SlideShare a Scribd company logo
1 of 53
Download to read offline
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Progress on Exceptional APN Functions and
Absolute Irreducibility of Infinite Classes of
Multivariate Polynomials
Heeralal Janwa
(Joint work with Moises Delgado)
IWSDA2015, Bengluru, India
18 September, 2015
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Definition 1
Let L = Fq, with q = pn. f : L → L is almost perfect nonlinear
(APN) on L if for all a, b ∈ L, a = 0
f (x + a) − f (x) = b (1)
has at most 2 solutions.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Equivalently, for p = 2, the cardinality of {f (x + a)−f (x) : x ∈ L}
is at least 2n−1 for each a ∈ L∗. APN functions are important in
applications to cryptography, coding theory, combinatorics, finite
geometries, and other related areas.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The best known examples of APN functions are
xd Exponent d Constraints
Gold 2r + 1 (r, n) = 1
Kasami-Welch 22r − 2r + 1 (r, n) = 1,n odd
Welch 2r + 3 n = 2r + 1
Niho 2r + 2r/2 − 1 n = 2r + 1, r even
2r + 2(3r+1)/2 − 1 n = 2r + 1, r odd
Inverse 22r − 1 n = 2r + 1
Dobbertin 24r + 23r + 22r + 2r − 1 n = 5r
Table: Monomial APN Functions
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Until 2006, the list of known APN functions on L = GF(2n) was
rather short.
Y. Edel, G. Kyureghyan and A. Pott established (by an exhaustive
search) the first example of an APN function not equivalent to any
monomial APN functions. Their example is
x3
+ ux36
∈ GF(210
)[x],
where u ∈ wGF(25)∗ ∪ w2GF(25)∗ and w has order 3, is APN on
GF(210). Since then, several new infinite families of polynomial
APN functions have been discovered.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Definition 2
Let L = Fq, q = pn. A function f : L → L is called exceptional
APN if f is APN on L and also on infinitely many extensions of L.
Aubry, McGuire and Rodier made the following conjecture:
Up to equivalence, the Gold and Kasami-Welch functions are
the only exceptional APN functions.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Proposition 1 (J and Wilson 93, Rodier 2009)
Let L = Fq, with q = 2n. A function f : L → L is APN if and only
if the affine surface X with equation
f (x) + f (y) + f (z) + f (x + y + z) = 0
has all its rational points contained in the surface
(x + y)(x + z)(y + z) = 0.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Definition 3
φ(x, y, z) :=
f (x) + f (y) + f (z) + f (x + y + z)
(x + y)(x + z)(y + z)
φj (x, y, z) :=
xj + yj + zj + (x + y + z)j
(x + y)(x + z)(y + z)
Theorem 1
If f = Σfj tj then
φ(x, y, z) = Σfj φj (x, y, z).
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
A First Bound
Second Bound
Theorem 2 (J and Wilson 93, Rodier 2009)
Let f be a polynomial from F2m to F2m , d its degree. Suppose that
the surface X with affine equation
φ(x, y, z) = 0 (2)
is absolutely irreducible. Then, if d < 0.45q1/4 + 0.5 and d ≥ 9 , f
is not APN.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
A First Bound
Second Bound
Proposition 2 (J and Wilson 93, Rodier 2009)
Let f be a polynomial of F2m to itself, d its degree. Let us suppose
that d is not a power of 2 and that the curve X∞ with equation
φd (x, y, z) = 0
is absolutely irreducible. Then the surface φ(x, y, z) = 0 of
equation (2) is absolutely irreducible.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
A First Bound
Second Bound
Theorem 3 (J and Wilson 93, Rodier 2009)
Let f be a polynomial from F2m to F2m , d its degree. Let us
suppose that d is not a power of 2 and that the surface X
φ(x, y, z) = 0
is regular in codimension one. Then if d ≥ 10 and d < q1/4 + 4, f
is not APN.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
A First Bound
Second Bound
PROOF: From an improvement of a theorem of Deligne on Weil’s
conjectures by Ghorpade-Lachaud,
|X(k)−q2
−q−1| ≤ (d −4)(d −5)q3/2
+(−82+57d −13d2
+d3
)q
If q > 183 − 230d + 94d2 − 16d3 + d4 and d ≥ 6, then
X(k) > 3((d − 3)q + 1) and so f is not APN. Q.E.D.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
