Title: Coding%20Theory,%20Card%20Tricks%20and%20Hat%20Problems
1Coding Theory, Card Tricks and Hat Problems
Ho Chi Minh City, September 30, 2007
- Michel Waldschmidt
- Université P. et M. Curie - Paris VI
- Centre International de Mathématiques Pures et
Appliquées - CIMPA
http//www.math.jussieu.fr/miw/coursHCMUNS2007.ht
ml
2Ho Chi Minh City
September 30, 2007
Coding Theory, Card Tricks and Hat Problems
Starting with card tricks, we show how
mathematical tools are used to detect and to
correct errors occuring in the transmission of
data. These so-called "error-detecting codes"
and "error-correcting codes" enable
identification and correction of the errors
caused by noise or other impairments during
transmission from the transmitter to the
receiver. They are used in compact disks to
correct errors caused by scratches, in satellite
broadcasting, in digital money transfers, in
telephone connexions, they are useful for
improving the reliability of data storage media
as well as to correct errors cause when a hard
drive fails. The National Aeronautics and Space
Administration (NASA) has used many different
error-correcting codes for deep-space
telecommunications and orbital missions.
http//www.math.jussieu.fr/miw/
3Ho Chi Minh City
September 30, 2007
Coding Theory, Card Tricks and Hat Problems
Most of the theory arises from earlier
developments of mathematics which were far
removed from any concrete application. One of the
main tools is the theory of finite fields, which
was invented by Galois in the XIXth century, for
solving polynomial equations by means of
radicals. The first error-correcting code
happened to occur in a sport newspaper in Finland
in 1930. The mathematical theory of information
was created half a century ago by Claude Shannon.
The mathematics behind these technical devices
are being developped in a number of scientific
centers all around the world, including in
Vietnam and in France.
http//www.math.jussieu.fr/miw/
4I know which card you selected
- Among a collection of playing cards, you select
one without telling me which one it is. - I ask you some questions where you answer yes or
no. - Then I am able to tell you which card you
selected.
52 cards
- You select one of these two cards
- I ask you one question and you answer yes or no.
- I am able to tell you which card you selected.
62 cards one question suffices
74 cards
8First question is-it one of these two?
9Second question is-it one of these two ?
104 cards 2 questions suffice
Y Y
Y N
N Y
N N
118 Cards
12First question is-it one of these?
13Second question is-it one of these?
14Third question is-it one of these?
158 Cards 3 questions
YYY
YYN
YNY
YNN
NYY
NYN
NNY
NNN
16Yes / No
- 0 / 1
- Yin / Yang - -
- True / False
- White / Black
- / -
- Heads / Tails (tossing a coin)
173 questions, 8 solutions
0 0 0
0 0 1
0 1 0
0 1 1
1 0 0
1 0 1
1 1 0
1 1 1
1816 Cards
- If you select one card among a set of 16, I
shall know which one it is, once you answer my 4
questions by yes or no.
19(No Transcript)
20Label the 16 cards
21In binary expansion
22Ask the questions so that the answers are
23The 4 questions
- Is the first digit 0 ?
- Is the second digit 0 ?
- Is the third digit 0 ?
- Is the fourth digit 0 ?
24More difficult
25One answer may be wrong
- Consider the same problem, but you are allowed to
give (at most) one wrong answer. - How many questions are required so that I am able
to know whether your answers are right or not?
And if they are right, to know the card you
selected?
26Detecting one mistake
- If I ask one more question, I shall be able to
detect if there is one of your answers which is
not compatible with the others. - And if you made no mistake, I shall tell you
which is the card you selected.
27Detecting one mistake with 2 cards
- With two cards I just repeat twice the same
question. - If both your answers are the same, you did not
lie and I know which card you selected - If your answers are not the same, I know that one
is right and one is wrong (but I dont know which
one is correct!).
284 cards
29First question is-it one of these two?
30Second question is-it one of these two?
31Third question is-it one of these two?
