INDIVIDUAL LEARNERS

SCHOOL OF ARTIFICIAL INTELLIGENCE

Artificial Intelligence
Nanodegree Program Syllabus
Overview
This program will teach learners how to become a better artificial intelligence or machine learning engineer by teaching them
classical AI algorithms applied to common problem types. They will complete projects and exercises incorporating search,
optimization, planning, and probabilistic graphical models which have been used in artificial intelligence applications for
automation, logistics, operations research, and more. These concepts form the foundation for many of the most exciting
advances in AI in recent years. Each project you build will be an opportunity to demonstrate what students have learned in
their lessons and become part of a career portfolio that will demonstrate their mastery of these skills to potential employers.

Program information

Estimated Time Skill Level

3 months at 12-15hrs/week* Advanced

Prerequisites

A well-prepared learner should have knowledge of algebra, calculus, statistics, and Python.

Required Hardware/Software

Learners will need a computer running a 64-bit operating system (most modern Windows, OS X, and Linux versions will work)
with at least 8GB of RAM, along with administrator account permissions sufficient to install programs including Anaconda
with Python 3.5 and supporting packages. Their network should allow secure connections to remote hosts (like SSH). We will
provide learners with instructions to install the required software packages. Udacity does not provide any hardware.

*The length of this program is an estimation of total hours the average student may take to complete all required
coursework, including lecture and project time. If you spend about 5-10 hours per week working through the program, you
should finish within the time provided. Actual hours may vary.

Artificial Intelligence 2
 Course 1

Introduction to Artificial Intelligence

In this course, students will learn about the foundations of AI. They’ll configure their programming environment to work on AI
problems with Python. At the end of the course, they’ll build a Sudoku solver and solve constraint satisfaction problems.

Course Project

Build a Sudoku Solver

Humans use reason to solve problems by decomposing the problem statement and incorporating domain
knowledge to limit the possible solution space. Use a technique called constraint propagation together
with backtracking search to make an agent that only considers reasonable solution candidates and
efficiently solves any Sudoku puzzle. This approach appears in many classical AI problems, and the solution
techniques have been extended and applied to diverse problems in bioinformatics, logistics, and operations
research. Demonstrate some basic algorithms knowledge and learn to use constraint satisfaction to solve
general problems.

Lesson 1

• Learn about the resources available to help one succeed.

Welcome to Artificial
Intelligence

• Consider the meaning of “artificial intelligence.”

Lesson 2 • Be able to define core concepts from AI including “agents,” “environments,” and
“states.”
Intro to Artificial Intelligence
• Learn the concept of “rational” behavior for AI agents.

Artificial Intelligence 3
Lesson 3 • Express logical constraints as Python functions.

Solving Sudoku with AI • Use constraint propagation and search to solve all Sudoku puzzles.

Lesson 4
• Use Conda to configure and manage Python packages and dependencies.
Setting Up Your Environment

• Learn to represent problems in terms of logical constraints.

Lesson 5 • Use constraint propagation to limit the potential solution space.

Constraint Satisfaction • Incorporate backtracking search to find a solution when the set of constraints

Problems is incomplete.

• Use Z3 Solver to solve constraint satisfaction problems.

Course 2

Classical Search
In this course students will learn classical graph search algorithms—including uninformed search techniques like breadth-first
and depth-first search and informed search with heuristics including A*. These algorithms are at the heart of many classical
AI techniques, and have been used for planning, optimization, problem solving, and more. Complete the lesson by teaching
PacMan to search with these techniques to solve increasingly complex domains.

Artificial Intelligence 4
Lesson 1
• Learn about the significance of search in AI
Introduction

Lesson 2 • Learn uninformed search techniques including Depth-First, Search,

Breadth-First Search, and Uniform Cost Search.
Uninformed Search

Lesson 3 • Learn informed search techniques (using heuristics) including A* Search.

Informed Search • Understand admissibility and consistency conditions for heuristics.

Lesson 4
• Implement informed and uninformed search for Pacman.
Classroom Exercise: Search

Lesson 5
• List of external resources for one to continue learning about search.
Additional Topics: Search

Artificial Intelligence 5
 Course 3

Automated Planning
In this course students will learn to represent general problem domains with symbolic logic and use search to find optimal
plans for achieving the agent’s goals. Planning and scheduling systems power modern automation and logistics operations, as
well as aerospace applications like the Hubble telescope & NASA Mars rovers.

