SlideShare una empresa de Scribd logo
1 de 34
Descargar para leer sin conexión
Generative Adversarial
Networks
Dai-­‐Hai	
  Nguyen
Bioinformatics	
  Center
Kyoto	
  University
28/04/2018 1
Outline
• Overview	
  of	
  generative	
  models
• Variational AutoEncoder (VAE)
• Generative	
  Adversarial	
  Networks	
  (GAN)
• GAN:	
  applications
• Conclusion
28/04/2018 2
Generative Models
28/04/2018 3
Supervised vs. Unsupervised Learning
• Supervised Learning
Data : (x, y)
x: data, y: label
Goal: Learn a function f to map x-> y
Tasks: Classification, Regression, Detection, etc
http://cs231n.stanford.edu/slides/2017/cs231n_2017_lecture13.pdf
28/04/2018 4
Supervised vs. Unsupervised Learning
• Unsupervised Learning
Data : only data x, no label y
Goal: Learn some underlying hidden structures of data
Tasks: clustering, dim reduction, density estimation, etc.
28/04/2018 5
Unsupervised Learning
• Taxonomy tree of unsupervised leanring
Source:	
  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/
28/04/2018 6
Generative models
• Goal:
Given training samples, generate new samples from the
same distribution
Training data ~ 𝑝"#$#(𝑥) Generated samples ~ 𝑝()"*+(𝑥)
In other words, try to learn a model 𝑝()"*+(𝑥) similar to 𝑝"#$#(𝑥)
28/04/2018 7
Generative models
• Maximum Likelihood Estimation (MLE):
Given training samples 𝑥,, 𝑥.,…, 𝑥/, how to learn 𝑝()"*+ 𝑥; 𝜃 from
which training samples are likely to be generated
𝜃∗
= 𝑎𝑟𝑔𝑚𝑎𝑥8 9 log	
   𝑝()"*+(𝑥>; 𝜃)
/
>?,
28/04/2018 8
Unsupervised Learning
• RECAP
Source:	
  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/
28/04/2018 9
Variational Auto-Encoder (VAE)
28/04/2018 10
Variational Autoencoder
• (probabilistic) generative model to generate samples from latent
variable.
• Assumption: training data {𝑥,, 𝑥.,…, 𝑥/} is generated from latent
variable 𝑧
Sample 𝑥~𝑝8(𝑥|𝑧)
Sample z~𝑝(𝑧)
Vary	
  z1
Vary	
  z2
Example:	
  
Samples	
  x	
  are	
  face	
  images
Latent	
  z	
  is	
  2d	
  vector:
Z1:	
  head	
  orientation
Z2	
  :	
  degree	
  of	
  smile
28/04/2018 11
Variational Autoencoder
• How to learn the model?
MLE again !
𝜃∗
= 𝑎𝑟𝑔𝑚𝑎𝑥8 9 log 𝑝8(𝑥>)
/
>?,
Where 𝑝8 𝑥 = ∫ 𝑝8 𝑥 𝑧 𝑝 𝑧 𝑑𝑧E
-> intractable to compute
• Solution: Variational Approximation
28/04/2018 12
Variational Autoencoder
• Variational approximation
log 𝑝8(𝑥) can be written as the following formulation:
Likelihood	
   term	
  to	
  quantifyhow
good	
  	
  the	
  sample	
  is	
  
reconstructed	
  from	
  z.	
  This	
  can	
  
be	
  estimated	
  by	
  a	
  network.
KL	
  divergence	
  term	
  to	
  estimate	
  the	
  
difference	
  between	
  two	
  distribution
This	
  has	
  good	
  form	
  if	
  both	
  of	
  
