Linear Algebra Quiz
Work through these questions to check your understanding of the key concepts from this course. Each question covers a different topic — take your time and use the hints when you need them.
Exercise
u = [1, 2] and v = [3, −1]. What is u · v?
Need a hint?
Multiply component-wise, then sum.
Exercise
Which vector is orthogonal to [2, 1]?
Need a hint?
Orthogonal means the dot product is zero.
Exercise
A = [[5, 7], [1, 3]]. What is A·e₂, where e₂ = [0, 1]?
Need a hint?
Multiplying by a standard basis vector selects something.
Exercise
A 4×4 matrix has every entry equal to 1. What is its rank?
Need a hint?
How many linearly independent columns are there?
Exercise
What are the eigenvalues of the diagonal matrix diag(3, −2)?
Need a hint?
What does a diagonal matrix do to the standard basis vectors?
Exercise
A 1000×500 matrix is effectively rank 10. Roughly how many numbers does the truncated SVD need to store it, versus 500,000 for the full matrix?
Need a hint?
A rank-k SVD keeps k left vectors, k right vectors, and k singular values.
Exercise
In the ML project loop (data → hypothesis space → objective → optimization → evaluation → feedback), replacing a full weight matrix with a low-rank (SVD / LoRA-style) factorization primarily changes which slot?
Need a hint?
Does it change what data you collect, what the model can express, or how you search for parameters?
Exercise
Transfer test — a technique this course did NOT teach: the Adam optimizer replaces plain gradient descent's fixed step with per-parameter adaptive steps. Which slot of the ML project loop does it modify?
Need a hint?
Adam changes how parameters are updated, not what the model can represent or what 'good' means.