324 cards 3 questions
Y Y Y
Y N N
N Y N
N N Y
334 cards 3 questions
0 0 0
0 1 1
1 0 1
1 1 0
34Correct triple of answers
Wrong triple of answers
One change in a correct triple of answers yields
a wrong triple of answers
35Boolean addition
- even even even
- even odd odd
- odd even odd
- odd odd even
36Parity bit
- Use one more bit which is the Boolean sum of the
previous ones. - Now for a correct answer the sum of the bits
should be 0. - If there is exactly one error, the parity bit
will detect it the sum of the bits will be 1
instead of 0.
378 Cards
384 questions for 8 cards
Use the 3 previous questions plus the parity bit
question
39First question is-it one of these?
40Second question is-it one of these?
41Third question is-it one of these?
42Fourth question is-it one of these?
4316 cards, at most one wrong answer 5 questions
to detect the mistake
44Ask the 5 questions so that the answers are
45Correcting one mistake
- Again I ask you questions where your answer is
yes or no, again you are allowed to give at most
one wrong answer, but now I want to be able to
know which card you selected - and also to tell
you whether and when you lied.
46With 2 cards
- I repeat the same question three times.
- The most frequent answer is the right one vote
with the majority.
47With 4 cards
I repeat my two questions three times each,
which makes 6 questions
48With 8 Cards
I repeat 3 times my 3 questions, which makes 9
questions
49With 16 cards, this process requires 3?412
questions
5016 cards, 7 questions
- Recall if you select one card among 16, only 3
questions are required if all answers are correct.
- With 4 questions, we detect one mistake.
- We shall see that 7 questions allow to recover
the exact result if there is at most one wrong
answer.
51The Hat Problem
52The Hat Problem
- Three people are in a room, each has a hat on his
head, the color of which is black or white. Hat
colors are chosen randomly. Everybody sees the
color of the hat on everyones head, but not on
their own. People do not communicate with each
other. - Everyone gets to guess (by writing on a piece of
paper) the color of their hat. They may write
Black/White/Abstain.
53- The people in the room win together or lose
together. - The team wins if at least one of the three people
did not abstain, and everyone who did not abstain
guessed the color of their hat correctly. - How will this team decide a good strategy with a
high probability of winning?
54- Simple strategy they agree that two of them
abstain and the other guesses randomly. - Probability of winning 1/2.
- Is it possible to do better?
55- Hint
- Improve the odds by using the available
information everybody sees the color of the hat
on everyones head but himself.
56- Better strategy if a member sees two different
colors, he abstains. If he sees the same color
twice, he guesses that his hat has the other
color.
57- The two people with white hats see one white
hat and one black hat, so they abstain.
The one with a black hat sees two white hats,
so he writes black.
They win!
58- The two people with black hats see one white
hat and one black hat, so they abstain.
The one with a white hat sees two black hats,
so he writes white.
They win!
59 Everybody sees two white hats, and therefore
writes black on the paper.
60 Everybody sees two black hats, and therefore
writes white on the paper.
61two white or two black
62three white or three black
63- The team wins exactly when the three hats do not
have all the same color, that is in 6 cases out
of a total of 8 - Probability of winning 3/4.
64- Now if the hats are no more distributed randomly
and if the master of the game knows the previous
strategy, he may select often the three hats all
of the same color.
- Is there another strategy giving the same
probability 3/4, for randomly distributed colors?
Answer YES!
65- The team selects one distribution of hats and
bets that it will not be this one,
neither the opposite distribution where black
is replaced by white and white by black
66- If the distribution of the colors is one of
these, they will lose.
- If the distribution of the colors is one of the 6
other possibilities, exactly one of the three
people will be able to guess correctly the color
of his hat, and the two others will abstain, so
the team will win.
67Hat problem with 7 people
For 7 people in the room in place of 3, which is
the best strategy and its probability of
winning?
Answer the best strategy gives a probability
of winning of 7/8
68Winning at the lottery
69Tails and Ends
- Toss a coin 3 consecutive times
- There are 8 possible sequences of results
- (0,0,0), (0,0,1), (0,1,0), (0,1,1),
- (1,0,0), (1,0,1), (1,1,0), (1,1,1).
70If you bet (0,1,0), you have
- All three correct results for (0,1,0).
- Exactly two correct results if the sequence is
either (0,1,1), (0,0,0) or (1,1,0), - Exactly one correct result if the sequence is
either (0,0,1), (1,1,1) or (1,0,0), - No correct result at all for (1,0,1).
