ENGG2013 Unit 10

n n determinant and
an application to cryptography
Feb, 2011.

Yesterday A formula
for matrix inverse using cofactors
cofactors

Usually called the adjoint of A

Suppose that det A is nonzero.

Three steps in computing above formula

1. for i,j = 1,2,3, replace each aij by cofactor Cij
2. Take the transpose of the resulting matrix.
3. divide by the determinant of A.
kshum

ENGG2013

Outline

nxn determinant
Caesar Cipher
Modulo arithmetic
Hill Cipher

kshum

ENGG2013

DETERMINANT IN GENERAL

kshum

ENGG2013

A pattern

Arrange the products so that the first

subscripts are in ascending order.
All possible orderings of the second subscripts
appear once and only once.
kshum

ENGG2013

Transposition
A transposition is an exchange of two objects
in a list of objects.
Examples:

ABCD

21453

ACBD

12453

Transposition is another
mathematical term, and is
not the same as matrix tranpose.
kshum

ENGG2013

Another pattern

The sign of each term is closely related to the

number of transpositions required to obtain
the second subscripts, starting from (1,2) for
the 2x2 case or (1,2,3) for the 3x3 case.
kshum

ENGG2013

The sign
Let p(1), p(2), , p(n) be an order of 1,2,,n.
For example p(1)=3, p(2) = 2, p(3)=1 is an ordering
of 1, 2, 3.

Starting from (1,2,,n), if we need an odd no.

of transpositions to get ( p(1), p(2), , p(n) ),
we define the sign of (p(1), p(2),,p(n)) be 1.
Otherwise, if we need an even no. of
transpositions to get ( p(1), p(2), , p(n) ), we
define the sign of (p(1), p(2),,p(n)) be +1.
kshum

ENGG2013

Definition of nn determinant
1

The summation is over all n! possible

orderings p = ( p(1), p(2), , p(n) ) of 1,2,,n.
There are n! terms.

sgn(p) is either +1 or 1, usually called the

signature or signum of p.
http://en.wikipedia.org/wiki/Determinant
kshum

ENGG2013

Properties of determinant
Determinant of nn identity matrix equals 1.
Exchange two rows (or columns) multiply
determinant by 1.
Multiply a row (or a column) by a constant k
multiply the determinant by k.
Add a constant multiple of a row (column) to
another row (column) no change
Additive property as in the 33 and 22 case.
kshum

ENGG2013

10

Cofactor and the adjoint formula

for matrix inverse
Cofactors are defined in a similar way as in the 3x3 case.
The cofactor of the (i,j)-entry of a matrix A, denoted by Cij, is
defined as (1)i+j Aij, where A is the determinant of the submatrix obtained by removing the i-th row and the j-th column.

We have similar expansion along a row or a column (also

called the Laplace expansion) as in the 3x3 case.
The adjoint formula:
transpose

adjoint of A

nxn identity

The formula in this form holds when det A = 0 also

kshum

ENGG2013

11

CAESAR CIPHER

kshum

ENGG2013

12

Caesar and his army

ATTACK

Soldier carrying the

message ATTACK

Message may be intercepted

by enemy
kshum

ENGG2013

13

Caesar cipher
http://en.wikipedia.org/wiki/Caesar_cipher

ATTACK

Soldier carrying the

encrypted message
DWWDFN

The encrypted message

looks random and meaningless
kshum

ENGG2013

14

Private key encryption

Key

Plain text

Plain text

The value of key is kept

secret
Encryption
function

Decryption
function

Ciphertext

Ciphertext

key
kshum

ENGG2013

15

Mathematical description
Caesar cipher is not secure
enough, because the number
of keys is too small.

Key =3

ATTACK

Shift to the right

by 3

ATTACK

Shift to the left

by 3

DWWDFN

DWWDFN

Key = 3
kshum

ENGG2013

16

MODULO ARITHMETIC

kshum

ENGG2013

17

Mod 12
Clock arithmetic

6+8= 2 mod 12
12
1

11
10

5+12 = 5 mod 12
2

4
8
7
kshum

6
ENGG2013

5
18

Mod 7
Week arithmetic

1+9 = 3 mod 7
2+3 = 5 mod 7

0
Sun

Mon

Tue

Wed

Thr

5
Fri

6
Sat
1

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31
6

kshum

ENGG2013

19

Mod 60
http://www.hko.gov.hk/gts/time/stemsandbranchesc.htm

arithmetic
1 2 3 4 5 6 7 8 9 10 11 12


13 14 15 16 17 18 19 20 21 22 23 24


25 26 27 28 29 30 31 32 33 34 35 36

Year of rabbit

37 38 39 40 41 42 43 44 45 46 47 48


49 50 51 52 53 54 55 56 57 58 59 60


kshum

ENGG2013

20

Mod n formal definition

n is a fixed positive integer
Definition: a mod n is the remainder of a after
division by n.
Example: 25 = 1 mod 12.