distributions	
   are	
  Gaussian-­‐>	
  easy	
  to	
  
estimate
This	
  KL	
  divergence	
  term	
  is	
  intractable	
  
because	
  p(z|x)	
  cannot	
  computed.
But	
  it	
  is	
  aways >=	
  0
28/04/2018 13
Variational Autoencoder
• Variational approximation
log 𝑝8(𝑥) can be written as the following formulation:
Tractable	
  lower	
  bound	
  (ELBO)
28/04/2018 14
Variational Autoencoder
• Variational approximation
log 𝑝8(𝑥) can be written as the following formulation:
Tractable	
  lower	
  bound	
  (ELBO) Strategy:
• Maxmizing ELBO instead of intractable logp(x)
• What to be modeled:
1. 𝑝8(𝑥|𝑧) by a network (decoder)
2. 𝑞J 𝑥 𝑥 by another network (encoder)
28/04/2018 15
Variational Autoencoder: model
q training
Make q(z|x)
close to p(z)Minimize
reconstruction
error
28/04/2018 16
Variational Autoencoder: model
q Sampling: use decoder network only and sample z from prior
Sample	
  z	
  from	
  N(0,	
  I)
28/04/2018 17
Variational Autoencoder: generated
samples
28/04/2018 18
Reminder
• Taxonomy tree of unsupervised leanring
Source:	
  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/
28/04/2018 19
Generative Adversarial Network (GAN)
28/04/2018 20
Generative Adversarial Network: Idea
Key	
  points:
q Belongs	
  to	
  “Implicit	
  density”	
  group	
  and
“hot”	
  method	
  in	
  ML	
  by	
  Goodfellow
q Motivated	
  by	
  game	
  theory
q Two	
  players:
1. Generator	
  tries	
  to	
  generate	
  “fake”	
  
samples	
  from	
  its	
  model
2. Discriminator	
  tries	
  to	
  distinguish	
  
“fake”	
  and	
  “real”	
  samples
28/04/2018 21
GAN: Two player game
Model:
q Generator	
  network:	
  try	
  to	
  fool	
  the	
  discriminator	
  by	
  generating	
  “like-­‐‑real”	
  
images
qDiscriminator	
  network:	
  try	
  to	
  distinguish	
  real	
  and	
  fake	
  samples
28/04/2018 22
GAN: Two player game
Objective	
  fucntion:
Loss	
  for	
  real	
  data	
  x Loss	
  for	
  fake	
  data	
  x
How	
  this	
  work
-­‐‑ D	
  tries	
  to	
  maximize	
  the	
  cost	
  such	
  that	
  D(x)	
  close	
  to	
  1	
  (for	
  real	
  x)	
  and	
  D(G(z))	
  
close	
  to	
  0	
  (fake)
-­‐‑ G	
  tries	
  to	
  minimize	
  the	
  cost	
  such	
  that	
  D(G(z))	
  is	
  close	
  to	
  1
(try	
  to	
  make	
  generated	
  samples	
  real-­‐‑looking	
  as	
  much	
  as	
  possible,	
  to	
  fool	
  D)
28/04/2018 23
GAN: Two player game
Objective	
  fucntion:
Loss	
  for	
  real	
  data	
  x Loss	
  for	
  fake	
  data	
  x
How	
  to	
  train:	
  alternative	
  approach
-­‐‑ Fix	
  G,	
  D	
  maximize	
  the	
  cost
-­‐‑ Fix	
  D,	
  G	
  minimize	
  the	
  cost
28/04/2018 24
GAN: density ratio estimation
Density	
  estimation	
  via	
  density	
  ratio	
  estimation:
28/04/2018 25
Generative Adversarial Network: Result
Some	
  generated	
  samples
28/04/2018 26
GAN: Applications
28/04/2018 27
GAN: applications
Image-­‐‑to-­‐‑Image	
  translation
Goal:	
  learn	
  a	
  mapping	
  from	
  
input	
  image-­‐‑>output	
  image
Ref:	
  	
  Image-­‐to-­‐Image	
  Translation	
  with	
  Conditional	
  Adversarial	
  Networks,	
  CVPR201628/04/2018 28
GAN: applications
Image-­‐‑to-­‐‑Image	
  translation
Generator:	
  
• In:	
  noise	
  +	
  input	
  image;	
  Out:	
  sample
Discriminator:	
  
§ In:	
  pairs	
  of	
  in/out	
  images;	
  Out:	
  fake/real
Optimization:
Where
Encourage	
  less	
  blurring
Similar	
  to	
  original	
  GAN
28/04/2018 29
GAN: applications
Image-­‐‑to-­‐‑Image	
  translation
Result:
Ref:	
  	
  Image-­‐to-­‐Image	
  Translation	
  with	
  Conditional	
  Adversarial	
  Networks,	
  CVPR201628/04/2018 30
GAN: applications
Text-­‐‑to-­‐‑Image	
  translation
Goal:	
  learn	
  a	
  mapping	
  from	
  
input	
  text>output	
  image
Ref:	
  	
  Generative	
  Adversarial	
  Text	
  to	
  Image	
  Synthesis,	
  ICML201628/04/2018 31
GAN: applications
Text-­‐‑to-­‐‑Image	
  translation
Generator:	
  
• In:	
  noise	
  +	
  text	
  ;	
  Out:	
  image
Discriminator:	
  
§ In:	
  pairs	
  of	
  text/	
  images;	
  Out:	
  fake/real
Model:
28/04/2018 32
GAN: applications
Text-­‐‑to-­‐‑Image	
  translation
Result:
Ref:	
  	