71Among 8 possibilities
- each bet
- is winning in exactly 1 case
- has exactly two correct results in 3 cases
- has exactly one correct result in 3 cases
- has no correct result at all in only 1 case
72- Goal To be sure of having at least 2 correct
results - Clearly, one bet is not sufficient
- Are two bets sufficient?
- Recall that there are 8 possible results, and
that each bet has at least 2 correct results in 4
cases.
73Answer YES, two bets
suffice!
- For instance bet
- (0,0,0) and (1,1,1)
- Whatever the result is, one of the two digits
- 0 and 1
- occurs more than once.
- Hence one (and only one) of the two bets
- has at least two correct results.
74- The 8 sequences of three digits
- 0 and 1
- split into two groups
- 4 having two or three 0s
- 000, 001, 010, 100
- and
- 4 having two or three 1s
- 111, 110, 101, 011
-
75Other solutions
- Select any two bets with all three different
digits, say - 1 0 1 and 0 1 0
- The result either is one of these, or else has
two common digits with one of these and just one
common digit with the other.
76- Select 101 and 010
- The 8 sequences of three digits
- 0 and 1
- split into two groups
- 4 having two or three digits which
are the same as 101 - 101, 100, 111, 001
- and
- 4 having two or three digits which are the
same as 010 - 010, 011, 000, 110
77Toss a coin 7 consecutive times
- There are 27 128 possible sequences of results.
- How many bets are required in such a way that you
are sure one at least of them has at most one
wrong answer?
78Tossing a coin 7 times
- Each bet has all correct answers once every 128
cases. - It has just one wrong answer 7 times either the
first, second, seventh guess is wrong. - So it has at most one wrong answer 8 times among
128 possibilities.
79Tossing a coin 7 times
- Now 128 8 ? 16. Indeed 128 is 27, and 8 is 23.
27 23 ? 24 - Therefore you cannot achieve your goal with less
than 16 bets. - Coding theory tells you how to select your 16
bets, exactly one of them will have at most one
wrong answer.
80SPORT TOTO
- A match between two players (or teams) may give
three possible results either player 1 wins, or
player 2 wins, or else there is a draw. - There is a lottery, and a winning ticket needs to
have at least three correct bets. How many
tickets should one buy to be sure to win?
814 matches, 3 correct forecasts
- For 4 matches, there are 34 81 possibilities.
- A bet on 4 matches is a sequence of 4 symbols
1, 2, x. Each such ticket has exactly 3 correct
answers 8 times. For instance the ticket
(1, 2, x, 1) has 3 correct answers for the
results - (2, 2, x, 1), (1, 1, x, 1), (1, 2, 1, 1), (1, 2,
x, 2) - (x, 2, x, 1), (1, x, x, 1), (1, 2, 2, 1), (1, 2,
x, x)
824 matches, 3 correct forecasts
- Each ticket has all correct answers once and 3
correct answers 8 times - Hence it has at least 3 correct answers in 9
cases. - Altogether there are 81 9 ? 9 possible results.
- So at least 9 tickets are required to be sure to
win. - How to select such 9 tickets?
839 tickets (in columns)
- x x x 1 1 1 2 2
2 - x 1 2 x 1 2 x
1 2 - x 1 2 1 2 x 2 x
1 - x 1 2 2 x 1 1 2
x
This is an error correcting code on the
alphabet 1,2,x with rate 1/2
84Error correcting codes
85Coding Theory
- Coding theory is the branch of mathematics
concerned with transmitting data across noisy
channels and recovering the message. Coding
theory is about making messages easy to read
don't confuse it with cryptography which is the
art of making messages hard to read!
86Claude Shannon
- In 1948, Claude Shannon, working at Bell
Laboratories in the USA, inaugurated the whole
subject of coding theory by showing that it was
possible to encode messages in such a way that
the number of extra bits transmitted was as small
as possible. Unfortunately his proof did not give
any explicit recipes for these optimal codes.
87Richard Hamming
- Around the same time, Richard Hamming, also at
Bell Labs, was using machines with lamps and
relays having an error detecting code. The digits
from 1 to 9 were send on ramps of 5 lamps with
two on and three out. There were very frequent
errors which were easy to detect and then one had
to restart the process.
