CS221: Artificial Intelligence: Principles and Techniques

Communication: We will use Ed for all communications. You will be automatically enrolled to the course Ed, which you can access via Canvas. SCPD students, please email scpd-gradstudents@stanford.edu if you need assistance. We encourage all students to use Ed, either through public or private posts. However, if you have an issue that you would like to discuss privately, you can also email us at cs221-spr2122-staff@lists.stanford.edu, which is read by only the faculty and the co-head CAs.

Academic accommodations: If you need an academic accommodation, you should initiate the request with the Office of Accessible Education (OAE). The OAE will evaluate the request, recommend accommodations, and prepare a letter for faculty. Students should contact the OAE as soon as possible and at any rate in advance of assignment deadlines, since timely notice is needed to coordinate accommodations. It is the student's responsibility to reach out to the teaching staff regarding the OAE letter.Please send your letters to cs221-spr2122-staff@lists.stanford.edu by Friday, April 15 (week 3).

Time / Location: Below is an overview of the course components. All class activities and office hours are in our class calendar. All following times are in Pacific Time (PT):

Lectures: Mon/Wed 1:30-2:50pm in NVIDIA auditorium.
Sections: Fridays 3:15-4:45pm, Huang 018.
Office Hours: CA office hours are in the Huang basement; see calendar for times; see Office Hour Logistics for logistics.

Instructor:

Tatsunori Hashimoto

Course Assistants:

Amelie Byun
Course Coordinator

Sam Lowe
Co-Head CA

Amrita Palaparthi
Co-Head CA

Weston Hughes
Homework CA

Yuchen Wang
Homework CA

Yilin Wu
Homework CA

Ishaan Gulrajani
Homework CA

Skanda Vaidyanath
General CA

Samar Khanna
General CA

Xiaoyuan Ni
Homework CA

Lantao Yu
Homework CA

Akash Velu
General CA

Sharan Ramjee
General CA

Ryan Cao
General CA

What is this course about? What do web search, speech recognition, face recognition, machine translation, autonomous driving, and automatic scheduling have in common? These are all complex real-world problems, and the goal of artificial intelligence (AI) is to tackle these with rigorous mathematical tools. In this course, you will learn the foundational principles that drive these applications and practice implementing some of these systems. Specific topics include machine learning, search, game playing, Markov decision processes, constraint satisfaction, graphical models, and logic. The main goal of the course is to equip you with the tools to tackle new AI problems you might encounter in life.

Prerequisites: This course is fast-paced and covers a lot of ground, so it is important that you have a solid foundation on both the theoretical and empirical fronts. You should have taken the following classes (or their equivalents):

Programming (CS 106A, CS 106B, CS 107)
Discrete math (CS 103)
Probability (CS 109)

Reading: There is no required textbook for this class, and you should be able to learn everything from the lecture notes and homeworks. However, if you would like to pursue more advanced topics or get another perspective on the same material, here are some books:

Russell and Norvig. Artificial Intelligence: A Modern Approach. A comprehensive reference for all the AI topics that we will cover.
Koller and Friedman. Probabilistic Graphical Models. Covers factor graphs and Bayesian networks (this is the textbook for CS228).
Sutton and Barto. Reinforcement Learning: An Introduction. Covers Markov decision processes and reinforcement learning. Available free online.
Hastie, Tibshirani, and Friedman. The elements of statistical learning. Covers machine learning. Available free online.
Tsang. Foundations of constraint satisfaction. Covers constraint satisfaction problems. Available free online.

Bear in mind that some of these books can be quite dense and use different notation terminology, so it might take some effort to connect up with the material from class.

Video access disclaimer: A portion of class activities will be given and recorded in Zoom. For your convenience, you can access these recordings by logging into the course Canvas site. These recordings might be reused in other Stanford courses, viewed by other Stanford students, faculty, or staff, or used for other education and research purposes. If you have questions, please contact a member of the teaching team.

