CSE 107 COURSE INFORMATION

Welcome to CSE107: Intro to Modern Cryptography!

Times and Places

Class: RWAC 0115; Mon,Tue,Wed,Thu 2pm-3:20pm

Discussion: DIB 122; Mon,Wed noon-12:50pm

Office Hours: See “Course Staff & How To Contact Us” section

Course website: https://cseweb.ucsd.edu/classes/s125/cse107-a/

(Tentative) Schedule, With Links to Homeworks

Note: Other than the Midterm and Final, the schedule is subject to change – some topics might take shorter or longer than expected, and we might end up having fewer mini-quizzes. We’ll be sure to announce any big changes.

What You’ll Learn

Cryptography is the science of secure communication.

Historically cryptography was only about secret codes and was more of an art than a science. But modern cryptosystems are built in ways that allow us to prove them secure assuming the hardness of well-studied computational problems. And today’s cryptography allows doing things that seem clearly impossible; e.g., how can two strangers with no common background knowledge agree on a shared secret value that an eavesdropper observing all their communication won’t also know? (Every time you visit a website, your computer does this with the webserver!)

Once you’ve taken this class, you’ll be able to give sound advice on various real-world questions about secure communication; e.g., “If I’m worried about X, will using Y be enough to protect me?”. Importantly, you’ll also be able to acknowledge when you don’t know the answer rather than giving unsafe or dangerous advice. You’ll also be able to…

• describe the building blocks (called cryptographic primitives) that modern cryptosystems are built from
• compare standard security definitions that capture various real world security properties
• evaluate in a given situation which primitive or security definition is called for
• explain how many real-world cryptographic systems and protocols work (What happens when you connect to an HTTPS website? How do those 2FA devices work?)
• carry out attacks that break the security of various plausible but broken cryptosystems

Along the way we’ll cover the mathematical foundations on which modern cryptography is based. You’ll learn, for example, how to solve equations involving things that aren’t numbers (groups and rings and fields!) and how to compute the last digit of 107²⁰²⁵ in your head (Euler’s theorem!).

Course Staff & How To Contact Us

The fastest way to contact us is to send a message to “Instructors” on Piazza.

Technology / Materials

We’ll be using Gradescope for submitting assignments and Piazza for class-related questions and announcements. Both are linked to from Canvas.

For coding assignments we’ll be using the Python programming language (version 3.9 or later) and you’ll need to be able to install Python libraries with pip. If you don’t have access to a computer on which you can install Python, you should apply to the Laptop Lending Program as soon as possible: https://basicneeds.ucsd.edu/financial-security/laptop-lending-program.html

There is no required textbook. We’ll very loosely be following “Introduction to Modern Cryptography” by Katz and Lindell. Everything you’re responsible for knowing will be covered in lecture or homework, so the textbook isn’t required, but you might find the book helpful to augment what you learn in lecture. PDFs of it are easily found online.

Feedback

This is my first time teaching CSE107 (although I’ve TA’d it four times) and I’ll be teaching it a little differently from how it’s usually taught. Let me know what’s working and what isn’t! I’ll post a feedback form each week.

Classroom Environment

My goal is to make the classroom a welcoming place where everyone feels safe asking questions when they’re confused, and where everyone is able to focus on learning. I’m interested in hearing from everyone (and especially from anyone in a group that’s underrepresented in computer science) if there’s anything I can do to better achieve this goal!

I will not tolerate any racism, transphobia, sexism, or any other form of bigotry in my classroom.

Collaboration Policy

You are strongly encouraged to collaborate with your fellow CSE107 students on the problem sets. Just like in real life, you can accomplish more working together than you can alone. And just like in real life, you must list your collaborators rather than taking all the credit yourself; intentional failure to list your collaborators is an academic integrity violation.

Collaboration means discussing ideas; it does NOT mean directly copying someone’s code, posting entire solutions to a discord chat, etc. In general, you shouldn’t be talking to a collaborator at the same time as you’re writing your solution, but it’s fine to work out the solution to a problem on a chalkboard with others and then go home and write/code the solution by yourself. The point of this policy is to make sure that what you’re turning in reflects your own understanding of the solution rather than that of your collaborators.

Likewise, you’re not allowed to copy (or paraphrase) code from the internet that solves a problem substantially similar to one on the homework. (But it’s fine to look at non-cryptography-related code online to see examples of basic Python syntax.)

LLM / “AI” Policy

LLMs like ChatGPT are great at producing plausible looking but subtly wrong text (since LLMs have no notion of truth or correctness). For code, this translates to plausible looking code with subtle, hard-to-find bugs (which might only trigger with low probability). Also, LLMs are famously bad at math.

Meanwhile cryptography is a math-heavy field where subtle, low-probability bugs can (and often do) lead to catastrophic loss of security, and it’s impossible to test for all classes of such bugs.

Hopefully it’s obvious why using LLMs to produce cryptographic code is a terrible idea.