88The first correcting codes
- For his researches, Hamming was allowed to have
the machine working during the week-end only, and
they were on the automatic mode. At each error
the machine stopped until the next monday
morning. - "If it can detect the error," complained
Hamming, "why can't it correct it! "
89The origin of Hammings code
- He decided to find a device so that the machine
would not only detect the errors but also correct
them. - In 1950, he published details of his work on
explicit error-correcting codes with information
transmission rates more efficient than simple
repetition. - His first attempt produced a code in which four
data bits were followed by three check bits which
allowed not only the detection but the correction
of a single error.
90(No Transcript)
91Codes and Geometry
- 1949 Marcel Golay (specialist of radars)
produced two remarkably efficient codes. - Eruptions on Io (Jupiters volcanic moon)
- 1963 John Leech uses Golays ideas for sphere
packing in dimension 24 - classification of
finite simple groups - 1971 no other perfect code than the two found by
Golay.
92Error Correcting Codes Data Transmission
- Telephone
- CD or DVD
- Image transmission
- Sending information through the Internet
- Radio control of satellites
93Applications of error correcting codes
- Transmitions by satellites
- Compact discs
- Cellular phones
94- Olympus Mons on Mars Planet
- Image from Mariner 2 in 1971.
95- Information was sent to the earth using an
error correcting code which corrected 7 bits
among 32. - In each group of 32 bits, 26 are control bits
and the 6 others contain the information.
96- Between 1969 and 1973 the NASA Mariner probes
used a powerful Reed--Muller code capable of
correcting 7 errors out of 32 bits transmitted,
consisting now of 6 data bits and 26 check bits!
Over 16,000 bits per second were relayed back to
Earth.
The North polar cap of Mars, taken by Mariner 9
in 1972.
97Voyager 1 and 2 (1977)
- Journey Cape Canaveral, Jupiter, Saturn, Uranus,
Neptune. - Sent information by means of a binary code which
corrected 3 errors on words of length 24.
98Mariner spacecraft 9 (1979)
- Sent black and white photographs of Mars
- Grid of 600 by 600, each pixel being assigned one
of 64 brightness levels - Reed-Muller code with 64 words of 32 letters,
minimal distance 16, correcting 7 errors, rate
3/16
99Voyager (1979-81)
- Color photos of Jupiter and Saturn
- Golay code with 4096212 words of 24 letters,
minimal distance 8, corrects 3 errors, rate 1/2. - 1998 lost of control of Soho satellite recovered
thanks to double correction by turbo code.
100NASA's Pathfinder mission on Mars
- The power of the radio transmitters on these
craft is only a few watts, yet this information
is reliably transmitted across hundreds of
millions of miles without being completely
swamped by noise.
Sojourner rover and Mars Pathfinder lander
101Listening to a CD
- On a CD as well as on a computer, each sound is
coded by a sequence of 0s and 1s, grouped in
octets - Further octets are added which detect and correct
small mistakes. - In a CD, two codes join forces and manage to
handle situations with vast number of errors.
102Coding the sound on a CD
On CDs the signal in encoded digitally. To guard
against scratches, cracks and similar damage,
two "interleaved" codes which can correct up to
4,000 consecutive errors (about 2.5 mm of track)
are used.
- Using a finite field with 256 elements, it is
possible to correct 2 errors in each word of 32
octets with 4 control octets for 28 information
octets.
103A CD of high quality may have more than 500
000 errors!
- After processing of the signal in the CD player,
these errors do not lead to any disturbing noise. - Without error-correcting codes, there would be no
CD.
1041 second of radio signal 1 411 200 bits
- The mathematical theory of error correcting codes
provides more reliability and at the same time
decreases the cost. It is used also for data
transmission via the internet or satellites
105Codes and Mathematics
- Algebra
- (discrete mathematics finite fields, linear
algebra,) - Geometry
- Probability and statistics
106Finite fields and coding theory
- Solving algebraic equations with
radicals Finite fields theory
Evariste Galois
(1811-1832) - Construction of regular polygons with rule and
compass - Group theory
107Coding Theory
transmission
Source Coded Text Coded Text Receiver
108Error correcting codes
Transmission with noise
Source Coded Text Coded Text Receiver
109 Principle of coding theory
- Only certain words are allowed (code
dictionary of valid words). - The useful letters carry the information,
the other ones (control bits) allow detecting or
correcting errors.