CA office hours are either in the Huang basement or online. See the calendar for the schedule.
- For online office hours, you will need to register an account on QueueStatus. When you wish to join the queue, click "Sign Up" at the CS 221 queue. Be sure to enter your email when you "Sign Up"; this is a way for the CA to contact you. For online office hours, you will need to install Zoom to video call with the CA. See instructions below.
Installing Zoom:
- Linux: Go to the Zoom downloads page and download appropriate Linux package from 'Zoom Client for Linux'. Open package (e.g. with 'Ubuntu Software Center' or other appropriate application) and install.
- Windows: Go to Stanford Zoom and click 'Launch Zoom'. Click 'host meeting'; nothing will launch but there will a link to 'download & run Zoom'. Click on 'download & run Zoom' to download 'Zoom_launcher.exe'. Run 'Zoom_launcher.exe' to install.

Grading

Homeworks (75%): There will be weekly homeworks with both written and programming parts. Each homework is centered around an application and will also deepen your understanding of the theoretical concepts. Some homeworks will have a competition component; winners will receive extra credit. Here are all the homework deadlines:

Exam (25%): The exam is a three-hour written exam that will test your knowledge and problem-solving skills on all preceding lectures and homeworks. You cannot use any external aids except one double-sided page of notes. If you have a major conflict (e.g., an academic conference), you should let us know privately via email by Friday, April 15 (week 3).
Exam date/time/location: June 7, 12:15pm-3:15pm, location TBA.

Please note that we will not be able to accommodate students with a final exam conflict due to enrollment in a conflicting course. Students should make sure that they are able to take the CS 221 final exam at the normal time if they are enrolled in a conflicting course.

For SCPD students, your exams will be distributed through the SCPD office once you have set up an exam monitor.

Project (optional, ungraded): The final project provides an opportunity for you to use the tools from class to build something interesting of your choice. Projects should be done in groups of one to four students. Each project group will be assigned a CA mentor who will give feedback and answer questions, and thus the project will be an opportunity for mentorship. We will award up to 2% of extra credit to projects that are done well.
The project will be something that you work on throughout the course and we have set up some milestones to help you along the way:
- Project interest form [p-interest] (due Apr 15)
- Project proposal [p-proposal] (due Apr 29)
- Project progress report [p-progress] (due May 20)
- Project final report and video [p-final] (due June 3)
See the project page for more details.

Regardless of the group size, all groups must submit the work detailed in each milestone and will be graded on the same criteria. Although we allow 1-2 person project groups, we encourage groups of 3-4 members. We encourage teams of 3-4 students because this size typically best fits the expectations for CS 221 projects. We expect each team to submit a completed project (even for team of 1 or 2). All projects require that students spend time gathering data, and setting up the infrastructure to reach an end result. A 3 or 4 person team can share these tasks much better, allowing the team to focus more on the interesting results and discussion in the project. Each member of the team should contribute in both technical and non-technical components of the project. We will provide resources on Ed and the project page that can help you find group members.

For inspiration, we have made some previous CS221 projects available for viewing.
Ed: You will be awarded with up to 2% extra credit if you answer other students' questions in a substantial and helpful way.

Homeworks

Written assignments: Homeworks should be written up clearly and succinctly; you may lose points if your answers are unclear or unnecessarily complicated. Here is an example of what we are looking for. Written homework must be typeset using LaTeX, Microsoft Word, Pages for Mac, or an equivalent program, and submitted as a PDF. We strongly encourage you to use LaTeX: there are user-friendly web interfaces like Overleaf. See this getting started guide. Figures may be hand-drawn, so long as they are included in a typeset PDF.

Programming assignments: The grader runs on Python 3.7, which is not guaranteed to work with alternative versions (e.g., Python 2.7). Please use Python 3.7 to develop your code.

The programming assignments are designed to be run in GNU/Linux environments. Most or all of the grading code may incidentally work on other systems such as MacOS or Windows, and students may optionally choose to do most of their development in one of these alternative environments. However, no technical support will be provided for issues that only arise on an alternative environment. Moreover, no matter what environment is used during development, students must confirm that their code (specifically, the student's submission.py) runs on Gradescope.