Course Project

Build a Forward Planning Agent

Intelligent agents are expected to act in complex domains where their goals and objectives may not be
immediately achievable. They must reason about their goals and make rational choices of actions to achieve
them. Build a system using symbolic logic to represent general problem domains and use classical search to
find optimal plans for achieving the agent’s goals. Demonstrate an understanding of classical optimization
and search algorithms, symbolic logic, and domain independent planning.

• Learn propositional logic (propositions and statements).

Lesson 1
• Learn first-order logic (quantifiers, variables, and objects).
Symbolic Logic & Reasoning
• Encode problems with symbolic constraints using First Order Logic.

Lesson 2

• Learn to define planning problems.

Introduction to
Automated Planning

Artificial Intelligence 6
 • Learn high-level features of automated planning techniques using search
Lesson 3 and symbolic logic including forward planning, backwards planning, and
hierarchical planning.
Classical Planning
• Explore planning heuristics and planning graphs.

Lesson 4
• List of external resources for you to continue learning about search.
Additional Topics: Search

Course 4

Optimization Problems
In this course students will learn about iterative improvement optimization problems and classical algorithms emphasizing
gradient-free methods for solving them. These techniques can often be used on intractable problems to find solutions that are
“good enough” for practical purposes and have been used extensively in fields like operations research and logistics. Students
will finish the lesson by completing a classroom exercise comparing the different algorithms’ performance on a variety of
problems.

Lesson 1 • Introduce iterative improvement problems that can be solved with

optimization.
Introduction

Lesson 2
• Learn random hill climbing for local search optimization problems.
Hill Climbing

Artificial Intelligence 7
Lesson 3
• Learn to use simulated annealing for global optimization problems.
Simulated Annealing

Lesson 4 • Explore and implement genetic algorithms that keep a pool of candidates to
solve optimization problems.
Genetic Algorithms

Lesson 5

• Compare optimization techniques on a variety of problems.

Classroom Exercise:
Optimization Problems

Lesson 6
• Learn about improvements and optimizations to optimization search including
Additional Optimization late acceptance hill climbing, basin hopping, and differential evolution.
Topics

Course 5

Adversarial Search
In this course students will learn how to search in multi-agent environments (including decision making in competitive
environments) using the minimax theorem from game theory. Then build an agent that can play games better than any
human.

Artificial Intelligence 8
 Course Project

Build an Adversarial Game Playing Agent

AI agents acting in the real world have to “hope for the best, but prepare for the worst.” Write an agent that
can play games with superhuman performance in an adversarial environment. The principles of adversarial
search provide a foundation for autonomous agents acting in the real world, and for understanding modern
advances in AI like DeepMind’s AlphaGo Zero.

• Understand “adversarial” problems and applications (e.g., multi-agent

Lesson 1 environments).

• Extend state space search techniques to domains the agents do not fully
Search in Multi-Agent
control.
Domains
• Learn the minimax search technique.

Lesson 2 • Apply depth-limiting to overcome limitations in basic minimax search.

Optimizing Minimax Search • Apply alpha-beta pruning to overcome limitations in basic minimax search.

Lesson 3 • Extend adversarial search to non-deterministic domains and domains with

more than two players.
Extending Minimax Search

• Understand other adversarial search techniques such as Monte Carlo Tree

Lesson 4
Search.
Additional Adversarial • List of external resources for one to continue learning about adversarial
Search Topics search.

Artificial Intelligence 9
 Course 6

Fundamentals of Probabilistic Graphical Models

In this course students will learn to use Bayes Nets to represent complex probability distributions and algorithms for sampling
from those distributions. Then learn the algorithms used to train, predict, and evaluate Hidden Markov Models for pattern
recognition. HMMs have been used for gesture recognition in computer vision, gene sequence identification in bioinformatics,
speech generation and part of speech tagging in natural language processing, and more.