110Detecting one error by sending twice the message
- Send twice each bit
- 2 code words among 422 possible words
- (1 useful letter among 2)
- Code words
- (two letters)
- 0 0
- and
- 1 1
- Rate 1/2
111- Principle of codes detecting one error
-
- Two distinct code words
- have at least two distinct letters
-
112Detecting one error with the parity bit
- Code words (three letters)
- 0 0 0
- 0 1 1
- 1 0 1
- 1 1 0
- Parity bit (x y z) with zxy.
- 4 code words (among 8 words with 3 letters),
- 2 useful letters (among 3).
- Rate 2/3
2
113Code Words Non Code Words
- 0 0 0 0 0 1
- 0 1 1 0 1 0
- 1 0 1 1 0 0
- 1 1 0 1 1 1
- Two distinct code words
- have at least two distinct letters.
2
114Check bit
- In the International Standard Book Number (ISBN)
system used to identify books, the last of the
ten-digit number is a check bit. - The Chemical Abstracts Service (CAS) method of
identifying chemical compounds, the United States
Postal Service (USPS) use check digits. - Modems, computer memory chips compute checksums.
- One or more check digits are commonly embedded in
credit card numbers.
115Correcting one errorby repeating three times
- Code words
- (three letters)
- 0 0 0
- 1 1 1
- Rate 1/3
- Send each bit three times
- 2 code words
- among 8 possible ones
- (1 useful letter among 3)
116- Correct 0 0 1 as 0 0 0
- 0 1 0 as 0 0 0
- 1 0 0 as 0 0 0
- and
- 1 1 0 as 1 1 1
- 1 0 1 as 1 1 1
- 0 1 1 as 1 1 1
117- Principle of codes correcting one error
-
- Two distinct code words have at least three
distinct letters -
118Hamming Distance between two words
- number of places in which the two words
- differ
- Examples
- (0,0,1) and (0,0,0) have distance 1
- (1,0,1) and (1,1,0) have distance 2
- (0,0,1) and (1,1,0) have distance 3
- Richard W. Hamming (1915-1998)
119Hamming distance 1
120Hammings unit sphere
- The unit sphere around a word includes the words
at distance at most 1
121At most one error
122Words at distance at least 3
123Decoding
124Hamming Distance
- The Hamming distance between two words is the
number of places where letters differ. - A code detects n errors iff the Hamming distance
between two distinct code words is at least 2n. - It corrects n errors iff the Hamming distance
between two distinct code words is at least
2n1.
125The code (0 0 0) (1 1 1)
- The set of words with three letters (eight
elements) splits into two balls - The centers are (0,0,0) and (1,1,1)
- Each of the two balls consists of elements at
distance at most 1 from the center
126Two or three 0
Two or three 1
- (0,0,1)
- (0,0,0) (0,1,0)
- (1,0,0)
- (1,1,0)
- (1,1,1) (1,0,1)
- (0,1,1)
127Another error correcting code on 3 letters
- Select two words with mutual distance 3
two words with three distinct letters, say
(0,0,1) and (1,1,0) - For each of them, consider the ball of elements
at distance at most 1
128- (0,0,0)
- (0,0,1) (0,1,1)
- (1,0,1)
- (1,1,1)
- (1,1,0) (1,0,0)
- (0,1,0)
129Back to the Hat Problem
- Replace white by 0 and black by 1
- hence the distribution of colors becomes a
word of three letters on the alphabet 0 , 1 - Consider the centers of the balls (0,0,0) and
(1,1,1). - The team bets that the distribution of colors is
not one of the two centers.
130Assume the distribution of hats does not
correspond to one of the centers (0, 0, 0) and
(1, 1, 1). Then
- One color occurs exactly twice (the word has both
digits 0 and 1). - Exactly one member of the team sees twice the
same color this corresponds to 0 0 in case he
sees two white hats, 1 1 in case he sees two
black hats. - Hence he knows the center of the ball (0 , 0 ,
0) in the first case, (1, 1, 1) in the second
case. - He bets the missing digit does not yield the
center.
