What does AI mean to you?

What is this course all about?

What is this course all about?

Computers are dumb...

...but we can make them do incredibly clever things

Let's take a look!

This course covers:

Search & games
Dealing with uncertainty
Machine learning, deep learning & RL
Ethics of AI

The TA Team

Andrei

Brent

Deeptha

Raahil

Andrei

Brent

Certain websites such as Campuswire can not be embedded as an iframe in a reveal JS presentation. This page should automatically redirect you (and auto forward to the next slide when you come back here), but if it doesn't, use the link above.

Components

Attendance Requirements

Mandatory attendance - exams + quizzes

No remote options for any lectures

Policies - Homework Submissions

Deadlines: 8:00 pm, Friday

Assignments will be released on Canvas

Submissions: Gradescope

Grades released on Gradescope

Regrade Requests: submit within 1 week of receiving grade

Policies - Homework Logistics

(Starting HW1) Before the Friday 8 pm deadline, you must:

1. Upload an initial working submission to Gradescope.

2. Meet with a TA during office hours to complete a comprehension check/walkthrough.

Between Fri 8 pm and a 'late' deadline on Sun 8pm:

Update/refine your submission, incorporate feedback and submit final version.

Late Policy

No penalty until Sunday 8:00 pm

Warning: No office hours between Fri night to Sun night.

If you have accommodations through DAS, get the notification letter sent ASAP

Life happens, come talk to me!

Policy on Collaboration and Generative AI

Discussing concepts is okay.
Discussing the actual problem on the assignment is not.
If in doubt, just ask me!

For assignments (except assignment 0), use whatever you'd like, but you must cite it.
Required Disclosure Statement for every assignment.

Failure to disclose AI use will be treated as an academic integrity violation.

Course Rigor

This is not an easy course.

AI is inherently complex and mathematical.
There will be equations, lots of them.

Most students who work hard get an A.
This is not because I'm lenient - they had to earn it.

Tips for Success

Be regular to class, and use pen and paper or a writing tablet to take notes. Avoid distractions.

Spend time with practice problems (on Canvas) right from the beginning.

Ask loads of questions! I love interactive lectures.

Come to office hours!

Academic Integrity

If I catch you cheating...

• You will fail this course


• You will be reported to OSCCR
  Penalties ranging from warnings/probation to dismissal


• [GRAD students] You will be reported to the Khoury Academic Integrity Committee
  Consequences include:
  
  
  • Immediate and permanent loss of TA/co-op
    (Most reinstatement applications are unsuccessful)
  • Dismissal

My Expectations From You

Come to class having studied notes in advance

Manage your time wisely - you will always have conflicting deadlines

Read all instructions, follow policies strictly, and respect academic integrity!

Come to office hours - get to know me outside class

Challenge yourselves - that's why you're here!

Ask questions - I'm here to help you learn!

My Commitments To You

I will always welcome student questions

I will foster a nonjudgmental environment

You can reach out to me for any reason*

*Faculty are mandatory reporters at Northeastern - see syllabus for details.

Required Background

Python 3, strong programming skills

Basics of Linear Algebra
(matrix and vector operations, dot products)

Basics of Probability & Statistics
(conditional probability, Bayes' theorem)

Basics of Calculus
(differentiation, partial derivatives)

Assignment 0 - out today!

Self-placement/evaluation

Locked until Syllabus Quiz is completed

Covers required background

Tests programming ability

Contains resources to build necessary skills

Due on May 15

Attempt this honestly - I don't expect anyone to know everything!

Any questions for/about me?

Where did it all start?

Can Machines Think?

Alan Turing, 1950

Some Philosophical Questions...

Is a machine intelligent if it merely mimics human thinking?

Is a machine intelligent if it can fool a human into believing so?

Could you construct a machine whose workings you couldn't explain?

What does intelligence have to do with free will?

Thinking = symbol manipulation

"Intelligence can be replicated
using a computer program"

- Mathematicians in the 1950s

1956, Dartmouth College, workshop on AI

"The conjecture that every aspect of learning or any other feature of intelligence can be,
in principle, so precisely described that a machine can be made to simulate it"

What is an intelligent agent?

A system or program that perceives and interacts with its environment
while making informed decisions to achieve a certain goal or maximize some utility

A system or program that perceives and interacts with its environment
while making informed decisions to achieve a certain goal or maximize some utility

How does a program perceive its environment?

• Computer Vision
• Depth Sensors
• Language
• Explicit Representations

A system or program that perceives and interacts with its environment
while making informed decisions to achieve a certain goal or maximize some utility

How does a program interact with its environment?

• Physical Interactions
• Image Generation
• Language

A system or program that perceives and interacts with its environment
while making informed decisions to achieve a certain goal or maximize some utility

How does a program learn?

• Rule Based Systems
• Learning from Computation
• Learning from Data
• Learning from Mistakes

A system or program that perceives and interacts with its environment
while making informed decisions to achieve a certain goal or maximize some utility

How does a program reason?

• Objective Functions
• Utility
• Fairness
• Biases

Consider a Roomba (a robot vacuum)

Think about

• Percepts
• Interactions
• Learning/Inference
• Rationality

How about Netflix/Youtube/Amazon Recommendations?

Think about

• Percepts
• Interactions
• Learning/Inference
• Rationality

Categorizing AI Environments

• Fully Observable v/s Partially Observable
• Single Agent v/s Multi Agent
• Deterministic v/s Non-Deterministic
• Episodic v/s Sequential
• Static v/s Dynamic
• Discrete v/s Continuous
• Known v/s Unknown

Let's now think about building an AI

Consider the Missionaries and Cannibals game...

Consider the Missionaries and Cannibals game...

SEARCH

Representation: State Space Graph

Nodes: States or World Configurations

Edges: Actions

Edge Weights: Cost of Action

Successor/Transition Function: Set of Edges & Edge Weights

Path: Sequence of Actions

Solution: A Path leading from Start to Goal

SEARCH

Let us try and formalize the following problems.