A First Bound
Second Bound
Absolute irreducibility testing (J and Wilson 91, J 92, J,
McGuire and Wilson 95 )
The main ingredient in all the proofs on APN functions (by
perhaps everyone) is the following absolute irreducibility testing.
Algorithm:
Assume f (x, y, z) factors as P(x, y, z)Q(x, y, z).
Compute and classify multiplicities of each singular point.
Find intersection multiplicities.
If the sum of intersection multiplicies exceeds that predicted
by Bezout’s theorem, then factorization can not occur.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Theorem 4
If the degree of f is odd and not a Gold or a Kasami-Welch
number, then f is not exceptional APN.
Theorem 5
If the degree of f is 2e with e odd, and if f contains an odd degree
term, then f is not exceptional APN.
Theorem 6
If the degree of f is 4e with e ≥ 7 and e ≡ 3 (mod 4), then f is
not exceptional APN.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Theorem 7
Suppose f (x) = x2k +1 + g(x), where deg(g) ≤ 2k−1 + 1. Let
g(x) = 2k−1+1
j=0 aj xj . Suppose that there exists a nonzero
coefficient aj of g such that φj (x, y, z) is absolutely irreducible.
Then φ(x, y, z) is absolutely irreducible and f is not exceptional
APN.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Theorem 8 (J and Wilson, 93)
If f (x) = x22k −2k +1 is a Kasami Welch function, then
φ(x, y, z) =
α∈F2k −F2
Pα(x, y, z), (3)
where Pα(x, y, z) is absolutely irreducible of degree 2k + 1 over
GF(2k).
PROOF:
Hensel lifting, Bezout’s theorem, Gauss’ lemma and so on for three
pages.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Theorem 9 (J and Wilson, 93)
If f (x) = x2k +1 is a Gold function, then
φ(x, y, z) =
α∈F2k −F2
(x + αy + (α + 1)z). (4)
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Lemma 1
For an integer k > 1, let n = 2k + 1, m = 22k − 2k + 1 be Gold
and Kasami-Welch, respectively, and d an integer with d ≡ 3
(mod 4). Then
a) If n1 = 2k1 + 1 and n2 = 2k2 + 1 are different Gold numbers
such that (k1, k2) = 1, then (φn1 , φn2 ) = 1.
b) (φm1 , φm2 ) = 1 for different Kasami-Welch numbers m1 and
m2.
c) (φn, φm) = 1, (φn, φd ) = 1, and (φm, φd ) = 1. In general, if
φd (x, y, z) is absolutely irreducible and is not one of the
factors of φm(x, y, z) (respectively of φn(x, y, z)), then
(φm, φd ) = 1 (respectively (φn, φd ) = 1).
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Theorem 10
For k ≥ 2 and α = 0, let f (x) = x2k +1 + αx2k−1+3 + h(x) ∈ L[x],
where h(x) = 2k−1+1
j=0 aj xj and either a5 = 0 or there is a non
zero aj φj for some j = 5.
Then φ(x, y, z) is absolutely irreducible.
PROOF:
Becomes four times as complicated if the technique of [AMR] is
used. We need new techniques.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Hyperplane sections
Hyperplane section y − z = 0.
Lemma 2
Let φj (x, y, z) be as in (3). Then
a) For n = 2k + 1 > 3, φj (x, y, y) = (x + y)2k −2;
b) For n > 3 and n ≡ 3 (mod 4), x + y does not divide φj (x, y, y).
c) For n > 5 and n ≡ 1 (mod 4), φj (x, y, y) = (x + y)2l −2S(x, y),
such that x + y does not divide S(x, y), where j = 1 + 2l m, l ≥ 2,
and m is an odd number greater than 1.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Main results: Gold case
Theorem 11
For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x], where deg(h) ≡ 3
(mod 4) and deg(h) < 2k + 1. Then, φ(x, y, z) is absolutely
irreducible.
Theorem 12
For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x]. Let d = deg(h). If
d≡ 1 (mod 4) and d < 2k + 1. If (φ2k +1, φd ) = 1, then φ(x, y, z)
is absolutely irreducible.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Theorem 13
Let f (x) = xd + h(x) ∈ L[x], where d = 22k − 2k + 1,
deg(h) ≤ 22k−1 − 2k−1 + 1. Suppose that there exist aj = 0 such
that (φj , φn) = 1. Then φ(x, y, z) is absolutely irreducible.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
The affine transformation φ(x + 1, y + 1) We consider the affine
transformation x ← x + 1, y ← y + 1 on the surface φ(x, y, z) = 0
(see (2)). Let
φj (x, y) = φj (x + 1, y + 1, 1).
Fact: For a Gold number n = 2k + 1 we have:
φn(x, y) =
α∈F2k −F2
(x + αy) (5)
The fact follows directly from equation
Next we will extend our results with two more theorems.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Theorem 14
Given a Gold number n, and a number m ≡ 5 (mod 8) and m > 5,
then (φn, φm) = 1.
As an application of Theorem 14:
Theorem 15
For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x] where d = deg(h) ≡ 5
(mod 8). Then φ(x, y, z) is absolutely irreducible.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Kasami exponents
Lemma 3
Let φj (x, y, z) be as before.
For j = 2k + 1 > 3, φj (x, 1, 1) = (x + 1)2k −2.
For j ≡ 3 (mod 4) > 3,
φj (x, 1, 1) = (x2m + x2m−1 + ... + x + 1)2, where j = 3 + 4m
for m ≥ 1.
For j ≡ 1 (mod 4) > 5,
φj (x, 1, 1) = (x + 1)2l −2(xm−1 + xm−2 + ... + x + 1)2l
, where
j = 1 + 2l m, l ≥ 2 and m > 1 is an odd number.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Theorem 16
For k ≥ 2, let f (x) = x22k −2k +1 + h(x) ∈ L[x] where
d = deg(h) ≡ 3 (mod 4) < 22k − 2k+1 + 3. Then:
If d ≤ 22k−1 − 2k−1 + 1 or
if d > 22k−1 − 2k−1 + 1 and (2k − 1, d−1
2 ) = 1,
then φ(x, y, z) is absolutely irreducible.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
PROOF: Suppose that φ factors as
φ(x, y, z) = P(x, y, z)Q(x, y, z), where P and Q are polynomials
defined on the algebraic closure of L. Then
22k −2k +1
j=3
aj φj (x, y, z) = (Ps +Ps−1 +...+P0)(Qt +Qt−1 +...+Q0)
(6)
where Pi , Qi the homogeneous components of P and Q. Assume
that s ≥ t, then
22k
− 2k
+ 1 > s ≥ 22k−1
− 2k−1
− 1 ≥ t > 0.
Let e = 22k − 2k + 1 − d > 2k − 2.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
PsQt = Pα(x, y, z) (7)
where α ∈ F2k , α = 0, 1. Since Pα are different absolutely
irreducible factors, Ps and Qt are relatively prime.
By the hypothesis on h(x), the homogeneous terms of degree r, for
d − 3 < r < s + t, equal zero. Then the terms of degree s + t − 1
are PsQt−1 + Ps−1Qt = 0. Hence, Ps divides Ps−1Qt and this
implies that Ps divides Ps−1. We conclude that Ps−1 = 0 as the
degree of Ps−1 is less than the degree of Ps. In addition Qt−1 = 0
as Ps = 0. Similarly,
Ps−2 = Qt−2 = 0, Ps−3 = Qt−3 = 0, ..., Ps−(e−1) = Qt−(e−1) = 0.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
The term of degree d − 3 is:
PsQt−e + Ps−eQt = ad φd (x, y, z) (8)
We show that Qt−e = 0. Suppose d ≤ 22k−1 − 2k−1 + 1. Then
e = 22k − 2k + 1 − d ≥ 22k−1 − 2k−1. Since t ≤ 22k−1 − 2k−1 − 1,
then t − e < 0. Thus Qt−e = 0.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
We assume that d > 22k−1 − 2k−1 + 1 and (2k − 1, d−1
2 ) = 1.
Now consider y = z = 1.
PsQt = (x + 1)2k −2
(x2k −2
+ x2k −3
+ ... + 1)2k
(9)
PsQt−e + Ps−eQt = ad (x2m
+ x2m−1
+ ... + 1)2
(10)
where Ps, Qt are functions of x of degree s, t respectively and
d = 3 + 4m.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
x + 1 x2m + x2m−1 + ... + 1. The roots of x2k −2 + x2k −3 + ... + 1,
x2m + x2m−1 + ... + 1 are also lth root of unity for l = 2k − 1 or
l = d−1
2 . Then, by hypothesis, the left hand side of (9) and (10)
are relatively prime, so (Ps, Qt) = 1 which implies that
s = 2k(2k − 2) and t = 2k − 2.
Thus, t − e < 0 implying that Qt−e = 0. Q.E.D.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Consider the particular subcase where deg(h) is a Gold number.
Theorem 17
For k ≥ 2, let f (x) = x22k −2k +1 + h(x) ∈ L[x] where
d = deg(h) = 2m + 1 < 22k − 2k + 1. Then φ(x, y, z) is absolutely
irreducible.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
A new conjecture on φi
Based on overwhelming evidence, we believe that (φ2k +1, φd ) = 1
is always the case, and therefore, theorem 12 is unconditionally
true. From the evidence, we propose the following conjecture:
Conjecture 1
If d≡ 1 (mod 4) and d is not a Gold or Kasami exponent, then,
φ2k +1, φd are relatively prime for all k ≥ 1.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Theorem 18
Given a Gold number n and any odd number m (not a Gold
number), then (φn, φm) = 1.
Proof:
Since (φn, φm) = 1 ⇔ (φn, φm) = 1, we will work with the
functions φ.
Let n = 2k + 1, m = 2i l + 1, where l > 1 is an odd integer.
