Lecture 2 · Problem formulation
Wolf, Goat, Cabbage
Turn a river-crossing problem into states, actions, and a goal test, then explore its state space.
Open visualization →About the course
Learn to recognize which AI method fits a computational problem, and gain experience implementing those methods on representative problems.
Designed for computer science undergraduate students with no prior exposure to artificial intelligence. Students from related fields with an interest in AI methods and their applications are also welcome.
Deterministic reasoning
- Heuristic and local search
- Adversarial search
- Constraint satisfaction problems
Probabilistic models
- Bayesian networks
- Hidden Markov models
- Kalman and particle filters
- Fully and partially observable Markov decision processes
Machine learning
- Linear models for regression and classification
- Neural networks and kernel methods
- Reinforcement learning
- Perception
Meeting times & locations
Lectures
Tuesday & Friday
12:10–1:30 PM
SEC 118 · Busch Campus
Science and Engineering Resource Center
118 Frelinghuysen Road, Piscataway ↗
Friday recitations
| Section | Time | Room |
|---|---|---|
| 03 | 5:55–6:50 PM | SEC 205 |
| 04 | 2:15–3:10 PM | SEC 202 |
| 08 | 2:15–3:10 PM | SEC 208 |
Science and Engineering Resource Center
Office hours
Instructor: Tuesday, 3:00–4:00 PM
One Spring Street, 3rd floor, New Brunswick
Join instructor office hours on Zoom ↗
ID: 926 5529 6309 · Passcode: 396084
Guoning Zhang: Thursday, 9:30–10:30 AM
Join Guoning’s office hours on Zoom ↗
ID: 940 4764 4953 · Passcode: 734360
Doga Diren: Wednesday, 9:00–10:00 AM
Join Doga’s office hours on Zoom ↗
ID: 913 8900 8922 · Passcode: 250851
Schedule & slides
Class poll: help choose the lecture topics for November. Vote in the November topics poll →
Tentative and subject to change. Slides and notes are available through Canvas. Interactive companions are open to everyone. Use the passcodes below to access the Zoom recordings.
| # | Topic | Course materials | Interactive companions |
|---|---|---|---|
| 01 | Introduction and Overview | Slides (Canvas) Recording (Zoom)Passcode: @D8TJ$Tf | — |
| 02 | Uninformed Search | Slides (Canvas) Recording (Zoom)Passcode: 7ZF.351# | Problem formulation Search algorithms |
| 03 | Heuristic Search | Slides (Canvas) Notes (Canvas)Recording unavailable | — |
| 04 | Local Search | Slides (Canvas) Notes (Canvas) Recording (Zoom)Passcode: t@!9^A+4 | — |
| 05 | Adversarial Search | Slides (Canvas) Recording (Zoom)Passcode: Kxv@C@30 | Alpha–beta pruning Minimax tic-tac-toe |
| 06 | Probabilistic Reasoning | Slides (Canvas) Notes (Canvas) Recording (Zoom)Passcode: L?m.#ZF4 | — |
| 07 | Bayesian Models | Notes (Canvas) Recording (Zoom)Passcode: P*CGia?3 | — |
| 08 | Introduction to Modern Deep Learning I | Posted after lecture | — |
| 09 | Introduction to Modern Deep Learning II | Posted after lecture | — |
| Midterm exam · Date to be announced | |||
| 10 | Temporal Models | Posted after lecture | — |
| 11 | Reinforcement Learning | Posted after lecture | — |
| 12 | Advanced Topics in Neural Networks I | Posted after lecture | — |
| 13 | Advanced Topics in Neural Networks II | Posted after lecture | — |
| Final exam · Date to be announced | |||
Interactive visualizations
Explore the examples, step through the algorithms, and answer the questions as you go.
Lecture 2 · Uninformed search
Search algorithms
Step through breadth-first, depth-first, depth-limited, iterative deepening, uniform-cost, and bidirectional search.
Open visualization →Lecture 5 · Adversarial search
Minimax & alpha–beta pruning
Build a tic-tac-toe game tree and see what each cutoff proves as alpha–beta pruning skips unnecessary branches.
Open visualization →Lecture 5 · Play against minimax
Minimax tic-tac-toe
Play first or second against a minimax agent, then work through three questions about its decisions.
Play the demo →Deep learning · Notebook
Generative deep learning
From naive Bayes to variational autoencoders, following chapters 1 to 3 of Foster's Generative Deep Learning, in PyTorch.
Read the notebook → Open in Colab ↗ Download .ipynbCourse staff & help
| Name | Role | |
|---|---|---|
| Kowndinya Boyalakuntla | Instructor | kowndinya.boyalakuntla@rutgers.edu |
| Doga Diren | Teaching Assistant | doga.diren@rutgers.edu |
| Guoning Zhang | Teaching Assistant | guoning.zhang@rutgers.edu |
- Lectures, homework, or exam material: ask on Piazza. Post publicly when possible so classmates can benefit from the discussion. Course announcements also go out through Piazza.
- Regrade requests or recitation questions: email your TAs. They will bring the matter to the instructor if needed.
- Illness, accommodations, scheduling conflicts, or personal matters: email the instructor directly.
Grading & exams
50%Homework25%Midterm25%Final
Some homework assignments involve programming. Python is the recommended language.
Both exams are closed-book.
Midterm: deterministic reasoning, plus Bayesian networks and inference.
Final: the remaining course topics.
Exam dates will be announced.
Letter grades
| Grade | Score |
|---|---|
| A | Above 89 |
| B+ | 80–89 |
| B | 70–79 |
| C+ | 60–69 |
| C | 50–59 |
| D | 40–49 |
| F | Below 40 |
Textbook
Artificial Intelligence: A Modern Approach
Stuart Russell and Peter Norvig · Fourth edition, 2021 · Pearson
Readings are assigned between lectures. Lecture slides follow the book’s structure closely.