Addition and multiplication: If the sum or

product of two integers is larger than or equal
to n, divide by n and take the remainder.
Example: 2+10 = 0 mod 12.
Example: 25 = 3 mod 12.
kshum

ENGG2013

21

More examples

10 mod 7 = 3
4+5 mod 7 = 2
6+7 mod 7 = 6
27 mod 7 = 0

kshum

ENGG2013

22

Mod 26
A

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

Fix a one-to-one correspondence between the English alphabets

and the integers mod 26.

Caesars cipher: shifting a letter to the right by 3

is the same as adding 3 in mod 26 arithmetic.
kshum

ENGG2013

23

Examples of mod 26 calculations

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

3+19 = ? mod 26
13+20 = ? mod 26
34 = ? Mod 26
134 = ? Mod 26

kshum

ENGG2013

24

Peculiar phenomena
in modulo arithmetic
Non-zero times non-zero may be zero
49 = 0 mod 12
22 = 0 mod 4

Multiplicative inverse may not exist

Cannot find an integer x such that 4x = 1 mod 12.
4-1 does not exist mod 12.

kshum

ENGG2013

25

No fraction in modulo arithmetic

In mod 12, dont write 1/3 or 3-1 because it
does not exist.
But 5-1 is well-defined mod 12, because we
can solve 5x=1 mod 12.
Indeed, we have 55 = 1 mod 12.
Therefore 5-1 = 5 mod 12.
Fact from number theory:
multiplicative inverse of x mod n exists
if and only the gcd of x and n is 1.
kshum

ENGG2013

Fraction
26

HILL CIPHER

kshum

ENGG2013

27

Hill cipher

http://en.wikipedia.org/wiki/Hill_cipher

Invented by L. S. Hill in 1929.

Inputs : String of English letters, A,B,,Z.
An nn matrix K, with entries drawn from 0,1,,25.
(The matrix K serves as the secret key. )

Divide the input string into blocks of size n.

Identify A=0, B=1, C=2, , Z=25.
Encryption: Multiply each block by K and then
reduce mod 26.
Decryption: multiply each block by the inverse of
K, and reduce mod 26.
kshum

ENGG2013

28

Note
The decryption must be the inverse function of
the encryption function.
It is required that K-1 K = In mod 26.

Provided that det(K) has a multiplicative inverse

mod 26, i.e., if det(K) and n has no common
factor, the inverse of K can be computed by the
adjoint formula for matrix inverse.
Inverse of an integer mod 26 can be obtained by
trial and error.
kshum

ENGG2013

29

Example
Plain text: LOVE, Secret Key:
LO
VE
2, 3, 16, 5 are transformed to cipher text
CDQF
A

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

kshum

ENGG2013

30

How to decode?
Given CDQF, and the encryption matrix
How do we decrypt?
We need to compute the inverse of

Remind that all arithmetic are mod 26. There

is no fraction and care should be taken in
computing multiplicative inverse mod 26.

kshum

ENGG2013

31

Determinant
The determinant of
equals 20(7)-3(15),
which is 17 mod 26.
Find the multiplicative inverse of 17 mod 26,
i.e., find integer x such that 17x = 1 mod 26.
Just try all 26 possibilities for x:
171 = 17 mod 26
172= 8 mod 26
173 = 25 mod 26
174 = 16 mod 26
175 = 7 mod 26
176 = 24 mod 26
177 = 15 mod 26
kshum

178 = 6 mod 26
179= 23 mod 26
1710 = 14 mod 26
1711 = 5 mod 26
1712 = 22 mod 26
1713 = 13 mod 26
1714 = 4 mod 26
ENGG2013

1715 = 21 mod 26
1716= 12 mod 26
1717 = 3 mod 26
1718 = 20 mod 26
1719 = 11 mod 26
1720 = 2 mod 26
1721 = 19 mod 26

1722 = 10 mod 26
1723= 1 mod 26
1724 = 18 mod 26
1725 = 9 mod 26
170 = 0 mod 26

32

Computing the inverse mod 26

From 1723= 1 mod 26, we know that the
multiplicative inverse of 17 mod 26 is 23.
Using the formula for 2 2 matrix inverse

we get

kshum

Replace (17)-1 mod 26 by 23

ENGG2013

33

Decryption
Given the ciphertext CDQF, we decrypt by
multiplying by

From the table in p.23, 11, 14, 21, 4 is LOVE.

kshum

ENGG2013

34