  Generative	
  Adversarial	
  Text	
  to	
  Image	
  Synthesis,	
  ICML201628/04/2018 33
Conclusion
• Taxonomy tree of unsupervised leanring, AGAIN!
Source:	
  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/
28/04/2018 34

Más contenido relacionado

La actualidad más candente

Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)Gaurav Mittal
 
Introduction to Diffusion Models
Introduction to Diffusion ModelsIntroduction to Diffusion Models
Introduction to Diffusion ModelsSangwoo Mo
 
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...WiMLDSMontreal
 
GANs and Applications
GANs and ApplicationsGANs and Applications
GANs and ApplicationsHoang Nguyen
 
Generative adversarial networks
Generative adversarial networksGenerative adversarial networks
Generative adversarial networksYunjey Choi
 
Finding connections among images using CycleGAN
Finding connections among images using CycleGANFinding connections among images using CycleGAN
Finding connections among images using CycleGANNAVER Engineering
 
Generative adversarial network and its applications to speech signal and natu...
Generative adversarial network and its applications to speech signal and natu...Generative adversarial network and its applications to speech signal and natu...
Generative adversarial network and its applications to speech signal and natu...宏毅 李
 
Generative Adversarial Networks (GAN)
Generative Adversarial Networks (GAN)Generative Adversarial Networks (GAN)
Generative Adversarial Networks (GAN)Manohar Mukku
 
Generative Adversarial Networks
Generative Adversarial NetworksGenerative Adversarial Networks
Generative Adversarial NetworksMustafa Yagmur
 
PR-231: A Simple Framework for Contrastive Learning of Visual Representations
PR-231: A Simple Framework for Contrastive Learning of Visual RepresentationsPR-231: A Simple Framework for Contrastive Learning of Visual Representations
PR-231: A Simple Framework for Contrastive Learning of Visual RepresentationsJinwon Lee
 
Variational Autoencoder
Variational AutoencoderVariational Autoencoder
Variational AutoencoderMark Chang
 
Transfer learning-presentation
Transfer learning-presentationTransfer learning-presentation
Transfer learning-presentationBushra Jbawi
 
Explicit Density Models
Explicit Density ModelsExplicit Density Models
Explicit Density ModelsSangwoo Mo
 
Deep Learning - Convolutional Neural Networks
Deep Learning - Convolutional Neural NetworksDeep Learning - Convolutional Neural Networks
Deep Learning - Convolutional Neural NetworksChristian Perone
 
1시간만에 GAN(Generative Adversarial Network) 완전 정복하기
1시간만에 GAN(Generative Adversarial Network) 완전 정복하기1시간만에 GAN(Generative Adversarial Network) 완전 정복하기
1시간만에 GAN(Generative Adversarial Network) 완전 정복하기NAVER Engineering
 
Introduction To Generative Adversarial Networks GANs
Introduction To Generative Adversarial Networks GANsIntroduction To Generative Adversarial Networks GANs
Introduction To Generative Adversarial Networks GANsHichem Felouat
 

La actualidad más candente (20)

Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)Convolutional Neural Networks (CNN)
Convolutional Neural Networks (CNN)
 
Introduction to Diffusion Models
Introduction to Diffusion ModelsIntroduction to Diffusion Models
Introduction to Diffusion Models
 
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
Fashion-Gen: The Generative Fashion Dataset and Challenge by Negar Rostamzade...
 
GANs and Applications
GANs and ApplicationsGANs and Applications
GANs and Applications
 
Generative adversarial networks
Generative adversarial networksGenerative adversarial networks
Generative adversarial networks
 
Finding connections among images using CycleGAN
Finding connections among images using CycleGANFinding connections among images using CycleGAN
Finding connections among images using CycleGAN
 
Generative adversarial network and its applications to speech signal and natu...
Generative adversarial network and its applications to speech signal and natu...Generative adversarial network and its applications to speech signal and natu...
Generative adversarial network and its applications to speech signal and natu...
 
