learning Summer School
GANs Deep Learning Summer School
- 1. Generative Adversarial Networks Rubens Zimbres DataScientist Machine Learning for IoT at Vecto Mobile
- 2. Generative Models Discriminative Models Game Theory Prisoner’s Dillemma Nash Equilibrium Rationalchoice Bounded Rationality Non cooperative Game Training GANs Some examples
- 3. + Noise Player 1(D) Player 2 (G)
- 4. Generative Models • Probabilistic,samples from joint distribution P(x,y) • Generate samples (latent variables) following a distribution • Maximum Likelihood • Range of values belong to a class
- 5. Generative Models • LatentDirichlet Allocation (NLP) • Mixture Gaussians • Hidden Markov Model • VAE • RBM • GANs
- 6.
- 7.
- 8. Variational Autoencoders Encoder Decoder Blurred Backprop Kullback–Leiblerdivergence KL (surprise)
- 9.
- 10. Generative Models X1 X2 BuyRecommend Y2 Y1
- 11. Discriminative Models • Observedvariables: target, infer outputs, conditional • Function maps from x to Y • Conditional density function • P(Y|x) subsample • Decision bounday • Computationally expensive
- 12. Discriminative Models • LogisticRegression • SVMs • Neural Networks
- 14.
- 15.
- 16. Generative vs Discriminative •Generative learns joint probability distribution = P(x,y) • Discriminative learns conditional probability distribution = P(y|x)
- 17. GAN Advantages overGenerative Models • D & G = Multi Layer Perceptron • Regular backpropagation like VAE (G and D differentiable) • Does not necessarily involve Maximum Likelihood estimation • No need to use Markov Chains • Subject but robust to overfitting
- 18. Game Theory • 2Player Game – Zero Sum • Minimax solution: Nash Equilibrium • Unique solution: D=1/2 everywhere
- 20. Nash Equilibrium • Suboptimal • Non cooperative • Emulate human behavior: bounded rationality (Herbert Simon) • Simultaneous but not equivalent ** (GANs)
- 21. GANs • Semi SupervisedLearning: missing labels • Inverse Reinforcement Learning
- 23. G D 1 Latent Variables (add noise) Fake TrainingData Real MLP MLP Encoder Decoder Latent Variables (add noise) MLP MLP Real MNIST ADVERSARIAL (0,0) COOPERATIVE (1,1)
- 24. Question • How canwe change GAN’s strategy to create e WIN-WIN situation with payoff equal to (1,1) ?
- 25. G D 1 Latent Variables Add (noise) Fake TrainingData Real MLP MLP ADVERSARIAL (0,0) G COOPERATIVE (1,1) Infiltrated Cop MLP Training Data Real
- 26. GAN Tuning • NeuralNetwork architecture • Hyperparameters • Game Theory strategies
- 28. GANs • Train Simultaneously(no freeze) • Train Discriminator one step • Train Generator k steps (no Nash) • Until pG=pdata (global optimum) • Discriminator cannot distinguish both distributions
- 29. GANs • x comesfrom pdata y=1 (Real) • x comes from pG y=0 (Fake) • Optimize θ rather than pG • G: Relu (vanishing gradient) • Dropout • Supervised Learning in D: • Beginning: high confidence • Usually D is bigger/deeper than G
- 30. Vanishing Gradient Image source:Andrej Karpathy -25% each activation
- 31. Coefficient and Intercept:Regularization
- 32.
- 33. • Generator: GradientDescent on V • Discriminator: Gradient Ascent on V • Minibatch • Batch Normalization in G *** • To minimize Cross Entropy • Minimax Game • Regular cross entropy Training GANs
- 34. Training GANs • Discriminatorminimize J(D)(θD, θG) controlling θD • Generator minimize J(G)(θD, θG) controlling θG • Cannot control each other (G still learns)
- 35. Training Challenges • Nonconvergence: Nash Equilibrium, harder to optimize than objective function • Mode collapse: G fails to output diversity (few good samples) • D converges to right distribution • G generates samples in the most probable point • Solution: k steps for D training for 1 step G training
- 36.
- 37. Tips and Tricksfor GANs • Normalize data (-1,1) • Activation function Tanh output of Generative • Sample from gaussian distribution instead of uniform distribution • Develop different mini-batches in Discriminator: • All real • All fake • Use Batch Normalization
- 38. Tips and Tricksfor GANs • Use DCGANs or VAE+GAN • Adam in Generator (exp. momentum decay) (Adaptive Moment Estimation) • SGD in Discriminator • If Generator loss high = garbage to Discriminator • Dropouts in Generator (overfitting)
- 41.
- 42. Adam Optimizer Momentum w/exponential decay according to derivative of error
- 43.
- 44. GAN Examples: VectorArithmetic
- 45. Vector Arithmetic Radford etal, 2015
- 46.
- 47. DCGAN • Deep ConvolutionalGAN • Batch Norm not in end of Generator and NOT in Discriminator (x covariate shift - Kaggle) • To increase dimension, Upsampling: convo2D.T with stride > 1 • Downsampling: average pooling + conv2D + stride
- 48. Convolutions Convolution2D Convo2DTranspose 3x3 kernel4x4 input stride=1 3x3 kernel 2x2 input stride=1
- 49.
- 50.
- 51.
- 52.
