SLIDE 1
Computational Systems Biology Deep Learning in the Life Sciences
6.802 6.874 20.390 20.490 HST.506
David Gifford Lecture 9 March 5, 2020
Generative Models
http://mit6874.github.io
1
Computational Systems Biology Deep Learning in the Life Sciences - - PowerPoint PPT Presentation
Computational Systems Biology Deep Learning in the Life Sciences - - PowerPoint PPT Presentation
Computational Systems Biology Deep Learning in the Life Sciences 6.802 6.874 20.390 20.490 HST.506 David Gifford Lecture 9 March 5, 2020 Generative Models http://mit6874.github.io 1 Why generative models? We can sample new examples from a
SLIDE 2
SLIDE 3
We can sample new examples from a generative models
- Generate new examples from model fit to
training data
- Sampling from input distribution
- Optionally optimized with respect to a metric
- Reveals what models understand
- What is the best example of a written digit?
- What is the best example of a celebrity?
- Transform examples with respect to one or more
metrics
- Improve sentiment of text
- Perform multi-objective optimization of antibodies
SLIDE 4
https://www.cs.toronto.edu/~rgrosse/courses/csc321_2018/slides/lec19.pdf
SLIDE 5
SLIDE 6
SLIDE 7
SLIDE 8
Three example generative models
- Variational autoencoders
- Generative Adversarial Networks
- CycleGANs
- For each you should understand the loss function
SLIDE 9
Variational Autoencoders can provide improved examples
SLIDE 10
SLIDE 11
SLIDE 12
SLIDE 13
SLIDE 14
Why is this important? Why does it make the task difficult?
SLIDE 15
Why is this important? Find plausible revisions Why does it make the task difficult? p(z | x) intractable
SLIDE 16
SLIDE 17
SLIDE 18
SLIDE 19
SLIDE 20
SLIDE 21
Overall VAE loss function
SLIDE 22
SLIDE 23
SLIDE 24
SLIDE 25
SLIDE 26
SLIDE 27
SLIDE 28
SLIDE 29
SLIDE 30
Generative Adversarial Networks
SLIDE 31
We wish to learn a generative model that matches the true data distribution
SLIDE 32
https://www.cs.toronto.edu/~rgrosse/courses/csc321_2018/slides/lec19.pdf
The Generative Adversarial Network (GAN) Game
SLIDE 33
D(x) is probability x is from the real-world
SLIDE 34
SLIDE 35
Update discriminator to maximize D(real) and minimize D(G(z))
SLIDE 36
Update generator to maximize D(G(z))
SLIDE 37
Summary of GAN objective functions
SLIDE 38
GANs have become a bit of a fad
SLIDE 39
SLIDE 40
SLIDE 41
SLIDE 42
GANs can fail
SLIDE 43
GAN Problems
- Non-convergence – model parameters oscillate and never
converge
- Mode collapse – limited variety of samples from generator
- Diminished gradient – Discriminator is too successful and
generator learns nothing
- Overfitting – imbalance between generator and discriminator
- Hyperparameter sensitivity – highly sensitive
https://medium.com/@jonathan_hui/gan-why-it-is-so-hard-to-train-generative-advisory-networks-819a86b3750b
SLIDE 44
Mode collapse shown in second row (all 6) and in images
https://arxiv.org/pdf/1611.02163.pdf https://arxiv.org/pdf/1703.10717.pdf
SLIDE 45
CycleGANs for style mapping
SLIDE 46
CycleGANs map between styles
SLIDE 47
CycleGANs permit style transfer without matched training data
SLIDE 48
CycleGANs permit style transfer without matched training data
https://arxiv.org/pdf/1703.10593.pdf
SLIDE 49
CycleGANs permit style transfer without matched training data
SLIDE 50
SLIDE 51
CycleGANs permit style transfer without matched training data
SLIDE 52