LeNet-5
LeNet-5LeNet-5
LeNet-5
 
Generative Adversarial Networks (GAN)
Generative Adversarial Networks (GAN)Generative Adversarial Networks (GAN)
Generative Adversarial Networks (GAN)
 
Generative Adversarial Networks
Generative Adversarial NetworksGenerative Adversarial Networks
Generative Adversarial Networks
 
PR-231: A Simple Framework for Contrastive Learning of Visual Representations
PR-231: A Simple Framework for Contrastive Learning of Visual RepresentationsPR-231: A Simple Framework for Contrastive Learning of Visual Representations
PR-231: A Simple Framework for Contrastive Learning of Visual Representations
 
Variational Autoencoder
Variational AutoencoderVariational Autoencoder
Variational Autoencoder
 
Transfer learning-presentation
Transfer learning-presentationTransfer learning-presentation
Transfer learning-presentation
 
Explicit Density Models
Explicit Density ModelsExplicit Density Models
Explicit Density Models
 
Adversarial training Basics
Adversarial training BasicsAdversarial training Basics
Adversarial training Basics
 
Deep Learning - Convolutional Neural Networks
Deep Learning - Convolutional Neural NetworksDeep Learning - Convolutional Neural Networks
Deep Learning - Convolutional Neural Networks
 
Cnn
CnnCnn
Cnn
 
Zero shot learning
Zero shot learning Zero shot learning
Zero shot learning
 
1시간만에 GAN(Generative Adversarial Network) 완전 정복하기
1시간만에 GAN(Generative Adversarial Network) 완전 정복하기1시간만에 GAN(Generative Adversarial Network) 완전 정복하기
1시간만에 GAN(Generative Adversarial Network) 완전 정복하기
 
Introduction To Generative Adversarial Networks GANs
Introduction To Generative Adversarial Networks GANsIntroduction To Generative Adversarial Networks GANs
Introduction To Generative Adversarial Networks GANs
 

Similar a Brief introduction on GAN

Scaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid ParallelismScaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid ParallelismParameswaran Raman
 
AlgorithmsModelsNov13.pptx
AlgorithmsModelsNov13.pptxAlgorithmsModelsNov13.pptx
AlgorithmsModelsNov13.pptxPerumalPitchandi
 
Dimensionality Reduction
Dimensionality ReductionDimensionality Reduction
Dimensionality ReductionSaad Elbeleidy
 
IRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification AlgorithmsIRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification AlgorithmsIRJET Journal
 
IRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification AlgorithmsIRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification AlgorithmsIRJET Journal
 
IRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms ComparisonIRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms ComparisonIRJET Journal
 
IRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms ComparisonIRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms ComparisonIRJET Journal
 
Model-Based Optimization / INFORMS International
Model-Based Optimization / INFORMS InternationalModel-Based Optimization / INFORMS International
Model-Based Optimization / INFORMS InternationalBob Fourer
 
DC02. Interpretation of predictions
DC02. Interpretation of predictionsDC02. Interpretation of predictions
DC02. Interpretation of predictionsAnton Kulesh
 
Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5
Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5
Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5TigerGraph
 
Using Graph Algorithms for Advanced Analytics - Part 5 Classification
Using Graph Algorithms for Advanced Analytics - Part 5 ClassificationUsing Graph Algorithms for Advanced Analytics - Part 5 Classification
Using Graph Algorithms for Advanced Analytics - Part 5 ClassificationTigerGraph
 
Neo4j MeetUp - Graph Exploration with MetaExp
Neo4j MeetUp - Graph Exploration with MetaExpNeo4j MeetUp - Graph Exploration with MetaExp
Neo4j MeetUp - Graph Exploration with MetaExpAdrian Ziegler
 
Neural Learning to Rank
Neural Learning to RankNeural Learning to Rank
Neural Learning to RankBhaskar Mitra
 
DWDM-AG-day-1-2023-SEC A plus Half B--.pdf
DWDM-AG-day-1-2023-SEC A plus Half B--.pdfDWDM-AG-day-1-2023-SEC A plus Half B--.pdf
DWDM-AG-day-1-2023-SEC A plus Half B--.pdfChristinaGayenMondal
 
Software tookits for machine learning and graphical models
Software tookits for machine learning and graphical modelsSoftware tookits for machine learning and graphical models
Software tookits for machine learning and graphical modelsbutest
 
Learning when to give up: theory, practice and perspectives
Learning when to give up: theory, practice and perspectivesLearning when to give up: theory, practice and perspectives
Learning when to give up: theory, practice and perspectivesGiuseppe (Pino) Di Fabbrizio
 
20MEMECH Part 3- Classification.pdf
20MEMECH Part 3- Classification.pdf20MEMECH Part 3- Classification.pdf
20MEMECH Part 3- Classification.pdfMariaKhan905189
 
Active Learning for Multi-relational Data Construction
Active Learning for Multi-relational Data ConstructionActive Learning for Multi-relational Data Construction
Active Learning for Multi-relational Data ConstructionHiroshi Kajino
 

Similar a Brief introduction on GAN (20)

Scaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid ParallelismScaling Multinomial Logistic Regression via Hybrid Parallelism
Scaling Multinomial Logistic Regression via Hybrid Parallelism
 