131- The two others see two different colors, hence
they do not know the center of the ball. They
abstain. - Therefore the team wins when the distribution of
colors does not correspond to one of the centers
of the balls. - This is why the team wins in 6 cases.
132- Now if the word corresponding to the distribution
of the hats is one of the centers, all members of
the team bet the wrong answer! - They lose in 2 cases.
133Another strategy
- Select two words with mutual distance 3
two words with three distinct letters, say
(0,0,1) and (1,1,0) - For each of them, consider the ball of elements
at distance at most 1
134- The team bets that the distribution of colors is
not one of the two centers (0,0,1), (1,1,0). - A word not in the center has exactly one letter
distinct from the center of its ball, and two
letters different from the other center.
135Assume the word corresponding to the distribution
of the hats is not a center. Then
- Exactly one member of the team knows the center
of the ball. He bets the distribution does not
correspond to the center. - The others do not know the center of the ball.
They abstain. - Hence the team wins.
136The binary code of Hamming (1950)
-
- It is a linear code (the sum of two code words
is a code word) and the 16 balls of radius 1 with
centers in the code words cover all the space of
the 128 binary words of length 7 - (each word has 7 neighbors (71)?16 256).
137Hamming code
- Words of 7 letters
- Code words (1624 among 12827)
- (a b c d e f g)
- with
- eabd
- facd
- gabc
- Rate 4/7
138Hammings code
139How to compute e , f , g from a , b , c , d.
eabd
d
a
b
facd
c
gabc
14016 code words of 7 letters
- 0 0 0 0 0 0 0
- 0 0 0 1 1 1 0
- 0 0 1 0 0 1 1
- 0 0 1 1 1 0 1
- 0 1 0 0 1 0 1
- 0 1 0 1 0 1 1
- 0 1 1 0 1 1 0
- 0 1 1 1 0 0 0
- 1 0 0 0 1 1 1
- 1 0 0 1 0 0 1
- 1 0 1 0 1 0 0
- 1 0 1 1 0 1 0
- 1 1 0 0 0 1 0
- 1 1 0 1 1 0 0
- 1 1 1 0 0 0 1
- 1 1 1 1 1 1 1
Two distinct code words have at least three
distinct letters
141The Hat Problem with 7 people
- The team bets that the distribution of the hats
does not correspond to the 16 elements of the
Hamming code - Loses in 16 cases (they all fail)
- Wins in 128-16112 cases (one bets correctly, the
6 others abstain) - Probability of winning 112/1287/8
142Tossing a coin 7 times
- There are 128 possible results
- Each bet is a word of 7 letters on the alphabet
0, 1 - How many bets do you need if you want to
guarantee at least 6 correct results?
143- Each of the 16 code words has 7 neighbors (at
distance 1), hence the ball of which it is the
center has 8 elements. - Each of the 128 words is in exactly one of these
balls.
144- Make 16 bets corresponding to the 16 code words
then, whatever the final result is, exactly one
of your bets will have at least 6 correct
answers. - Notice that 16/1281/8. This is the probability
of losing for the hat problem with 7 people using
the Hamming code.
1457 questions to find the selected card among 16,
with one possible wrong answer
- Replace the cards by labels from 0 to 15 and
write the binary expansions of these - 0000, 0001, 0010, 0011
- 0100, 0101, 0110, 0111
- 1000, 1001, 1010, 1011
- 1100, 1101, 1110, 1111
1467 questions to find the selected number in
0,1,2,,15 with one possible wrong answer
- Is the first binary digit 1?
- Is the second binary digit 1?
- Is the third binary digit 1?
- Is the fourth binary digit 1?
- Is the number in 1,2,4,7,9,10,12,15?
- Is the number in 1,2,5,6,8,11,12,15?
- Is the number in 1,3,4,6,8,10,13,15?
147Sphere Packing
- While Shannon and Hamming were working on
information transmission in the States, John
Leech invented similar codes while working on
Group Theory at Cambridge. This research included
work on the sphere packing problem and culminated
in the remarkable, 24-dimensional Leech lattice,
the study of which was a key element in the
programme to understand and classify finite
symmetry groups.
148Sphere packing
The kissing number is 12
149Sphere Packing
- Kepler Problem maximal density of a packing of
identical sphères - p / Ö 18 0.740 480 49
- Conjectured in 1611.
