Practice Problems

DSC 40A at UC San Diego

This site contains practice problems taken from old DSC 40A exams. We’ve organized problems in two ways:

1. By topic. This makes it easy to practice old exam problems by topic.
2. By past exam. This makes it easy to simulate taking an entire exam.

In all cases, you should work on these problems on paper, since your exams will also be on paper. Even though you will see some multiple choice bubbles, you intentionally cannot select them.

Note that some offerings in the past had one Midterm Exam and one Final Exam, while other offerings had two Midterm Exams and a two-part Final Exam. Typically:

• Midterms, Midterm 1s, and Final Part 1s cover the first 5 weeks of the course (loss functions, empirical risk minimization, linear regression, gradient descent), while
• Midterm 2s and Final Part 2s cover the latter 5 weeks of the course (mostly probability).
• Offerings with a single Final Exam were typically cumulative.

Note that in Spring, Summer, and Fall 2024, 40A covered more linear algebra concepts (e.g. projecting a vector onto the span of other vectors) than were covered in the past, and our exams will reflect this. Note that gradient descent was not in scope for the midterm, but we did cover it so gradient descent will appear on the final. We did not cover clustering; clustering will not appear on the final.

In Fall 2024 center and spread will not appear on the midterm.

Furthermore, note that different exams were offered in different formats (remote vs. in-person) with different time limits (50 minutes vs. 80 minutes vs. 180 minutes) and different levels of allowed resources (student-created notes sheets allowed vs. instructor-created reference sheets only vs. no notes). In Fall 2024, the Midterm Exam will be 50 minutes long, and students are allowed to bring one two-sided piece of paper (11 inches by 8.5 inches) of notes.

To get a rough sense of the format of the exam, you can see this past exam PDF.

📊 Topics

In scope for the Fall 2024 Midterm Exam:

• Summary Statistics and the Constant Model
• Simple Linear Regression
• Multiple Linear Regression and the Normal Equations
• Matrices, Vectors, and their properties

Topics not in scope for the Midterm Exam, but in scope for the Final Exam:

• Gradient Descent and Convexity
• Foundations of Probability
• Conditional Probability and Naive Bayes
• Combinatorics
• Independence and Conditional Independence
• Bayes’ Theorem

📝 Exams

👋 Feedback: Find an error? Still confused? Have a suggestion? Let us know here.