Lecture 1: Introduction to Fourier Analysis

p.2, line 16: we now at a new instead of we now look at an new

instead of X C
p.3, line 12: An 1 is missing in the denominator, i.e. <
j
,
k
>=
(
jk
)
n
1

jk
1
p.3, last line: change the font on X, i.e. : X T
p.3, footnote: Maybe, normalize your scalarproduct (is more consistent
with the rest of the script) and write the integral w.r.t. the measure ,
i.e. < f, g >=
1
|X|

xX
f(x)g(x) and < f, g >=
1
(X)
_
X
f(x)g(x)d(x)
p.5, Theorem 1.2: clarify Let f : T C b a continuous function ... for
every t T at which f is continuous
p.6, Proposition 1.5: correct the denominator, i.e.
1
n + 1
_
sin
n+1
2
s
sin
s
2
_2
Lecture 2: Introduction to Some Convergence
Theorems
p.8, fejer kernel property 2: normalize
1
2

_
T
k
n
dt = 1
p.9, normalized measure ... state that (G) = 1 explicitly
p.11, secondlastline of this proof should be
_
b
a
|h(x)(P(x) h(x)|dx
c |b a|
p.11, end of the proof of Theorem 2.4: You want to say that h 0 in
L
2
[a, b] and not that h = 0 pointwise. Right?
p.11, Theorem 2.5: normalization factor and formulation, i.e.
lim
n
1
n
n

r=1
f(2r) =
1
2
_
T
f(t)dt
Perhaps change also the domain of your function f appropriate.
p.12, line 1: It is also used...
p.12, line -3: continuous function f such that instead of continuous
function such that
1
p.13, Denition 2.3: Write for example

p
=
_
(x
i
)
iN
:

iN
|x
i
|
p
<
_
p.13, line -3: third SUM should have a factor 2 infront and secondlast e
i
should be e
j
2

< f, e
i
>< f
n

j=1
< f, e
k
> e
j
, e
i
> 2
n

i
< f
n

j=1
< f, e
j
> e
j
, e
i
>
p.13, line -2: Replace innter by inner
Lecture 3: Harmonic Analysis on the Cube and
Parsevals Identity
p.16, line 3 from the bottom: Weierstrass instead of Weierstrauss
p.17, just before Theorem 3.3: f should be continuous as well
p.17, Theorem 3.3: It is not clear where this theorem comes from. I think
you stated here the theorem without a proof. Is this Dirichlets Theorem?
Section 3.3: The notation of the 2-Norm is not always the same. It would
be clearer if there would always be
2
.
p.20, equation (3.4): maybe more space in between the formulas (around
the comma)
p.20, second line of proof of Theorem 3.11: write < f, f > instead of ||f||
2
2
just as before and afterwards
p.21, proof of Corollary 3.13 second line: it is not immediately clear from
where the factor 2 comes, maybe still sum over 1 |r| n, and change
in next line
p.22, line -5, C is a circle, new line for this
Lecture 4: Applications of the Harmonic Analysis
p.24: The way you write Parsevals Identity is a bit misleading. It is not
clear at rst view where the normalization factors are. Perhaps you can
write these norms and scalarproducts out as sums.
2
p.24, proof of Corollary 4.2: mention something about the scalar product
(real-, complex-valued), because < f + g, f + g >=< f, f > + < f, g >
+ < g, f > + < g, g > and in complex case < f, g > might not be the
same as < g, f >.
p.25, line 4, in the integral
_
g(u)ds the terme
iru
is missing and integra-
tion should be w.r.t u
p.25f, Hurwitz proof: I get in most of the calculations the other sign as in
the script, e.g. p.25, line -1, should be a +n
2
not n
2
(...
p.26, equation (4.2): The second arguments in the scalarprodutcs in the
rst line should be conjugate and analog for the second line.
p.28, MacWilliams identity: factor
1
2
n
is missing
p.29, also factors
1
2
n
missing in last and thirdlastline of proof
p.30, line 8, matrix dimensions are n k not k n
p.30, line 11, Variable cannot be zero i.e. write as for a given x
{0, 1}
n
, x = 0
p.30, secondlastline of proof: |Ax| < , and in the last exponent there is
something strange with the brackets, in the exponent there is still an o(n)
not 1; however than why is Pr[..] 2
negative+o(n)
< 1?
p.30, line -1, sum over
_
C
2
_
rather than C C
p.31, middle, where z
i
is the number of WORDS IN C WITH zeros in
the...
p.32, title of section 4.4: Change the name to Erd os
p.32, line -10, intersecting instead of itersecting
Lecture 5: Isoperimetric Problems
p.34, line 8: The edge problem ... |S| = k (instead of R), how small
CAN e(S, S

) be?
p. 34: The denition of shadow is strange. I think you want the following
(f) =
_
y 2
[n]
|x f : x y
_
p.36, line 7: We need to know

1
L
j
(T) if |T| = t.
p.36, line 8,9: The factor
1
2
n
should be on the right hand side.
p.36, line 10: The left hand side should be K
(n)
j
(t).
3
p.37, line 8: Write 1
L
p
instead of 1
p
and the same for q.
p.37, line 9,10: In the Parsevals identity you have also a factor 2
n
which
you should add somewhere to get a correct statement.
p.37, Lemma 5.1: The Krawtchouk polynomials satises a 3-term recur-
rence (the other statement is Theorem 5.2)
Lecture 6: MRRW Bound and Isoperimetric Prob-
lems
p.39, line 2: denition of f; factor missing: f = 2
n

gg
|C|
p.39, line 1,2 from bottom: Add on the left hand side a factor 2
n
. In line
2 from the bottom the summation should go up to n instead of r.
p.40, proof of fact 6.1: The summation should go up to n.
p.40, line -1, sums from k = 0 to n, instead of k = 1, since there are n+1
constraints
p.41, line 1: Let (x) :=

n
k=0

k
K
k
(x).
p.41, line 3: There should be a (0) instead of a (x) on the right hand
side.
p.41, line 9, sum from k = 0 and 3 lines later the same
Lecture 7: The Brunn-Minkowski Theorem and
Inuences of Boolean Variables
p.50, dictatorship inuence: 1 if 1 S instead of i S, same next line
p.51, line 5: The power should be
n
b
1.
p.51, line 8: The fraction should be
1
2
b
1
. And in the big-omega we should
write
log n
n
.
p.51, line 2 from the bottom: The inuence of the x variable in the function
f is equal to the number of mixed edges in x-direction divided by 2
n1
.
p.52, line 2, f(T) missing on the right side of the equation
p.52, last line: The right hand side and the middle term should be divided
by 2
n
.
p.53, line 9: Write

f instead of hatf.
p.53, line -2, Inf
f
(()S) too many brackets
4
Lecture 8: More on the inuence of the variables
on boolean function
p.55, Theorem 8.1: f : {1, 1}
n
... the n is missing
p.58, line 2: say that proof starts here
p.60, line 8: A factor 2 is missing, i.e. for i S it holds that

f(i)(S) =
2

f(S).
p.60, middle, ignore the 0 term because f() = 0 instead of what is written
p.60, line 10,17,18,19,20,21: A factor 4 is missing.
p.62, line 6: Friedgut instead of Freidgut
p.63, line -5, also Friedguts instead of Freidguts
p.70, line -5, again Friedgut
Lecture 9: Threshold Phenomena
(found nothing yet)
5