Let a ∈ F2k , a = 0, a = 1. By (6), we will prove the theorem by
showing that no term (x + ay) divides φm. By definition of the
functions φj , φj , in (3) and (6), (x + ay) divides φm(x, y) if and
only if (x + ay) divides f (x, y) = φm(x, y)(x)(y)(x + y).
Writing f (x, y) = (x + 1)m + (y + 1)m + 1 + (x + y + 1)m as a
sum of homogeneous terms:
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
f (x, y) = Fm(x, y) + Fm−1(x, y) + ... + F2i +1(x, y) (11)
Then (x + ay)|f (x, y) if and only if (x + ay) divides each
homogeneous term Fr in (7).
From (3), by direct computations, we have:
Fm−1(x, y) = xm−1
+ ym−1
+ (x + y)m−1
(12)
Fm(x, y) = xm
+ ym
+ (x + y)m
(13)
Then, for (x + ay) to divides (8) and (9), it should happen that
Fm−1(ay, y) = (ay)m−1 + ym−1 + (ay + y)m−1 = 0 and
Fm(ay, y) = (ay)m + ym + (ay + y)m = 0; which respectively
implies that
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
(a + 1)l
+ al
+ 1 = 0 (14)
(a + 1)l+1
+ al+1
+ 1 = 0 (15)
Substituting (10) in (11)
(al
+ 1)(a + 1) = al+1
+ 1
al+1
+ al
+ a + 1 = al+1
+ 1
al−1
= 1 (16)
ie, a is a (l − 1)-th root of unity. Furthermore, using this in (10)
(a + 1)l−1
= 1 (17)
ie, a + 1 is also a (l − 1)-th root of unity.
Thus, if a does not satisfy either (12) or (13), the theorem is
proved.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Suppose now that a satisfy (12) and (13). Then, let us consider
the term Fm−(2i +1) in (7).
Fm−(2i +1)(x, y)= m
2i +1 (xm−(2i +1) + ym−(2i +1) + (x + y)m−(2i +1))
First, using Lucas’s theorem, let us prove that m
2i +1 =1.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
This theorem uses the convention that a
b = 0 if a < b.
Let l = ar 2r + ar−12r−1 + ... + a12 + 1 be the base 2 expansion of
l. Then, the expansion of m is
m = 21l + 1 = ar 2i+r + ar−12i+r−1 + ... + a12i+1 + 2i + 1. Using
the mentioned theorem we have that m
2i +1 =1.
Now, let us prove that x + ay does not divide Fm−(2i +1). For
x + ay to divide it, it should happen that Fm−(2i +1)(ay, y) = 0,
however,
Fm−(2i +1)(ay, y) = (ay)m−(2i +1) + ym−(2i +1) + (ay + y)m−(2i +1)
= ((a + 1)m−(2i +1) + am−(2i +1) + 1)ym−(2i +1)
= ((a + 1)2i (l−1) + a2i (l−1) + 1)ym−(2i +1)
= ym−(2i +1),
since a, a + 1 are (l − 1)-th roots of unity.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Using this results we can generalize theorems 7 and 8 in the
following theorem.
Theorem 19
For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x], where deg(h) is any
odd number (nor a Gold number). Then φ(x, y, z) is absolutely
irreducible.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Evidence for the new conjecture
We define φj (x, y) = φj (x + 1, y + 1, 1). Then:
For a Gold number j = 2k + 1 we have:
φj (x, y) =
α∈F2k −F2
(x + αy) (18)
For any number j ≡ 5 (mod 8):
φj (x, y) = x4
y + xy4
+
1+4l
r=6
Fr (x, y) (19)
where j = 1 + 4l, l > 1 is any odd number and Fr is zero or
homogeneous of degree r.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Theorem 20
Given a Gold number n and a number m ≡ 5 (mod 8), then
(φn, φm) = 1.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
PROOF: Since (φn, φm) = 1 ⇔ (φn, φm) = 1, we will work with
the functions φ.
Let n = 2k + 1, α ∈ Fk, α = 0, 1 and m = 4l + 1, where l > 1 is
an odd integer.
By eq. (18), (φn, φm) = 1 ⇔ (x + αy) φm(x, y). We will prove
that no factor of the form (x + αy) divides φm.
Suppose that (x + αy) divides φm(x, y). Then (x + αy) divides
F(x, y) = φm(x, y)xy(x + y).
Writing F(x, y) as a sum of homogeneous terms:
F(x, y) = F5(x, y) + F6(x, y) + ... + F1+4l (x, y)
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Then, (x + αy)|F(x, y) if and only if (x + αy)|Fr (x, y) for all r.
By eq. (19), (x − α)|x4y + xy4 implies α3 + 1 = 0, then α is a
3rd -root of unity.
Consider the following three cases to prove the theorem:
l ≡ 0 (mod 3)
l ≡ 1 (mod 3)
l ≡ 2 (mod 3)
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Case l ≡ 0 (mod 3)
Let l ≡ 0 (mod 3), then l = 3q, m = 4(3q) + 1.
If (x + αy) divides φm, by the Remainder theorem φm(αy, y) = 0.
However,