In this class, you are not allowed to use LLMs or other “AI” tools to write your solutions or code. We won’t use LLMs for making or grading the assignments, and we expect you not to use them for solving the assignments.

Extension policy

You can email me and I’ll give you a one day extension, no questions asked. (Make the subject line “[CSE107] One Day Extension” or it will get lost in my inbox and I won’t see it.) But if you’re doing this on every assignment, I’ll check in with you to try to figure out what’s going on and what we can do about it. (Note that using a one-day extension on every assignment leaves you with on average the same number of days to do each assignment as you’d have without extensions; it just shifts everything by one day. You might as well turn everything in on the actual due date so that you have the one-day extension available as a backup if something goes wrong.)

If something comes up and you need more than one day, email me to let me know what’s going on. I don’t intend to be super strict about granting extensions – I myself often start things too close to the deadline, and I know you all have other things going on in your lives besides this class. That said, any extension longer than a few days would make it hard for you to catch up again – this is a quarter-long class squeezed into a single month, so it moves pretty fast.

Attendance policy

A decent fraction of “lecture” time will actually be labs / learning activities, often done in groups, because studies consistently show these to be more effective than traditional lectures. I won’t be taking attendance, and I’ll make lecture recordings available upon request (email me if you need one), but the recordings won’t really be able to replace the labs/activities, so you should plan on attending the lectures. (Missing a few won’t be a problem, but if you’re planning to skip every lecture, then this is not the class for you. This class is in-person, not remote.)

I understand how much of a hassle it can be to commute to campus – I myself have a one-hour commute each direction – and I know that some of you will have other commitments like jobs. So while I’m hoping most of you will attend most lectures, I’ll do my best to communicate in advance what activities are happening on what days.

Illness

DO NOT UNDER ANY CIRCUMSTANCES COME TO CLASS IF YOU ARE SICK WITH ANY CONTAGIOUS ILLNESS. Your health and well-being are more important than any class, including this one. Also, I don’t want to catch whatever you have, and neither do your classmates. Stay home and email me for the lecture recordings.

If you’re sick for the midterm or final, stay home, email me, and we’ll work something out.

Grading

Your overall grade is based on the homeworks, the final, the midterm, and optionally on “mini-quizzes”.

Once per week we’ll have a short in-class “mini-quiz”. These are optional and low-stakes: low mini-quiz scores won’t hurt your overall grade, but high quiz scores can slightly improve your overall grade. These are opportunities for you to get practice with the types of questions that will appear on the midterm and final. (The weekly mini-quizzes are something new we’re trying, and we might decide partway through the class to stop doing them.)

The overall grade is computed as follows:

• 40% of the grade is homeworks (there were originally planned be 4 homeworks [plus “Homework 0” which weighs significantly less than the other homeworks] but there will instead be five)
• 60% of the grade is “combined test average”, computed as a weighted average of the following:
  • Midterm: weighs 40 points
  • Final: weighs 60 points
  • Mini-quizzes where you do better than your Midterm+Final average: weigh 10 points each
  • Mini-quizzes where you do worse than your Midterm+Final average: not counted (i.e., they weigh 0 points)

For example, assuming we have 4 mini-quizzes, if you skip or get low scores on all of them, then your grade is computed as 40% homeworks, 24% midterm, 36% final. If you get 100% on all 4 mini-quizzes, then your grade will be roughly 40% homeworks, 17% mini-quizzes, 17% midterm, 26% final.

The point of all this is just to ensure that a mini-quiz can slightly help your grade (and reduce the weight of the midterm and final) but getting a low score or skipping it entirely won’t hurt your grade (the midterm and final will just be weighted slightly more to make up for it).

Tentative letter grade cutoffs

• A: 90% - 100%
• B: 80% - 90%
• C: 70% - 80%
• D: 60% - 70%
• F: 0% - 60%

However, this is my first time teaching the class and the assignments are all (mostly) new, so I may need to adjust the cutoffs if I’ve accidentally made things too hard. In any case your letter grade will be at least as good as what’s given by the cutoffs above (e.g., 75% will never be worse than a C)

Grading scheme for “Real-world advice” questions

Some short-answer questions on homeworks / mini-quizzes / exams will be marked as “Real-world advice” questions and will be graded as follows:

• +3 Correct, useful advice
• +2 Mostly correct / somewhat useful advice
• +1 “I don’t know”
• +0 Technically correct but unhelpful / useless advice
• -1 Advice that’s wrong but won’t lead to danger
• -3 Dangerous/unsafe advice

You’ll get more points for acknowledging when you don’t know something than you will for pretending you do when you don’t.

Occasionally, to mix things up, you might get a question where someone wants your advice on something clearly and uncontroversially unethical (e.g. making ransomware to be unleashed on a hospital). In this case, the grading scheme is more or less reversed: positive points for “wrong” advice that averts the bad outcome or for refusing to give any advice, and negative points for giving advice that is actually useful. For full credit, your answer should point out what part of the advice is “wrong” / how it prevents the bad outcome.