📖 Syllabus

Table of Contents

1. About 🧐
2. Communication 💬
3. Technology 💻
4. Course Structure 🍎
   1. Homework
5. Exams 🧪
6. Policies ✏️
   1. Grading
   2. Late Policy and Drops
   3. Regrade Requests
   4. Incomplete Grades
   5. Academic Integrity
   6. A note on letter grades
7. Support 🤝
   1. Accommodations
   2. Diversity and Inclusion
8. Acknowledgements 🙏

---

Students on the waitlist or joining late: You are expected to keep up with all work and submit assignments by the deadlines from the start of the quarter.

---

About 🧐

DSC 40A answers the question: “How do we learn from data?” We will see that virtually every rigorous learning method involves two steps: turning the abstract problem of learning into a concrete math problem, and then solving that math problem. After this class, you will understand the basic theoretical principles underlying almost every machine learning and data science method, from simple linear regression to deep neural networks.

---

Communication 💬

All course announcements and communication will happen on Piazza. You should be added automatically. If you have questions about course content, homework, or logistics, post on Piazza. Make posts private if they include any part of your homework solution.

I will not respond to emails. All questions must go through Piazza so the entire course staff can help and other students can benefit from the discussion.

---

Technology 💻

• Course Website: All content posted here
• Piazza: Announcements and communication (auto-enrolled)
• Gradescope: Submit assignments and view grades (auto-enrolled)
• Google Colab: HW requiring a coding component

We will not use Canvas (except for auto-enrollment in Gradescope and UCSD egrades).

---

Course Structure 🍎

Lectures: Pre-recorded lectures cover theory (suggest watching at 1.2-1.5x speed). In-class time is for live problem-solving sessions (HOPS - Hands-On Problem Solving). We work through practice problems together step-by-step. Attendance is not required but highly recommended. If you do come to HOPS, make sure you have watched the corresponding (preceding) lecture.

Discussion: Work on problems in groups of 2-4 students. We will not grade these assignments, however if your team does turn them in a good faith effort to Gradescope by the end of the discussion day, we will use them as extra credit points for those students. The following discussion section will be dedicated to working through the answers as a discussion section class.

Office Hours: The homework is challenging and most students cannot complete it from lecture alone. Come to office hours! See the Calendar tab for times and locations. Or come to chat about industry, work as a data scientist or the data science degree at UCSD. We are happy to chat about any and all of it!

Homework

Submission:

• HWs must be typed in LaTeX. We provide templates and the first discussion section will be dedicated to showing you how to do HW 0 typed up in LaTeX should you have any doubts. With modern day LLMs getting help with LaTeX formatting should not be an issue and it is an important skill to develop. The DSC Capstone 180AB also requires you to use LaTeX.
• This helps us to grade faster, and removes the ambiguity of handwritten math answers, of which every year we have numerous issues.

Due Dates:

• Always released at 5:59 AM, and always due at 11:59 PM. Gradescope is source of truth for due dates and release dates. Though we will try to keep the course homepage up to date.
• Each HW and Groupwork comes with a 6-hour grace period until 5:59 AM next morning.
• Turning in your HW after the grace period will result in a 0.

Grading:

• Correct answer + correct work = 100%
• Correct approach + minor arithmetic or algebraic error = 50%
• Wrong approach or no work = 0%

Extensions: If you have a valid reason for missing a deadline (illness, family emergency, etc.), contact the TAs via Piazza with documentation (e.g., doctor’s note). TAs have full authority to approve extensions without checking with the instructor. If approved, you have two options: (1) submit within 48 hours of the original deadline for full credit, or (2) don’t submit and receive the mean score of your other homework assignments. But you must get TAs approval, so they can mark down who you are and how we need to alter your grade.

Collaboration: You can discuss solution strategies with other students, but do not share any written work. You can tell someone how to approach a problem, but you cannot show them your solution. Think of it this way: your collaboration should be able to happen over the phone. Cite the names of anyone you discussed problems with on your submission in the field on page 1.

Getting Help: Come to office hours! The homework is challenging. Use the resources on the course website. Post questions on Piazza about approaches (make posts private if they include any part of your answer). We cannot tell you if your answer is correct.

---

Exams 🧪

There will be one Midterm Exam and one Final Exam, both held in person.

Midterm: 20 points, covers material from the first half of the course Final: 30 points, cumulative (covers all course material)

Redemption: If you score higher on the Final than the Midterm, your Midterm grade will be replaced with your Final grade. This gives you a second chance to demonstrate mastery of early course material.

You must take the Final Exam to pass the course.

---

Policies ✏️

Grading

In an effort for full transparency, I will be providing you all with the python script (as a downloadable zip with a README) that will allow you to insert your grades and calculate a final score. This will be the same exact script I run, and thus should not incur any surprises or questions from all of you come final grades. This script will be available soon, once I finish updating it for this quarter.

Your grade will be computed out of 100 points as follows:

Grading Scale (no rounding):

These cutoffs are the strictest they will be. After the midterm, I may lower them (make them easier), but I will not raise them.

Grades will not be rounded. Extra credit opportunities are built into the midterm and final exams. You can show up LATE to an exam, but once the first person walks out, you will not be able to begin.

Late Policy and Drops

See Homework section for grace period and extension details. Lowest HW score replaced with highest (if submitted and graded). If you have extenuating circumstances, contact TAs ASAP.

Regrade Requests

You have 24 hours after HW grades are posted to request a regrade. This can be done via gradescope. There will be no regrade requests for the final exam.

Incomplete Grades

If you experience a significant setback outside your control (illness, loss, etc.), you may be eligible for an Incomplete grade. Contact the TAs, the Tutors, or myself ASAP. Note: Incompletes are for work not yet completed, not for redoing completed work and must be accepted by the department.

Academic Integrity

Work hard, use allowed resources, and act with integrity. See UCSD Policy on Integrity of Scholarship.

Not allowed:

• Sharing written homework solutions
• Looking for homework answers online or using AI tools like ChatGPT/Copilot for solutions.
• Collaborating on exams or using unauthorized resources (not in the exam instructions) will result in a failure.

Allowed:

• Discussing homework strategies with classmates
• I highly encourage you to use AI/LLMs to help you explain and refine concepts. But know they can mislead occasionally. Especially in the probability type problems, or often have arithmetic issues.
• Reading about concepts (but cite in the HW on page 1 if you accidentally find related material)

As always violations will be reported to the Academic Integrity Office and result in failing the course. Let’s not have to go through that embarrassment or paperwork… please.

A note on letter grades

Grading is never penalized. If everyone masters the material, everyone can get an A+. The grading scale (95+ for A+, 89+ for A, etc.) is built in with the idea that partial credit only can count for 1/2 points, and that this is challenging material. Bonus points on exams and group work accounts for any rounding issues that you would argue for.

Try your best not to worry too much about grades, and we’ll reciprocate by being fair. We’re in this together 😎.

---

Support 🤝

Accommodations

If you have accommodations through OSD, share your AFA letter with the instructor and dscstudent@ucsd.edu by the end of Week 2.

Diversity and Inclusion

We are committed to an inclusive learning environment. Honor and respect your classmates per the UCSD Principles of Community.

---

Acknowledgements 🙏

Thanks to other instructors of this course who have made various contributions, including but not limited to Janine Tiefenbruck, Aobo Li, Yian Ma, Gal Mishne, Justin Eldridge, and Suraj Rampure. Thanks also to the many tutors and TAs who have supported this course since its inception!