The Nature of Code Part 2 (Spring 2017)

Syllabus for Part 2 of Nature of Code: "Intelligence and Learning" at ITP Spring 2017

This syllabus is very much in progress. I'm drawing inspiration from this Coding Train community list of resources.

Class Info

• Daniel Shiffman: daniel.shiffman@nyu.edu
• Section 1: Tuesdays, 9:00-11:30am
• Section 2: Wednesdays, 9:00-11:30am
• Office Hours
• Sign up for Mailing List
• Resources and References

Prerequisites?

• Basic familiary with p5.js and Processing.
• Fundamentals of computer programming, equivalent to ICM
• While taking Nature of Code Part 1 is not required, I recommend you familiarize yourself with the following chapters before the first day of class.
  • Chapter 1: Vectors
  • Chapter 2: Forces
  • Chapter 3: Oscillation and Trig
  • Chapter 6: Steering Behaviors
• I assume no knowledge whatsoever about AI, machine learning, deep learning and the accompanying math required for the algorithms listed below. After all, I barely know this stuff myself.

Week 1 - Introduction (March 21/22)

• Week 1 Notes
• Week 1 Homework
• Class Intro / Overview
• Algorithms
  • Big O notation
  • Graphs
    • Binary Tree
    • Breadth-first Search
    • Dijktra's Algorithm
    • A* search
    • Traveling Salesperson
    • plus steering agents!

Week 2 - Genetic Algorithms (March 28/29)

• Week 2 Notes
• Week 2 Homework
• Search
• Evolutionary Design
• Evolutionary Ecosystem

Week 3 - Classification and Regression (April 4/5)

• Week 3 Notes
• Week 3 Homework
• What is Machine Learning
• What is Supervised Learning
• Classification and Regression
• KNN
• Linear Regression and Gradient Descent

Week 4 - Neural Networks (April 11/12)

• Week 4 Notes
• Perceptron
• Multi-Layered Perceptron
• inputs and outputs
• Backpropogation
• Training vs. Testing (MNIST data set)
• What is "Deep Learning"?

Week 5 - Adding Tensorflow: Convolutional Neural Networks (April 18/19)

• Environment Setup
• Week 5 Notes
• Assignment: Project Step 1
• Overview of libraries and frameworks for Deep Learning
• Convolutional Neural Networks for Image Classification (and more)
• Keras and Tensorflow
• Python and Flask
• Flask and p5.js

Week 6 - Recurrent Neural Networks, NeuroEvolution/Reinforcement Learning (April 25/26)

• Week 6 RNN Notes
• Week 6 Bonus NeuroEvolution Notes
• Recurrent Neural Networks for Sequences (text generation)
• Overview of Reinforcement Learning
• Neuro Evolution (evolving ANN weights)

Week 7 - Project Presentations (May 2/3)

• Project Presentations + Documentation

Policies

• Submit assignments by the evening before class to the extent possible.
• Come prepared with questions.
• Put away screens during others' presentations.
• Participate!
• Document!
• Grading:
  • 40% Class Participation
  • 40% Quality of assignments
  • 20% Final project
• For a 2-point class, 2 or more unexcused absences is grounds for failure.