AlgorithmsModelsNov13.pptx
AlgorithmsModelsNov13.pptxAlgorithmsModelsNov13.pptx
AlgorithmsModelsNov13.pptx
 
Dimensionality Reduction
Dimensionality ReductionDimensionality Reduction
Dimensionality Reduction
 
presentationIDC - 14MAY2015
presentationIDC - 14MAY2015presentationIDC - 14MAY2015
presentationIDC - 14MAY2015
 
IRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification AlgorithmsIRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification Algorithms
 
IRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification AlgorithmsIRJET- Performance Evaluation of Various Classification Algorithms
IRJET- Performance Evaluation of Various Classification Algorithms
 
IRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms ComparisonIRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms Comparison
 
IRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms ComparisonIRJET- Supervised Learning Classification Algorithms Comparison
IRJET- Supervised Learning Classification Algorithms Comparison
 
Model-Based Optimization / INFORMS International
Model-Based Optimization / INFORMS InternationalModel-Based Optimization / INFORMS International
Model-Based Optimization / INFORMS International
 
DC02. Interpretation of predictions
DC02. Interpretation of predictionsDC02. Interpretation of predictions
DC02. Interpretation of predictions
 
Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5
Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5
Graph Gurus Episode 32: Using Graph Algorithms for Advanced Analytics Part 5
 
Using Graph Algorithms for Advanced Analytics - Part 5 Classification
Using Graph Algorithms for Advanced Analytics - Part 5 ClassificationUsing Graph Algorithms for Advanced Analytics - Part 5 Classification
Using Graph Algorithms for Advanced Analytics - Part 5 Classification
 
Neo4j MeetUp - Graph Exploration with MetaExp
Neo4j MeetUp - Graph Exploration with MetaExpNeo4j MeetUp - Graph Exploration with MetaExp
Neo4j MeetUp - Graph Exploration with MetaExp
 
Neural Learning to Rank
Neural Learning to RankNeural Learning to Rank
Neural Learning to Rank
 
DWDM-AG-day-1-2023-SEC A plus Half B--.pdf
DWDM-AG-day-1-2023-SEC A plus Half B--.pdfDWDM-AG-day-1-2023-SEC A plus Half B--.pdf
DWDM-AG-day-1-2023-SEC A plus Half B--.pdf
 
A detailed analysis of the supervised machine Learning Algorithms
A detailed analysis of the supervised machine Learning AlgorithmsA detailed analysis of the supervised machine Learning Algorithms
A detailed analysis of the supervised machine Learning Algorithms
 
Software tookits for machine learning and graphical models
Software tookits for machine learning and graphical modelsSoftware tookits for machine learning and graphical models
Software tookits for machine learning and graphical models
 
Learning when to give up: theory, practice and perspectives
Learning when to give up: theory, practice and perspectivesLearning when to give up: theory, practice and perspectives
Learning when to give up: theory, practice and perspectives
 
20MEMECH Part 3- Classification.pdf
20MEMECH Part 3- Classification.pdf20MEMECH Part 3- Classification.pdf
20MEMECH Part 3- Classification.pdf
 
Active Learning for Multi-relational Data Construction
Active Learning for Multi-relational Data ConstructionActive Learning for Multi-relational Data Construction
Active Learning for Multi-relational Data Construction
 

Más de Dai-Hai Nguyen

Advanced machine learning for metabolite identification
Advanced machine learning for metabolite identificationAdvanced machine learning for metabolite identification
Advanced machine learning for metabolite identificationDai-Hai Nguyen
 
Metrics for generativemodels
Metrics for generativemodelsMetrics for generativemodels
Metrics for generativemodelsDai-Hai Nguyen
 
Hierarchical selection
Hierarchical selectionHierarchical selection
Hierarchical selectionDai-Hai Nguyen
 
Semi-supervised learning model for molecular property prediction
Semi-supervised learning model for molecular property predictionSemi-supervised learning model for molecular property prediction
Semi-supervised learning model for molecular property predictionDai-Hai Nguyen
 

Más de Dai-Hai Nguyen (8)

Advanced machine learning for metabolite identification
Advanced machine learning for metabolite identificationAdvanced machine learning for metabolite identification
Advanced machine learning for metabolite identification
 
Metrics for generativemodels
Metrics for generativemodelsMetrics for generativemodels
Metrics for generativemodels
 
IBSB tutorial
IBSB tutorialIBSB tutorial
IBSB tutorial
 
Hierarchical selection
Hierarchical selectionHierarchical selection
Hierarchical selection
 