φm(αy, y) = (αy + 1)m
+ (y + 1)m
+ 1 + (α2
y + 1)m
the term Fm−1 of degree m − 1 is:
Fm−1 = m
1 [αm−1 + 1 + (α2)m−1]ym−1 = ym−1, contradicting
that φm(αy, y) = 0.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Case l ≡ 1 (mod 3)
Let l ≡ 1 (mod 3), then l = 1 + 3q,
m = 4(1 + 3q) + 1 = 5 + 4(3q).
If (x + αy) divides φm, φm(αy, y) = 0. However the term Fm−5 of
degree m − 5 in φm(αy, y) is:
Fm−5 = m
5 [αm−5 + 1 + (α2)m−5]ym−5 = ym−5. Contradiction.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Case l ≡ 2 (mod 3)
If l ≡ 2 (mod 3), then l = 2 + 3q, m = 4(2 + 3q) + 1 = 9 + 4(3q).
If (x + αy) divides φm, φm(αy, y) = 0; but computing the term
Fm of degree m in φm(αy, y):
Fm = m
0 [αm + 1 + (α2)m]ym = ym. Q.E.D.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
The Kasami-Welch Degree Case
The case deg(h) ≡ 3 (mod 4).
The case deg(h) ≡ 1 (mod 4).
The affine transformation φ(x + 1, y + 1)
Theorem 21
For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x] where d = deg(h) ≡ 5
(mod 8) < 2k + 1. Then φ(x, y, z) is absolutely irreducible.
Theorem 22
All polynomials of the form f (x) = x65 + h(x) are not exceptional
APN for all odd degree polynomials h.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Some Remarks
In Theorem 12, (φ2k +1, φd ) = 1 is a necessary condition for f (x)
not to be exceptional APN.
This observation allows us to apply several key results. Janwa,
McGuire and Wilson proved the absolute irreducibility of φd , when
d ≡ 3 (mod 4).
They also established absolute irreducibility for several other
infinite cases with d ≡ 1 (mod 4). Moreover for d ≡ 5
(mod 8) > 13, if the maximal cyclic code of odd length m, Bm
(where d = 2l m + 1, l > 0), has no codewords of weight 4 then,
φd (x, y, z) is absolutely irreducible. For many values of m it is
possible that Bl has no codewords of weight 4, for example, if m is
a prime congruent to ±3 (mod 8). For more details and infinite
classes of examples, see [1].
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Some Remarks
1 Until recently, it was thought that φd (x, y, z) was absolutely
irreducible for the values of d ≡ 5 (mod 8). F. Hernando and
G. McGuire [2], with the help of MAGMA, found that the
polynomial φ205(x, y, z) factors in F2[x, y, z].
2 Janwa and Wilson [3] proved, using different methods
including Hensel’s lemma implemented on a computer, that
φd (x, y, z) is absolutely irreducible for 3 < d < 100, provided
that d is not a Gold or a Kasami-Welch number.
3 Subsequent results, (F´erard, Oyono, and Rodier [4] and
Delgado and Janwa [5]) supplement our results for the
Kasami degree polynomials.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Some Remarks
Thank you for your attention.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Some Remarks
Heeralal Janwa, Gary M. McGuire, and Richard M. Wilson.
Double-error-correcting cyclic codes and absolutely irreducible
polynomials over GF(2).
J. Algebra, 178(2):665–676, 1995.
Fernando Hernando and Gary McGuire.
Proof of a conjecture on the sequence of exceptional numbers,
classifying cyclic codes and APN functions.
J. Algebra, 343:78–92, 2011.
H. Janwa and R. M. Wilson.
Hyperplane sections of Fermat varieties in P3
in char. 2 and
some applications to cyclic codes.
In Applied algebra, algebraic algorithms and error-correcting
codes (San Juan, PR, 1993), volume 673 of Lecture Notes in
Comput. Sci., pages 180–194. Springer, Berlin, 1993.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
Introduction
Exceptional APN functions
Lower Bounds
Recent results
Large Classes of Gold degree families that are not exceptional APN
Main results
Open Problems and Future Directions
Some Remarks
Eric F´erard, Roger Oyono, and Franccois Rodier.
Some more functions that are not APN infinitely often. The
case of Gold and Kasami exponents.
In Arithmetic, geometry, cryptography and coding theory,
volume 574 of Contemp. Math., pages 27–36. Amer. Math.
Soc., Providence, RI, 2012.
H. Janwa M. Delgado.
On the conjecture on apn functions.
2012.
Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit

More Related Content

Similar to IWSDA2015-talk-16Sept2015

Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Avichai Cohen
 
Regret Minimization in Multi-objective Submodular Function Maximization
Regret Minimization in Multi-objective Submodular Function MaximizationRegret Minimization in Multi-objective Submodular Function Maximization
Regret Minimization in Multi-objective Submodular Function Maximization
Tasuku Soma
 
slides_low_rank_matrix_optim_farhad
slides_low_rank_matrix_optim_farhadslides_low_rank_matrix_optim_farhad
slides_low_rank_matrix_optim_farhad
Farhad Gholami
 

Similar to IWSDA2015-talk-16Sept2015 (11)

WMC2016-230pm-presented
WMC2016-230pm-presentedWMC2016-230pm-presented
WMC2016-230pm-presented
 
Fractional programming (A tool for optimization)
Fractional programming (A tool for optimization)Fractional programming (A tool for optimization)
Fractional programming (A tool for optimization)
 
Euclid in a Taxicab: Sparse Blind Deconvolution with Smoothed l_1/l_2 Regular...
Euclid in a Taxicab: Sparse Blind Deconvolution with Smoothed l_1/l_2 Regular...Euclid in a Taxicab: Sparse Blind Deconvolution with Smoothed l_1/l_2 Regular...
Euclid in a Taxicab: Sparse Blind Deconvolution with Smoothed l_1/l_2 Regular...
 
Subgradient Methods for Huge-Scale Optimization Problems - Юрий Нестеров, Cat...
Subgradient Methods for Huge-Scale Optimization Problems - Юрий Нестеров, Cat...Subgradient Methods for Huge-Scale Optimization Problems - Юрий Нестеров, Cat...
Subgradient Methods for Huge-Scale Optimization Problems - Юрий Нестеров, Cat...
 
