Question 1
Consider a vector such that . Let , where is the identity matrix. Which of the following is true?
is Hermitian and unitary.
is Hermitian but not unitary.
is unitary but not Hermitian.
is neither Hermitian nor unitary.

The IIT Madras BS Machine Learning Foundations (MLF) Quiz 2 paper sat on 12 Apr 2026, in the January 2026 term: 14 questions for 41 marks in 120 minutes. Every question is below with its answer. Take it as a timed mock test to be marked, or read it through first.
Consider a vector such that . Let , where is the identity matrix. Which of the following is true?
is Hermitian and unitary.
is Hermitian but not unitary.
is unitary but not Hermitian.
is neither Hermitian nor unitary.
Correct answer
is Hermitian and unitary.
If is a matrix with eigenvalues and , what is det ? Note: det denotes the determinant.
—
—
—
—
Correct answer
—
Let
be a Hermitian matrix. Which of the following options is correct for the values of and ?
—
—
—
—
Correct answer
—
Let
. Find the condition on and for which the matrix is positive-definite.
and
—
and
and
Correct answer
—
Which of the following represents the Taylor series expansion up to the third degree term of
around ?
—
—
—
—
Correct answer
—
If is a real matrix of shape , which of the following is/are true?
is a positive semi-definite matrix.
is a positive semi-definite matrix.
is positive semi-definite if and only if is a symmetric matrix.
is positive semi-definite if and only if is a square matrix.
—
Correct answers
is a positive semi-definite matrix.
is a positive semi-definite matrix.
Consider the following matrix for :
Which of the following options is/are correct?
is unitary for all real values of other than .
is unitary for .
is unitary for .
If , then all eigenvalues of satisfy .
If is any non-zero real value, then all eigenvalues of satisfy .
Correct answers
is unitary for .
If , then all eigenvalues of satisfy .
Consider a function whose graph is shown below. and correspond to local minima of .
We run two separate instances of gradient descent with the same learning but different initial points. (1) Starting from the point (2) Starting from the point Assume that and are very small positive quantities. Which of the following is/are true if we run gradient descent for a large number of iterations?
Instance (1) will converge to the point
Instance (2) will converge to the point
Both instances (1) and (2) will converge to
Both instances (1) and (2) will converge to
Instance (1) will not converge to , but will diverge to negative infinity
Instance (2) will not converge to , but will diverge to infinity
Correct answers
Instance (1) will converge to the point
Instance (2) will converge to the point
Inscript is a real matrix of shape with the following singular values: . Find the rank of .
A written answer, not marked automatically.
Consider the matrix
. Find the sum of all the singular values of .
A written answer, not marked automatically.
A function is minimized using the gradient descent algorithm. Let denote the value of at which is minimized.Starting from the initial point , one iteration of gradient descent is performed with a learning rate of . If denotes the point obtained after this iteration, find the value of . Enter the answer correct to one decimal place.
A written answer, not marked automatically.
Consider the following dataset in Run PCA on this dataset. Based on the above data, answer the given subquestions.
If
is the covariance matrix of the dataset, find . Enter the nearest integer as your answer.
A written answer, not marked automatically.
Consider the following dataset in Run PCA on this dataset. Based on the above data, answer the given subquestions.
Let be the unit-norm eigenvector corresponding to the smallest eigenvalue of the covariance matrix. Find the variance of the dataset along the direction . Enter the nearest integer as your answer.
A written answer, not marked automatically.
Consider the following dataset in Run PCA on this dataset. Based on the above data, answer the given subquestions.
A written answer, not marked automatically.