Semi-supervised learning model for molecular property prediction
Semi-supervised learning model for molecular property predictionSemi-supervised learning model for molecular property prediction
Semi-supervised learning model for molecular property prediction
 
DL for molecules
DL for moleculesDL for molecules
DL for molecules
 
Seminar
SeminarSeminar
Seminar
 
Collaborative DL
Collaborative DLCollaborative DL
Collaborative DL
 

Último

100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptx100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptxAnupama Kate
 
Halmar dropshipping via API with DroFx
Halmar  dropshipping  via API with DroFxHalmar  dropshipping  via API with DroFx
Halmar dropshipping via API with DroFxolyaivanovalion
 
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...Delhi Call girls
 
VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...
VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...
VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...SUHANI PANDEY
 
Ravak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptxRavak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptxolyaivanovalion
 
Determinants of health, dimensions of health, positive health and spectrum of...
Determinants of health, dimensions of health, positive health and spectrum of...Determinants of health, dimensions of health, positive health and spectrum of...
Determinants of health, dimensions of health, positive health and spectrum of...shambhavirathore45
 
Week-01-2.ppt BBB human Computer interaction
Week-01-2.ppt BBB human Computer interactionWeek-01-2.ppt BBB human Computer interaction
Week-01-2.ppt BBB human Computer interactionfulawalesam
 
Schema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdfSchema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdfLars Albertsson
 
BigBuy dropshipping via API with DroFx.pptx
BigBuy dropshipping via API with DroFx.pptxBigBuy dropshipping via API with DroFx.pptx
BigBuy dropshipping via API with DroFx.pptxolyaivanovalion
 
VidaXL dropshipping via API with DroFx.pptx
VidaXL dropshipping via API with DroFx.pptxVidaXL dropshipping via API with DroFx.pptx
VidaXL dropshipping via API with DroFx.pptxolyaivanovalion
 
Mature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptxMature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptxolyaivanovalion
 
Introduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptxIntroduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptxfirstjob4
 
April 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's AnalysisApril 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's Analysismanisha194592
 
FESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdfFESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdfMarinCaroMartnezBerg
 
Midocean dropshipping via API with DroFx
Midocean dropshipping via API with DroFxMidocean dropshipping via API with DroFx
Midocean dropshipping via API with DroFxolyaivanovalion
 
Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...
Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...
Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...Valters Lauzums
 

Último (20)

100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptx100-Concepts-of-AI by Anupama Kate .pptx
100-Concepts-of-AI by Anupama Kate .pptx
 
Halmar dropshipping via API with DroFx
Halmar  dropshipping  via API with DroFxHalmar  dropshipping  via API with DroFx
Halmar dropshipping via API with DroFx
 
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
Call Girls in Sarai Kale Khan Delhi 💯 Call Us 🔝9205541914 🔝( Delhi) Escorts S...
 
Abortion pills in Doha Qatar (+966572737505 ! Get Cytotec
Abortion pills in Doha Qatar (+966572737505 ! Get CytotecAbortion pills in Doha Qatar (+966572737505 ! Get Cytotec
Abortion pills in Doha Qatar (+966572737505 ! Get Cytotec
 
VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...
VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...
VIP Model Call Girls Hinjewadi ( Pune ) Call ON 8005736733 Starting From 5K t...
 
Ravak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptxRavak dropshipping via API with DroFx.pptx
Ravak dropshipping via API with DroFx.pptx
 
Determinants of health, dimensions of health, positive health and spectrum of...
Determinants of health, dimensions of health, positive health and spectrum of...Determinants of health, dimensions of health, positive health and spectrum of...
Determinants of health, dimensions of health, positive health and spectrum of...
 
Week-01-2.ppt BBB human Computer interaction
Week-01-2.ppt BBB human Computer interactionWeek-01-2.ppt BBB human Computer interaction
Week-01-2.ppt BBB human Computer interaction
 
Schema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdfSchema on read is obsolete. Welcome metaprogramming..pdf
Schema on read is obsolete. Welcome metaprogramming..pdf
 
BigBuy dropshipping via API with DroFx.pptx
BigBuy dropshipping via API with DroFx.pptxBigBuy dropshipping via API with DroFx.pptx
BigBuy dropshipping via API with DroFx.pptx
 
VidaXL dropshipping via API with DroFx.pptx
VidaXL dropshipping via API with DroFx.pptxVidaXL dropshipping via API with DroFx.pptx
VidaXL dropshipping via API with DroFx.pptx
 
Mature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptxMature dropshipping via API with DroFx.pptx
Mature dropshipping via API with DroFx.pptx
 