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
 
Problem Understanding through Landscape Theory
Problem Understanding through Landscape TheoryProblem Understanding through Landscape Theory
Problem Understanding through Landscape Theory
 
Introduction to Deep Neural Network
Introduction to Deep Neural NetworkIntroduction to Deep Neural Network
Introduction to Deep Neural Network
 
Regret Minimization in Multi-objective Submodular Function Maximization
Regret Minimization in Multi-objective Submodular Function MaximizationRegret Minimization in Multi-objective Submodular Function Maximization
Regret Minimization in Multi-objective Submodular Function Maximization
 
Anti-differentiating Approximation Algorithms: PageRank and MinCut
Anti-differentiating Approximation Algorithms: PageRank and MinCutAnti-differentiating Approximation Algorithms: PageRank and MinCut
Anti-differentiating Approximation Algorithms: PageRank and MinCut
 
slides_low_rank_matrix_optim_farhad
slides_low_rank_matrix_optim_farhadslides_low_rank_matrix_optim_farhad
slides_low_rank_matrix_optim_farhad
 
Ilya Shkredov – Subsets of Z/pZ with small Wiener norm and arithmetic progres...
Ilya Shkredov – Subsets of Z/pZ with small Wiener norm and arithmetic progres...Ilya Shkredov – Subsets of Z/pZ with small Wiener norm and arithmetic progres...
Ilya Shkredov – Subsets of Z/pZ with small Wiener norm and arithmetic progres...
 