- Proved in 1999 by Thomas Hales.
- Connections with crystallography.
150Current trends
- In the past two years the goal of finding
explicit codes which reach the limits predicted
by Shannon's original work has been achieved. The
constructions require techniques from a
surprisingly wide range of pure mathematics
linear algebra, the theory of fields and
algebraic geometry all play a vital role. Not
only has coding theory helped to solve problems
of vital importance in the world outside
mathematics, it has enriched other branches of
mathematics, with new problems as well as new
solutions.
151Research going on in France
- INRIA Rocquencourt, Projet Codes
http//www-rocq.inria.fr/codes/ - Marseille, Institut de Mathématiques de Luminy
iml.univ-mrs.fr - Limoges http//www.xlim.fr
- Toulon Groupe de Recherche en Informatique et
Mathématiques GRIM http//grim.univ-tln.fr
152Research going on in France
- Bordeaux Institut de Mathématiques de Bordeaux
A2X Théorie des nombres et Algorithmique
Arithmétique - Equipe Codes et Réseaux
- http//www.math.u-bordeaux1.fr/maths/
spip.php?article9
153Research going on in France
- Toulouse Groupe de Recherche en Informatique et
Mathématiques du Mirail (GRIMM) - http//www.univ-tlse2.fr/LB028/0/
- fiche___laboratoire/RH02Equipes
- ENST (Télécom) Bretagne
- http//departements.enst-bretagne.fr/sc/recherche
/turbo/
154GDR IM ''Informatique Mathématique''
- Groupe de travail C2 Cryptographie et Codage
- http//www.gdr-im.fr/
- http//www-rocq.inria.fr/codes/Claude.Carlet/c
omp.html
155Directions of research
- Theoretical questions of existence of specific
codes - connection with cryptography
- lattices and combinatoric designs
- algebraic geometry over finite fields
- equations over finite fields
156A last card trick
- The best card trick
- Michael Kleber, Mathematical Intelligencer 24
(2002)
157A card trick
- Among 52 cards, select 5 of them, do not show
them to me, but give them to my assistant. - After looking at these 5 cards, my assistant
gives me four of them, one at a time. - There is one card left which is known to you and
to my assistant. - I am able to tell you which one it is.
158Which information did I receive?
- I received 4 cards, one at a time. With my
assistant we agreed beforehand with an ordering. - I receive these 4 cards in one of 24 possible
orders. - There are 4 choices for the first card, once the
first card is selected there are 3 choices for
the second card, and then 2 choices for the third
one. And finally no choice for the last one. - 24 4 ? 3 ? 2 ? 1
15924 possible orders for 4 cards
- I receive these 4 cards in one of the 24
following orders - 1234, 1243, 1324, 1342, 1423, 1432,
- 2134, 2143, 2314, 2341, 2413, 2431,
- 3124, 3142, 3214, 3241, 3412, 3421,
- 4123, 4132, 4213, 4231, 4312, 4321.
- So the information I receive can be converted
into a number from 1 to 24
160But there are 52 cards!
- Therefore this idea is not sufficient.
- There are four colors
- Spade, Hart, Diamond, Club
My assistant received 5 cards.
161Pigeonhole or Box Principle
- If there are more pigeons than holes, one at
least of the holes hosts at least two pigeons. - If there are more holes than pigeons, one at
least of the holes is empty.
My assistant received 5 cards, there are 4
colours So one at least of the colors occurs
twice.
We agree that the first card I receive will tell
me the colour of the hidden card.
162Information I receive from the next 3 cards
- Once I receive the first card, I know the color
of the hidden card. I need to find out which one
it is among the 12 others of the same color. - Next I receive 3 cards in one of the 6 possible
orders. I convert this information into a number
from 1 to 6. - My assistant had the choice between two cards for
the first I received.
163Count from 1 to 6
164Coding Theory, Card Tricks and Hat Problems
Ho Chi Minh City, September 30, 2007
- Michel Waldschmidt
- Université P. et M. Curie - Paris VI
- Centre International de Mathématiques Pures et
Appliquées - CIMPA
http//www.math.jussieu.fr/miw/coursHCMUNS2007.ht
ml