The Neural Aesthetic @ ITP-NYU, Fall 2019
Lectures Syllabus
03 Sep 2019
The whole class "in 60 minutes"
Description
- Course goals, logistics, and resources
- Introduction to AI, machine learning, and deep learning
- The whole class "in 60 minutes"
- Getting setup with Runway and ml5.js
Practical
Supplementary / links
Description
- How convolutional neural networks work (forward pass)
- Transfer learning
- Demos and applications of transfer learning
Supplementary / links
Description
- How neural nets are trained (backward pass)
- Overfitting, regularization, optimization
- ml4a-ofx demos: ConvnetPredictor, AudioClassifier, DoodleClassifier
Practical
Supplementary / links
Description
- Feature extraction
- Reverse image search and X degrees
- t-SNE of images/sounds & visualization
- Transfer learning
Supplementary / links
27 Sep 2019
Terminal velocity
Description
- Basic terminal navigation
- Python shell and bash scripting
- Connecting to remote VM
Practical
Supplementary / links
Description
- Visualizing convnet features
- Pixel-optimization, Deepdream, neural-synth
- Style transfer and textur synthesis
Practical
Supplementary / links
Description
- Principal component analysis
- Paperspace tutorial
- neural-style
Practical
Supplementary / links
Description
- Autoencoders & Generative adversarial networks
- Compiling, scraping, and processing datasets
Practical
Supplementary / links
Description
- Review of generative models and GANs
- BigGAN tutorial
- Scraping and preparing datasets
Supplementary / links
Description
- Image-to-image translation (pix2pix/CycleGAN)
- Tutorial on StyleGAN and DCGAN
Supplementary / links
Description
- Recurrent neural networks
- Natural language processing
- Music information retrieval
Description
- Agent-environment systems & games
- GLOW tutorial
- Abraham and autonomous artificial artists
Practical
Supplementary / links
10 Dec 2019
Final presentations
Description
- tbd
Practical
Supplementary / links