Call Girls In Shalimar Bagh ( Delhi) 9953330565 Escorts Service
Call Girls In Shalimar Bagh ( Delhi) 9953330565 Escorts ServiceCall Girls In Shalimar Bagh ( Delhi) 9953330565 Escorts Service
Call Girls In Shalimar Bagh ( Delhi) 9953330565 Escorts Service
 
Introduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptxIntroduction-to-Machine-Learning (1).pptx
Introduction-to-Machine-Learning (1).pptx
 
April 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's AnalysisApril 2024 - Crypto Market Report's Analysis
April 2024 - Crypto Market Report's Analysis
 
CHEAP Call Girls in Saket (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Saket (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICECHEAP Call Girls in Saket (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
CHEAP Call Girls in Saket (-DELHI )🔝 9953056974🔝(=)/CALL GIRLS SERVICE
 
FESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdfFESE Capital Markets Fact Sheet 2024 Q1.pdf
FESE Capital Markets Fact Sheet 2024 Q1.pdf
 
Midocean dropshipping via API with DroFx
Midocean dropshipping via API with DroFxMidocean dropshipping via API with DroFx
Midocean dropshipping via API with DroFx
 
Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...
Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...
Digital Advertising Lecture for Advanced Digital & Social Media Strategy at U...
 
(NEHA) Call Girls Katra Call Now 8617697112 Katra Escorts 24x7
(NEHA) Call Girls Katra Call Now 8617697112 Katra Escorts 24x7(NEHA) Call Girls Katra Call Now 8617697112 Katra Escorts 24x7
(NEHA) Call Girls Katra Call Now 8617697112 Katra Escorts 24x7
 