The submitted code will not be graded if it has one of the following issues:

The original grader.py script (operating on the submitted submission.py) may not exit normally if you use calls such as quit(), exit(), sys.exit(), and os._exit(). Also note that Python packages outside the standard library are not guaranteed to work. Therefore, do not use packages like numpy, scikit-learn, and pandas.
The code reads external resources other than the files given in the assignment.
The code is malicious. This is considered a violation of the honor code. The score of the assignment will be zero (0) and the incident will be reported to the Office of Judicial Affairs.

Collaboration policy and honor code: You are free to form study groups and discuss homeworks and projects. However, you must write up homeworks and code from scratch independently, and you must acknowledge in your submission all the students you discussed with. The following are considered to be honor code violations:

Looking at the writeup or code of another student.
Showing your writeup or code to another student.
Discussing homework problems in such detail that your solution (writeup or code) is almost identical to another student's answer.
Uploading your writeup or code to a public repository (e.g. github, bitbucket, pastebin) so that it can be accessed by other students.
Looking at solutions from previous years' homeworks - either official or written up by another student.

When debugging code together, you are only allowed to look at the input-output behavior of each other's programs (so you should write good test cases!). It is important to remember that even if you didn't copy but just gave another student your solution, you are still violating the honor code, so please be careful. We periodically run similarity-detection software over all submitted student programs, including programs from past quarters and any solutions found online on public websites. Anyone violating the honor code will be referred to the Office of Judicial Affairs. If you feel like you made a mistake (it can happen, especially under time pressure!), please reach out to the instructor or the head CA; the consequences will be much less severe than if we approach you.

Submission

Electronic Submission: All homework assignments are due at 11pm (23:00, not 23:59) Pacific time on the due date.
Refer to coursework for exam deadlines. Assignments are submitted through Gradescope. Do not submit your assignment via email. If anything goes wrong, please ask a question on Ed or contact a course assistant. If you need to sign up for a Gradescope account, please use your @stanford.edu email address. You can submit as many times as you'd like until the deadline: we will only grade the last submission. Submit early to make sure your submission runs properly on the Gradescope servers. Gradescope will run grader.py on the programming questions and give you feedback on non-hidden test cases. You are responsible for checking that your program runs properly on these cases. You will not get credit otherwise. If anything goes wrong, please ask a question on Ed or contact a course assistant. Do not email us your submission. Partial work is better than not submitting any work.

For assignments with a programming component, we will automatically sanity check your code in some basic test cases, but we will grade your code on additional test cases. Important: just because you pass the basic test cases, you are by no means guaranteed to get full credit on the other, hidden test cases, so you should test the program more thoroughly yourself!

Unless the assignment instructs otherwise, all of your code modifications should be in submission.py and all of your written answers in <assignment ID>.pdf. Upload the former to Gradescope under the "Programming" section, and the latter under the "Written" section.

For the project milestones, make sure that all members of your group submit. The submission should include a group.txt file which should contain the SUNetIDs of the entire group, one per line.

Late days: A homework is $\lceil d \rceil$ days late if it is turned in $d$ days past the due date (note that this means if you are $1$ second late, $\lceil d \rceil = 1$ and it is 1 day late). You have seven (7) late days in total that can be distributed among the homeworks without penalty. After that, the maximum possible grade is decreased by 25% each day (so the best you can do with $d = 1$ is 75%). As an example, if you are out of late days and submit one day late, a 90 will be capped at a 75, but a 72 will not be changed. Note that we will only allow a max of $d = 2$ late days per homework, though, so if $d > 2$ then we will not accept your submission. Gradescope is set up to accept written and programming submissions separately. Late days are calculated per-assignment, with the number of late days used depending on the later submission. So, if the programming part is $1$ day late and the written part is $2$ days late, then that will count as using $2$ late days.

