SIKKIM MANIPAL UNIVERSITY

III Semester B. Tech (AI& DS) Degree

Examination
MATHEMATICAL FOUNDATIONS FOR MACHINE LEARNING
Duration: 3 hrs
AD1305 [CGPA-(21)(3CR)]
Max. Marks: 100
6OEC 2022
Note: Answer allquestions.
Any misprinted and missing data may be suitablyassumed.

Q. Sub
Description Marks
No. Q.No. Allotted
(a) Determine the pivot and free variables of the following system [1+1+4]
2x -3y- 6z- 5u + 2t = 7
z + 3u - 7t =6
u- 2t=1
Solve the echelon system.
(b) Consider f(t) =3t -5 and g(t) = t² in polynomial space P(t) [2+2]
with inner product < f.g >= ff(t)g (t)dt. Find
i) <f,g>
ii)
(c) Using the Gram-Schmidt method, turn the basis B = (b, bz) of [10}
atwo dimensional subspace U R' into an ONB C =
(C1, Cz) of U, where

b,:= and

2 (a) Find the inverse of matrix [6]

2 -1
1
2
(b) Compute the Singular Value Decomposition (SVD)of matrix [14]
2

3 (a) For the functions x(r,0, 0) = rcos(0)sin (Ø), y(r,0, 0) = [8]

rsin(0)sin (Ø), and z(r,6,9) = rcos(0). Calculate the
Jacobian matrix J.
(b) Compute the Hessian of f(x, y, z) = - 2xyz- yo + 3yz' at (8]
the point (1,2,1).
(c) [4]
Find the outer product between the vector u=|2 and v =|-3|

Page 1 of 2
 SIKKIM MANIPAL UNIVERSITY
II Semester B. Tech (AI& DS) Degree Examination
MATHEMATICAL FOUNDATIONS FOR MACHINE LEARNING
Duration: 3 hrs AD1305 [CGPA-(21)(3CR)]
Note: Answer all questions. Max. Marks: 100
Any misprinted and missing datamay be
suitably assumed.
Q. Sub JUL 2023
No. Q.No. Description Marks
1 (a) Allotted
Using Gaussian elimination, find all solutions of the
system Ax = bwith inhomogeneous equation [7]

1 -3 -2|
6
A= 2 -4
-3 6
(b) Are the following set of vectors ffñearly
independent?

(c)
Compute the orthogonal projection of x=onto the line L [6]
u=
spanned by
and find the distance from x to L.
2 (a) Determine the lower triangular matrix used in Cholesky
given matrix factorisation for the [10]
4 10 8
A=10 26 26|
8 26 61J

(b) Compute the eigendecomposition of matrix

[10]

3 (a) For the function u(x,y, z) = 2x + 3y, v(x,y, z) = cos(x) sin(z) and [8]
w(x, y, z) =e*e'e' ,Calculate the Jacobian matrix J.
(b) 1 4.
Find the outer product between the vector u= 2 and v
[4]

(c) Consider the polynomial f(%)= x°.Compute Taylor Polynomial T,(x) [8)
evaluated at x =1.
(a) Suppose the joint PMF is given by the insurance company in the [3+3+4]
accompanying joint probability table:
p(x, y) 100 200

100 .20 .10 .20

250 .O5 .15 .30

i) Findthe marginal PMFs.

ii) Find mean of random variable X and Y i.e. uy and py.
iii) What is 1he covariance between the two random variables X and Y?
Page 1 of 2
 SIKKIM MANIPAL UNIVERSITY
IIISemester B. Tech (Al&DS) Degree Examination
MATHEMATICAL FOUNDATIONS FORMACHINE LEARNING
Duration: 3hrs ADI305 |CGPA-(21) (3CR)|
Max. Marks: 100

Note: Answer all questions.

Any misprinted and missing data may be suitably
assumcd.
Q. Sub
No. Q.No. Description Marks
(a) Allotted
Solve using the Gaussian Elimination method. [6]
x + y- z=-2
2x -y +z =5
(b) Find the rank of the matrix + 2y + 2z =1
-x
[6]
1 2

(c)
4=23 3
5
Assume that
[8]

and v,
Apply Gram-Schmidt
2
determine orthonormalOrthonormalization
basis. Process to
(a) Explain the geometric intuition for the
(b)
eigendecomposition. [6]
Compute the Singular Value Decomposition (SVD) of
matrix [14]

3 (a) For f(x,y, z) = e cos(y) +

row vector J. xy'z,Calculate the Jacobian (6]
(b) Compute the Hessian of f(<,y) = x'- 2xy - y° atthe point
(1,2). [8]
(c) Consider the polynomial f«) = x. Compute Taylor
Polynomial T,(x) evaluated at x = 1. [6]
4 (a) Suppose that X and Y have the following joint probability
mass functions (PMFs): [3+3+4]
y

f(a,y) 1 2
1 0.25 0,25
2 0.25 0.25
i) Find the marginal PMEs.
ii) Find mean of random variable X and Y i.e. u, and
Hy:
ii) What is the covariance between the (wo random
variables X and Y?

(b) Acontinuous random variable X has probability density given [3+3]

by
fr) = e x>0
X<0
Find the )mcan and standard deviation of random variable X
(c) Find the probability hat in successive tosses ola fair die, a3 14|
will conneup for the fiust time on the ilh toss
(a) Explain Gradient Descent (GD) optinization algorithmwith
suitable illustvation
(b) Derive the dual incar program using Lagrange duality from [10]
primal forn of linear program.