AI Nanodegree Program Syllabus: Term 1, In Depth
Here at Udacity, we are about to kick off the first-ever term of our Artificial Intelligence Nanodegree, and we couldn’t be more excited!
I am personally excited that we’re providing a depth of education online that is commensurate with a university education, but making it available to so many more people across the globe, and at a cost that makes an AI education realistic for all aspiring learners.
You can apply to be a part of our AI Nanodegree program here! Our AI Nanodegree program is a 6-month project-based program covering both classical and modern techniques in Artificial Intelligence.
So! Here is our first 3-month curriculum in depth, including all the projects that you will build:
Term 1
Introduction
Meet the instructors: Sebastian Thrun, Peter Norvig, and Thad Starner. These are some of the biggest names in the field of AI. Sebastian is the creator of Stanley, one of the first Self-Driving Cars. Peter is the author of the most authoritative AI textbook in the field, Artificial Intelligence: A Modern Approach. Finally, Thad Starner is a Professor at Georgia Tech and a researcher at Google X.
Learn about what we’ll be covering in this program, and what you get from a Udacity Nanodegree program.
Project: Build an AI to solve Sudoku
You’ll build a simple AI using Constraint Propagation and Search to solve Sudoku puzzles. You’ll extend this to solve Diagonal Sudokus and implement advanced Sudoku strategies such as the Naked Twins strategy.
Deterministic AIs
Game Playing
Game Playing is about teaching AI Agents to win at adversarial games ranging from Chess to Starcraft. In many ways, Game Playing is one of the hallmarks of intelligence, especially in extremely complex games such as Go. This field has long been one of the most exciting areas of exploration for AI, with landmark AIs such as IBM’s DeepBlue, and Google’s latest AlphaGo. You’ll learn about Minimax game trees and how to use alpha-beta pruning to dramatically improve results for your agent. In addition, you’ll apply iterative deepening to understand efficient ways to expand the game tree. This will be in the context of the game Isolation.
Project: Build a Game-Playing AI Agent to win the board game Isolation
You’ll apply what you just learned about game-playing strategies to build an AI to play Isolation. We’ll have your AI compete against other AIs to see how well you’ve performed!
Search
We’ll cover Depth First Search, Breadth First Search, A* Search, and how to analyze heuristics.
Project: Build a Pac-Man AI that finds the most efficient path through its world
We’ll be using the wonderful AI projects created by the AI Department of UC Berkeley to teach Pac-Man to navigate his world, and eat the most food in the shortest time. You’ll write your own implementations of BFS, DFS,and A* to complete this project.
Simulated Annealing
You’ll learn about how to explore large state spaces using the biologically-inspired techniques of Simulated Annealing.
Constraint Satisfaction
You’ll return to the techniques you used to solve Sudoku and understand how constraint satisfaction can be used to solve puzzles such as the map-coloring problem.
Logic and Planning
You will learn how to build systems that can arrive at new, logical conclusions from a given set of facts. In particular, you’ll explore First Order Logic, Propositional Logic, and how to use such logic to solve planning problems. This will all be taught by Peter Norvig.
Project: Cargo Route Planning
You’ll combine your knowledge of Logic, Planning and Search to implement a system that efficiently moves cargo from their origins to their destinations using the least number of flights. The system will use propositional logic to find a path to its goals given its start state and valid actions.
Probabilistic AI
Bayes Nets
First, you’ll learn how to use Probabilistic Inference to calculate the probability of certain events occurring. We’ll cover Bayesian Networks, Conditional Probability, and Bayes’ Rule. This section will be taught by Sebastian Thrun.
Hidden Markov Models
You’ll extend your knowledge of Bayesian Networks to cover Hidden Markov Models where intermediate states can be unobserved. Such models are often used in temporal pattern recognition tasks such as speech, and handwriting recognition.
Project: Translate Sign Language to Text
In the final project, you’ll use HMMs to translate Sign Language into their English Language characters. Your AI will just use images of people signing to automatically translate those into English! You’ll train your model on an available dataset of sign language and use it to classify new images of signing. Advanced students will be able to translate full sentences as well.
Feel free to reach out to us if you have any questions, and see you in the program!
Term 2: Sneak Peek!
In the second term of the program, you’ll explore the cutting edge advancements of AI - Deep Learning. You’ll learn the foundations of neural networks, understand gradient descent and backpropagation, and learn to make architecture choices. We’ll cover Convolutional Networks, Reinforcement Learning, and Recurrent Neural Networks. You’ll also choose a concentration in either Speech, Computer Vision, or Natural Language Processing. Here are our concentration partners:
Concentration 1: Speech with Amazon Alexa
Concentration 2: Natural Language Processing with IBM Watson
Concentration 3: Computer Vision with Affectiva
Stay tuned for an in-depth post on Term 2 curriculum, coming soon!
With this curriculum, we believe students will receive an exciting, hands-on exposure to both the foundations—and the cutting-edge advancements—of AI that are rapidly changing technology today. In addition, you’ll get to see how engineers in AI labs such as IBM Watson and Amazon Alexa actually use AI in practice, giving you the chance to go beyond theory and into application. Through our AI program, we want to provide you the best entry possible into the space of AI, online or offline. Join us!