Course Project

Part of Speech Tagging

Probabilistic models allow agents to better handle the uncertainty of the real world by explicitly modeling
their belief state as a distribution over all possible states. Use a Hidden Markov Model (HMM) to perform
part of speech tagging, a common pre-processing step in natural language processing. HMMs have been
used extensively in NLP, speech recognition, bioinformatics, and computer vision tasks.

Lesson 1
• Model probability distributions based on a given set of parameters in a
Introduction to real-world use case using discrete distributions.
Probabilistic Models

Lesson 2 • Review key concepts in probability including discrete distributions, joint

probabilities, and conditional probabilities.
Probability

Artificial Intelligence 10
Lesson 3
• Efficiently encode joint probabilities in Bayes networks.
Bayes Nets

• Learn about inference in Bayes networks through exact enumeration with

Lesson 4
optimizations.
Inference in Bayes Nets • Learn techniques for approximate inference in more complex Bayes networks.

• Learn parameters to maximize the likelihood of model parameters to training

data.
Lesson 5
• Determine the likelihood of observing test data given a fixed model.
Hidden Markov Models
• Learn an algorithm to Identify the most likely sequence of states in a model
given some data.

Lesson 6
• Learn the dynamic time warping algorithm for time-series analysis.
Dynamic Time Warping

Artificial Intelligence 11
Meet your instructors.

Peter Norvig
Research Director at Google

Peter Norvig is a director of research at Google and is co-author of Artificial Intelligence: A

Modern Approach, the leading textbook in the field.

Sebastian Thrun
Founder and Executive Chairman at Udacity

As the founder and president of Udacity, Sebastian’s mission is to democratize education. He is

also the founder of Google X, where he led projects including the self-driving car, Google Glass,
and more.

Thad Starner
Professor of Computer Science at Georgia Tech

Thad Starner is the director of the Contextual Computing Group (CCG) at Georgia Tech and is also
the longest-serving technical lead/manager on Google’s Glass project.

Artificial Intelligence 12
Udacity’s learning
experience

Hands-on Projects Quizzes

Open-ended, experiential projects are designed Auto-graded quizzes strengthen comprehension.
to reflect actual workplace challenges. They aren’t Learners can return to lessons at any time during
just multiple choice questions or step-by-step the course to refresh concepts.
guides, but instead require critical thinking.

Knowledge Custom Study Plans

Find answers to your questions with Knowledge, Create a personalized study plan that fits your
our proprietary wiki. Search questions asked by individual needs. Utilize this plan to keep track of
other students, connect with technical mentors, movement toward your overall goal.
and discover how to solve the challenges that
you encounter.

Workspaces Progress Tracker

See your code in action. Check the output and Take advantage of milestone reminders to stay
quality of your code by running it on interactive on schedule and complete your program.
workspaces that are integrated into the platform.

Artificial Intelligence 13
Our proven approach for building
job-ready digital skills.
Experienced Project Reviewers

Verify skills mastery.

• Personalized project feedback and critique includes line-by-line code review from
skilled practitioners with an average turnaround time of 1.1 hours.

• Project review cycle creates a feedback loop with multiple opportunities for
improvement—until the concept is mastered.

• Project reviewers leverage industry best practices and provide pro tips.

Technical Mentor Support

24/7 support unblocks learning.

• Learning accelerates as skilled mentors identify areas of achievement and potential
for growth.

• Unlimited access to mentors means help arrives when it’s needed most.

• 2 hr or less average question response time assures that skills development stays on track.

Personal Career Services

Empower job-readiness.
• Access to a Github portfolio review that can give you an edge by highlighting your
strengths, and demonstrating your value to employers.*

• Get help optimizing your LinkedIn and establishing your personal brand so your profile
ranks higher in searches by recruiters and hiring managers.

Mentor Network

Highly vetted for effectiveness.

• Mentors must complete a 5-step hiring process to join Udacity’s selective network.

• After passing an objective and situational assessment, mentors must demonstrate

communication and behavioral fit for a mentorship role.

• Mentors work across more than 30 different industries and often complete a Nanodegree
program themselves.

*Applies to select Nanodegree programs only.

Artificial Intelligence 14
Learn more at
www.udacity.com/online-learning-for-individuals →

12.22.22 | V1.0