Ensemble of GANs as a Data Augmentation Technique for Alzheimer research

Data scarcity is a regular problem in research, and in medicine it’s especially difficult to find datasets publicly available. The main reason is its rarity, by definition images of anomalies are scattered and/or not common, and also there are a lot of legal issues that prevent sharing personal information about patients. In this talk, I’m going to give an introduction to ensembles, and more specifically to Generative Adversarial Networks (GANs) ensembles, applied to this problem of data scarcity, generating MRI images of demented brains that could be use in Alzheimer research.

Presented at Pycon Bolivia 2022 - https://bo.pycon.org
Repo: https://github.com/p1nox/gan_ensembles

Transcript

1. Ensemble of GANs as a Data Augmentation Technique for Alzheimer research by Raul Pino PyCon Bolivia 2022
2. Using PyTorch to Generate Demented Brains by Raul Pino PyCon Bolivia 2022
3. Designing and Developing your own Capstone Project by Raul Pino PyCon Bolivia 2022
4. • Alzheimer, motivation, and research. • Capstone project: Idea + value added + metrics. • Bases: ◦ Intro to Deep Learning. ◦ Original DCGAN paper. ◦ Ensembles of GANs paper. • Solution & Demo: ◦ Dataset exploration. ◦ GAN base architecture. ◦ Ensemble techniques implementation. ◦ Results. • Takeaways & beyond. Agenda PyConBo22
5. About me PyConBo22 • Born in Venezuela. • +10 years of exp as Software Engineer & AI enthusiast. • Living in Chile. ◦ Elementus - uBiome - Groupon. • <3 AI, Coffee, Diving, … Empanadas Enthusiast! Salteñas!
6. Alzheimer, motivation & research PyconBo22 • Data scarcity. • Anonymization. • Sampling Bias. • Art & awareness. • Most common case of dementia. • Neurodegenerative, loss of neurons. • Difficult diagnosis. • No cure.
7. Capstone project: idea + value added + metrics PyconBo22 Deep Learning & Machine Learning Engineer programs. • Something related to Alzheimer? Datasets? • Background research? Previous work? • Anything new? Any papers? Any interesting mix? ◦ Ensembles of GANs? #wat? • Any metrics to test solution?
8. Basics: Intro Deep Learning PyconBo22 Neural Network on structured data like a csv.
9. Basics: Intro Deep Learning PyconBo22 Neural Network on structured data like a csv.
10. Basics: Intro Deep Learning PyconBo22 Neural Network on structured data like a csv.
11. Basics: Intro Deep Learning PyconBo22 Unstructured data like an img. (But what is a neural network? | Chapter 1, Deep learning - 3Blue1Brown)
12. Basics: Intro Deep Learning PyconBo22 Supervised Learning
13. Basics: GANs & DCGAN Paper PyconBo22 Unsupervised Learning
14. Basics: GANs & DCGAN Paper PyconBo22 Architecture guidelines for stable Deep Convolutional GANs (DCGANs): • Replace any pooling layers with strided convolutions (discriminator) and fractional-strided convolutions (generator). • Use batchnorm in both the generator and the discriminator. • Remove fully connected hidden layers for deeper architectures. • Use ReLU activation in generator for all layers except for the output, which uses Tanh. • Use LeakyReLU activation in the discriminator for all layers. .
15. Basics: Ensembles of GANs PyconBo22 Ensemble learners - Udacity
16. Suggest three different training techniques: • Standard Ensemble of GANs (eGANs). • Self-ensemble of GANs (seGANs). • Cascade of GANs (cGANs). Basics: Ensembles of GANs PyconBo22
17. Standard Ensemble of GANs (eGANs) Basics: Ensembles of GANs PyconBo22
18. Self-ensemble of GANs (seGANs) Basics: Ensembles of GANs PyconBo22
19. Cascade of GANs (cGANs) Basics: Ensembles of GANs PyconBo22
20. Solution & Demo: Dataset exploration PyconBo22 Go to https://github.com/p1nox/gan_ensembles/blob/master/data_exploration_viz.ipynb
21. Solution & Demo: GAN base architecture PyconBo22 Go to https://github.com/p1nox/gan_ensembles/blob/master/dcgan_control_model.ipynb
22. Solution & Demo: Ensemble techniques implementation PyconBo22 Go to https://github.com/p1nox/gan_ensembles/blob/master/data_exploration_viz.ipynb
23. Solution & Demo: Dataset exploration PyconBo22 Go to https://github.com/p1nox/gan_ensembles/blob/master/data_exploration_viz.ipynb
24. Takeaways & beyond Stable Diffusion 01 Brain Imaging Generation with Latent Diffusion Models 02 PyConBo22
25. PyConBo22 Resources • Ensemble of Generative Adversarial Networks as a Data Augmentation Technique for Alzheimer research https://github.com/p1nox/gan_ensembles/blob/master/capstone_project_docs/proposal.pdf • Hello Future (2022) Generative AI: a new approach to overcome data scarcity. https://hellofuture.orange.com/en/generative-ai-a-new-approach-to-overcome-data-scarcity/ • 3Blue1Brown: “But what is a neural network? | Chapter 1, Deep learning” - https://www.youtube.com/watch?v=aircAruvnKk • Ian Good fellow at Lex Fridman Podcast (2019) Ian Goodfellow: Generative Adversarial Networks (GANs) | Lex Fridman Podcast #19. https://www.youtube.com/watch?v=Z6rxFNMGdn0&t=3013s • Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter (2018) GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium. https://arxiv.org/pdf/1706.08500.pdf • Kaggle (2019) Alzheimer's Dataset ( 4 class of Images) - Images of MRI Segementation. https://www.kaggle.com/datasets/tourist55/alzheimers-dataset-4-class-of-images • Kaggle (2020) I'm Something of a Painter Myself: Use GANs to create art - will you be the next Monet?. https://www.kaggle.com/competitions/gan-getting-started/overview/evaluation • Can Kocagil (2021) DCGAN paper implementation using PyTorch to generate faces. https://github.com/cankocagil/DCGAN • Steven Lang (2021) FID score for PyTorch. https://github.com/mseitzer/pytorch-fid • Udacity Ensembles learners https://www.youtube.com/watch?v=Un9zObFjBH0 • GAN Structure- Google Developers https://developers.google.com/machine-learning/gan/gan_structure
26. Gracias “mri happy brain“ by @p1nox