- 53. DCGAN – FaceRecognition
- 54. Info GAN • Basedin Mutual Information and Entropy (uncertainty) Coin toss
- 55. G D Latent Variables (add noise) Fake TrainingData Real MLP MLP Relu B A b b=[1,1,1,1,1,1,1,1,1,1] N=10 B=[0,0,0,0,0,1,1,1,1,1] P(B=b) = 0.5 H(B)=-SumP(B=b).logP(B=b) H(B)=-0.5.log(0.5) H(B)=-0.15 H(A|B=b)=-0.8.log(0.8) H(A|B=b)=-0.04 B=[0,0,0,0,0,1,1,1,1,1] A=[0,0,0,0,0,0,1,1,1,1] P(A|B=b) = 4/5= 0.8 (subsample)
- 56.
- 57.
- 58.
- 59.
- 60. GAN Examples • Nextframe prediction • Image-to-image translation Lotter et al, 2016 Isola et al, 2016
- 61. GAN Examples • Text-to-Imagegeneration Image Inpainting Reed et al, 2016 Pathak et al, 2016
- 62. Cycle GAN: StyleTransfer Zhu et al, 2017
- 66. Different Game Strategies •“Rather than” tuning architecture and hyperparameters • Freeze layers (unofficial implementation) • Discriminator : Generator • 1:1 • 5:1 • 1:5 • Cooperation • Random strategy update • Tit-for-Tat strategy
- 67.
- 68.
- 69. One Pixel Attackto Fool Neural Nets
- 71. Generated Samples 𝒏𝒐𝒊𝒔𝒆 =𝒙 + 𝟎. 𝟏 ∗ 𝒓𝒏𝒅 𝟎, 𝟏 𝒙 = 𝟎 𝝈 = 𝟏
- 72.
- 73. 1:1 Update (VanillaGAN) G D 1 Latent Variables (add noise) Fake Training Data Real MLP MLP Train Simultaneously Accuracy Training Set: .92 Accuracy Test Set: .914 Nash Equilibrium: Sub-optimal Train SimultaneouslyLearning Rate: 0.008 Epochs: 1,000 each Generator optmiz: Adam Discriminator optimiz: SGD Decay Rate: 5e-5 Momentum: 0.9
- 74. 5:1 Update G D 1 Latent Variables (addnoise) Fake Training Data Real MLP MLP Train 1 time Learning Rate: 0.008 Epochs: 1,000 each Generator optmiz: Adam Discriminator optimiz: SGD Decay Rate: 5e-5 Momentum: 0.9 Accuracy Training Set: .91 Accuracy Test Set: .91 Train 5 times
- 75. 1:5 Update G D 1 Latent Variables (addnoise) Fake Training Data Real MLP MLP Train 5 times Learning Rate: 0.008 Epochs: 1,000 each Generator optmiz: Adam Discriminator optimiz: SGD Decay Rate: 5e-5 Momentum: 0.9 Accuracy Training Set: .92 Accuracy Test Set: .91 Train 1 time
- 76. Cooperation G D 1 Latent Variables (add noise) Fake MLP MLP Trainat once Learning Rate: 0.008 Epochs: 1,000 each Generator optmiz: Adam Discriminator optimiz: SGD Decay Rate: 5e-5 Momentum: 0.9 Accuracy Training Set: .99 Accuracy Test Set: .93 Train at once x sigmoid Vanishing gradient Training Data Real
- 77. Random Strategy G D 1 Latent Variables (addnoise) Fake MLP MLP Train at chance Learning Rate: 0.008 Epochs: 1,500 Generator optmiz: Adam Discriminator optimiz: SGD Decay Rate: 5e-5 Momentum: 0.9 Accuracy Test Set: .98 Accuracy Test Set: .96 Polarization of Game until Train at chance 𝑁 = ∞ 𝑥 ≈ 0.5 Training Data Real x sigmoid Vanishing gradient Relu
- 78. Tit-For-Tat Strategy G D 1 Latent Variables (addnoise) Fake MLP MLP Train with penalty if provides noisy samples Learning Rate: 0.008 Epochs: 1,500 Generator optmiz: Adam Discriminator optimiz: SGD Decay Rate: 5e-5 Momentum: 0.9 𝒏𝒐𝒊𝒔𝒆 = 𝒙 + 𝟎. 𝟏 ∗ 𝒓𝒏𝒅 𝟎, 𝟏 𝒙 = 𝟎 𝝈 = 𝟏 Accuracy Test Set: .96 Accuracy Test Set: .92 Train Training Data Real In GAN training: % noisy samples 𝒏𝒐𝒊𝒔𝒆 = 𝒙 + 𝟎. 𝟒 ∗ 𝒓𝒏𝒅 𝟎, 𝟏 𝒙 = 𝟎 𝝈 = 𝟏 Accuracy Test Set: .92 Accuracy Test Set: .91
- 79. Freeze Layers G D 1 Latent Variables (addnoise) Fake Training Data Real MLP MLP Train First Freeze weights Train Second Learning Rate: 0.008 Epochs: 1,000 each Generator optmiz: Adam Discriminator optimiz: SGD Decay Rate: 5e-5 Momentum: 0.9 Accuracy Training Set: .63 Accuracy Test Set: .12 x sigmoid Vanishing gradient Relu * SYNCHRONICITY
- 80.