Regrades: If you believe that the course staff made an objective error in grading, then you may submit a regrade request for the written part of your homework or exam. Remember that even if the grading seems harsh to you, the same rubric was used for everyone for fairness, so this is not sufficient justification for a regrade. It is also helpful to cross-check your answer against the released solutions. If you still choose to submit a regrade request, click the corresponding question on Gradescope, then click the "Request Regrade" button at the bottom. Any requests submitted over email or in person will be ignored. Regrade requests for a particular assignment are due by Monday 4pm, one week after the grades are returned. Note that we may regrade your entire submission, so that depending on your submission you may actually lose more points than you gain.

	Monday	Tuesday	Wednesday	Thursday	Friday	Saturday	Sunday
Week 1	Mar 28	Mar 29	Sep 30	Sep 31	Apr 1	Apr 2	Apr 3
	Lecture: Intro / Overview 1:30-3pm [foundations] homework release		Lecture: Machine Learning 1 (Linear Models) 1:30-3pm		Section 1 3:15-4:45pm [Problems] [Solutions] [Video]
Week 2	Apr 4	Apr 5	Apr 6	Apr 7	Apr 8	Apr 9	Apr 10
	Lecture: Machine Learning 2 (Neural Networks) 1:30-3pm [foundations] homework due [sentiment] homework release		Lecture: Machine Learning 3 (Unsupervised Learning) 1:30-3pm		Section 2 3:15-4:45pm [Problems] [Solutions] [Video]
Week 3	Apr 11	Apr 12	Apr 13	Apr 14	Apr 15	Apr 16	Apr 17
	No lecture		Lecture: Search 1:30-3pm		Project interest form due Section 3 3:15-4:45pm [Slides] [Problems] [Solution] [Video]
Week 4	Apr 18	Apr 19	Apr 20	Apr 21	Apr 22	Apr 23	Apr 24
	Lecture: Search and Heuristics 1:30-3pm [foundations] solutions release [sentiment] homework due [reconstruct] homework release		Lecture: Policy 1:30-3pm		Section 4 3:15-4:45pm [Slides] [Problems] [Solution] [Video]
Week 5	Apr 25	Apr 26	Apr 27	Apr 28	Apr 29	Apr 30	May 1
	Lecture: Reinforcement Learning / SARSA 1:30-3pm [sentiment] solutions release [reconstruct] homework due [blackjack] homework release		Lecture: Minimax Problems 1:30-3pm		Project proposal due Section 5 3:15-4:45pm [Problems] [Solutions] [Video]
Week 6	May 2	May 3	May 4	May 5	May 6	May 7	May 8
	Lecture: TD Learning 1:30-3pm [reconstruct] solutions release [blackjack] homework due [pacman] homework release		Lecture: Constraints 1:30-3pm		Section 6 3:15-4:45pm [Problems] [Solutions] [Video]
Week 7	May 9	May 10	May 11	May 12	May 13	May 14	May 15
	Lecture: Beam Search 1:30-3pm [blackjack] solutions release [pacman] homework due [scheduling] homework release		Lecture: Bayesian networks 1:30-3pm		Section 7 3:15-4:45pm [Problems] [Solutions] [Video]
Week 8	May 16	May 17	May 18	May 19	May 20	May 21	May 22
	Lecture: Forward-Backward Algorithm 1:30-3pm [pacman] solutions release [scheduling] homework due [car] homework release		Lecture: Bayes Learning and EM 1:30-3pm		Project progress report due Section 8 3:15-4:45pm [Problems] [Solutions] [Video]
Week 9	May 23	May 24	May 25	May 26	May 27	May 28	May 29
	Lecture: Logic (semantics) 1:30-3pm [scheduling] solutions release [car] homework due [logic] homework release		Lecture: Logic (first-order) 1:30-3pm		Section 9 3:15-4:45pm [Problems] [Solutions] [Video]
Week 10	May 30	May 31	June 1	June 2	June 3	June 4	June 5
	No classes (Memorial day) [car] solutions release [logic] homework due		Lecture: Summary and future of AI 1:30-3pm		Project final report due Section 10 3:15-4:45pm	[logic] solutions release
Week 11	June 6	June 7	June 8	June 9	June 10	June 11	June 12
		Final Exam 12:15-3:15pm