Brief introduction on GAN

  • 2. Outline • Overview  of  generative  models • Variational AutoEncoder (VAE) • Generative  Adversarial  Networks  (GAN) • GAN:  applications • Conclusion 28/04/2018 2
  • 4. Supervised vs. Unsupervised Learning • Supervised Learning Data : (x, y) x: data, y: label Goal: Learn a function f to map x-> y Tasks: Classification, Regression, Detection, etc http://cs231n.stanford.edu/slides/2017/cs231n_2017_lecture13.pdf 28/04/2018 4
  • 5. Supervised vs. Unsupervised Learning • Unsupervised Learning Data : only data x, no label y Goal: Learn some underlying hidden structures of data Tasks: clustering, dim reduction, density estimation, etc. 28/04/2018 5
  • 6. Unsupervised Learning • Taxonomy tree of unsupervised leanring Source:  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/ 28/04/2018 6
  • 7. Generative models • Goal: Given training samples, generate new samples from the same distribution Training data ~ 𝑝"#$#(𝑥) Generated samples ~ 𝑝()"*+(𝑥) In other words, try to learn a model 𝑝()"*+(𝑥) similar to 𝑝"#$#(𝑥) 28/04/2018 7
  • 8. Generative models • Maximum Likelihood Estimation (MLE): Given training samples 𝑥,, 𝑥.,…, 𝑥/, how to learn 𝑝()"*+ 𝑥; 𝜃 from which training samples are likely to be generated 𝜃∗ = 𝑎𝑟𝑔𝑚𝑎𝑥8 9 log   𝑝()"*+(𝑥>; 𝜃) / >?, 28/04/2018 8
  • 9. Unsupervised Learning • RECAP Source:  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/ 28/04/2018 9
  • 11. Variational Autoencoder • (probabilistic) generative model to generate samples from latent variable. • Assumption: training data {𝑥,, 𝑥.,…, 𝑥/} is generated from latent variable 𝑧 Sample 𝑥~𝑝8(𝑥|𝑧) Sample z~𝑝(𝑧) Vary  z1 Vary  z2 Example:   Samples  x  are  face  images Latent  z  is  2d  vector: Z1:  head  orientation Z2  :  degree  of  smile 28/04/2018 11
  • 12. Variational Autoencoder • How to learn the model? MLE again ! 𝜃∗ = 𝑎𝑟𝑔𝑚𝑎𝑥8 9 log 𝑝8(𝑥>) / >?, Where 𝑝8 𝑥 = ∫ 𝑝8 𝑥 𝑧 𝑝 𝑧 𝑑𝑧E -> intractable to compute • Solution: Variational Approximation 28/04/2018 12
  • 13. Variational Autoencoder • Variational approximation log 𝑝8(𝑥) can be written as the following formulation: Likelihood   term  to  quantifyhow good    the  sample  is   reconstructed  from  z.  This  can   be  estimated  by  a  network. KL  divergence  term  to  estimate  the   difference  between  two  distribution This  has  good  form  if  both  of   distributions   are  Gaussian-­‐>  easy  to   estimate This  KL  divergence  term  is  intractable   because  p(z|x)  cannot  computed. But  it  is  aways >=  0 28/04/2018 13
  • 14. Variational Autoencoder • Variational approximation log 𝑝8(𝑥) can be written as the following formulation: Tractable  lower  bound  (ELBO) 28/04/2018 14
  • 15. Variational Autoencoder • Variational approximation log 𝑝8(𝑥) can be written as the following formulation: Tractable  lower  bound  (ELBO) Strategy: • Maxmizing ELBO instead of intractable logp(x) • What to be modeled: 1. 𝑝8(𝑥|𝑧) by a network (decoder) 2. 𝑞J 𝑥 𝑥 by another network (encoder) 28/04/2018 15
  • 16. Variational Autoencoder: model q training Make q(z|x) close to p(z)Minimize reconstruction error 28/04/2018 16
  • 17. Variational Autoencoder: model q Sampling: use decoder network only and sample z from prior Sample  z  from  N(0,  I) 28/04/2018 17
  • 19. Reminder • Taxonomy tree of unsupervised leanring Source:  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/ 28/04/2018 19
  • 20. Generative Adversarial Network (GAN) 28/04/2018 20
  • 21. Generative Adversarial Network: Idea Key  points: q Belongs  to  “Implicit  density”  group  and “hot”  method  in  ML  by  Goodfellow q Motivated  by  game  theory q Two  players: 1. Generator  tries  to  generate  “fake”   samples  from  its  model 2. Discriminator  tries  to  distinguish   “fake”  and  “real”  samples 28/04/2018 21
  • 22. GAN: Two player game Model: q Generator  network:  try  to  fool  the  discriminator  by  generating  “like-­‐‑real”   images qDiscriminator  network:  try  to  distinguish  real  and  fake  samples 28/04/2018 22
  • 23. GAN: Two player game Objective  fucntion: Loss  for  real  data  x Loss  for  fake  data  x How  this  work -­‐‑ D  tries  to  maximize  the  cost  such  that  D(x)  close  to  1  (for  real  x)  and  D(G(z))   close  to  0  (fake) -­‐‑ G  tries  to  minimize  the  cost  such  that  D(G(z))  is  close  to  1 (try  to  make  generated  samples  real-­‐‑looking  as  much  as  possible,  to  fool  D) 28/04/2018 23
  • 24. GAN: Two player game Objective  fucntion: Loss  for  real  data  x Loss  for  fake  data  x How  to  train:  alternative  approach -­‐‑ Fix  G,  D  maximize  the  cost -­‐‑ Fix  D,  G  minimize  the  cost 28/04/2018 24
  • 25. GAN: density ratio estimation Density  estimation  via  density  ratio  estimation: 28/04/2018 25
  • 26. Generative Adversarial Network: Result Some  generated  samples 28/04/2018 26
  • 28. GAN: applications Image-­‐‑to-­‐‑Image  translation Goal:  learn  a  mapping  from   input  image-­‐‑>output  image Ref:    Image-­‐to-­‐Image  Translation  with  Conditional  Adversarial  Networks,  CVPR201628/04/2018 28
  • 29. GAN: applications Image-­‐‑to-­‐‑Image  translation Generator:   • In:  noise  +  input  image;  Out:  sample Discriminator:   § In:  pairs  of  in/out  images;  Out:  fake/real Optimization: Where Encourage  less  blurring Similar  to  original  GAN 28/04/2018 29
  • 30. GAN: applications Image-­‐‑to-­‐‑Image  translation Result: Ref:    Image-­‐to-­‐Image  Translation  with  Conditional  Adversarial  Networks,  CVPR201628/04/2018 30
  • 31. GAN: applications Text-­‐‑to-­‐‑Image  translation Goal:  learn  a  mapping  from   input  text>output  image Ref:    Generative  Adversarial  Text  to  Image  Synthesis,  ICML201628/04/2018 31
  • 32. GAN: applications Text-­‐‑to-­‐‑Image  translation Generator:   • In:  noise  +  text  ;  Out:  image Discriminator:   § In:  pairs  of  text/  images;  Out:  fake/real Model: 28/04/2018 32
  • 33. GAN: applications Text-­‐‑to-­‐‑Image  translation Result: Ref:    Generative  Adversarial  Text  to  Image  Synthesis,  ICML201628/04/2018 33
  • 34. Conclusion • Taxonomy tree of unsupervised leanring, AGAIN! Source:  https://allenlu2007.wordpress.com/2018/01/10/variational-autoencoder-的原理/ 28/04/2018 34