IWSDA2015-talk-16Sept2015

  • 1. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 2. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Progress on Exceptional APN Functions and Absolute Irreducibility of Infinite Classes of Multivariate Polynomials Heeralal Janwa (Joint work with Moises Delgado) IWSDA2015, Bengluru, India 18 September, 2015 Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 3. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Definition 1 Let L = Fq, with q = pn. f : L → L is almost perfect nonlinear (APN) on L if for all a, b ∈ L, a = 0 f (x + a) − f (x) = b (1) has at most 2 solutions. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 4. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Equivalently, for p = 2, the cardinality of {f (x + a)−f (x) : x ∈ L} is at least 2n−1 for each a ∈ L∗. APN functions are important in applications to cryptography, coding theory, combinatorics, finite geometries, and other related areas. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 5. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The best known examples of APN functions are xd Exponent d Constraints Gold 2r + 1 (r, n) = 1 Kasami-Welch 22r − 2r + 1 (r, n) = 1,n odd Welch 2r + 3 n = 2r + 1 Niho 2r + 2r/2 − 1 n = 2r + 1, r even 2r + 2(3r+1)/2 − 1 n = 2r + 1, r odd Inverse 22r − 1 n = 2r + 1 Dobbertin 24r + 23r + 22r + 2r − 1 n = 5r Table: Monomial APN Functions Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 6. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Until 2006, the list of known APN functions on L = GF(2n) was rather short. Y. Edel, G. Kyureghyan and A. Pott established (by an exhaustive search) the first example of an APN function not equivalent to any monomial APN functions. Their example is x3 + ux36 ∈ GF(210 )[x], where u ∈ wGF(25)∗ ∪ w2GF(25)∗ and w has order 3, is APN on GF(210). Since then, several new infinite families of polynomial APN functions have been discovered. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 7. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Definition 2 Let L = Fq, q = pn. A function f : L → L is called exceptional APN if f is APN on L and also on infinitely many extensions of L. Aubry, McGuire and Rodier made the following conjecture: Up to equivalence, the Gold and Kasami-Welch functions are the only exceptional APN functions. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 8. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Proposition 1 (J and Wilson 93, Rodier 2009) Let L = Fq, with q = 2n. A function f : L → L is APN if and only if the affine surface X with equation f (x) + f (y) + f (z) + f (x + y + z) = 0 has all its rational points contained in the surface (x + y)(x + z)(y + z) = 0. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 9. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Definition 3 φ(x, y, z) := f (x) + f (y) + f (z) + f (x + y + z) (x + y)(x + z)(y + z) φj (x, y, z) := xj + yj + zj + (x + y + z)j (x + y)(x + z)(y + z) Theorem 1 If f = Σfj tj then φ(x, y, z) = Σfj φj (x, y, z). Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 10. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions A First Bound Second Bound Theorem 2 (J and Wilson 93, Rodier 2009) Let f be a polynomial from F2m to F2m , d its degree. Suppose that the surface X with affine equation φ(x, y, z) = 0 (2) is absolutely irreducible. Then, if d < 0.45q1/4 + 0.5 and d ≥ 9 , f is not APN. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 11. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions A First Bound Second Bound Proposition 2 (J and Wilson 93, Rodier 2009) Let f be a polynomial of F2m to itself, d its degree. Let us suppose that d is not a power of 2 and that the curve X∞ with equation φd (x, y, z) = 0 is absolutely irreducible. Then the surface φ(x, y, z) = 0 of equation (2) is absolutely irreducible. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 12. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions A First Bound Second Bound Theorem 3 (J and Wilson 93, Rodier 2009) Let f be a polynomial from F2m to F2m , d its degree. Let us suppose that d is not a power of 2 and that the surface X φ(x, y, z) = 0 is regular in codimension one. Then if d ≥ 10 and d < q1/4 + 4, f is not APN. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 13. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions A First Bound Second Bound PROOF: From an improvement of a theorem of Deligne on Weil’s conjectures by Ghorpade-Lachaud, |X(k)−q2 −q−1| ≤ (d −4)(d −5)q3/2 +(−82+57d −13d2 +d3 )q If q > 183 − 230d + 94d2 − 16d3 + d4 and d ≥ 6, then X(k) > 3((d − 3)q + 1) and so f is not APN. Q.E.D. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 14. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions A First Bound Second Bound Absolute irreducibility testing (J and Wilson 91, J 92, J, McGuire and Wilson 95 ) The main ingredient in all the proofs on APN functions (by perhaps everyone) is the following absolute irreducibility testing. Algorithm: Assume f (x, y, z) factors as P(x, y, z)Q(x, y, z). Compute and classify multiplicities of each singular point. Find intersection multiplicities. If the sum of intersection multiplicies exceeds that predicted by Bezout’s theorem, then factorization can not occur. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 15. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Theorem 4 If the degree of f is odd and not a Gold or a Kasami-Welch number, then f is not exceptional APN. Theorem 5 If the degree of f is 2e with e odd, and if f contains an odd degree term, then f is not exceptional APN. Theorem 6 If the degree of f is 4e with e ≥ 7 and e ≡ 3 (mod 4), then f is not exceptional APN. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 16. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Theorem 7 Suppose f (x) = x2k +1 + g(x), where deg(g) ≤ 2k−1 + 1. Let g(x) = 2k−1+1 j=0 aj xj . Suppose that there exists a nonzero coefficient aj of g such that φj (x, y, z) is absolutely irreducible. Then φ(x, y, z) is absolutely irreducible and f is not exceptional APN. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 17. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Theorem 8 (J and Wilson, 93) If f (x) = x22k −2k +1 is a Kasami Welch function, then φ(x, y, z) = α∈F2k −F2 Pα(x, y, z), (3) where Pα(x, y, z) is absolutely irreducible of degree 2k + 1 over GF(2k). PROOF: Hensel lifting, Bezout’s theorem, Gauss’ lemma and so on for three pages. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 18. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Theorem 9 (J and Wilson, 93) If f (x) = x2k +1 is a Gold function, then φ(x, y, z) = α∈F2k −F2 (x + αy + (α + 1)z). (4) Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 19. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Lemma 1 For an integer k > 1, let n = 2k + 1, m = 22k − 2k + 1 be Gold and Kasami-Welch, respectively, and d an integer with d ≡ 3 (mod 4). Then a) If n1 = 2k1 + 1 and n2 = 2k2 + 1 are different Gold numbers such that (k1, k2) = 1, then (φn1 , φn2 ) = 1. b) (φm1 , φm2 ) = 1 for different Kasami-Welch numbers m1 and m2. c) (φn, φm) = 1, (φn, φd ) = 1, and (φm, φd ) = 1. In general, if φd (x, y, z) is absolutely irreducible and is not one of the factors of φm(x, y, z) (respectively of φn(x, y, z)), then (φm, φd ) = 1 (respectively (φn, φd ) = 1). Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 20. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Theorem 10 For k ≥ 2 and α = 0, let f (x) = x2k +1 + αx2k−1+3 + h(x) ∈ L[x], where h(x) = 2k−1+1 j=0 aj xj and either a5 = 0 or there is a non zero aj φj for some j = 5. Then φ(x, y, z) is absolutely irreducible. PROOF: Becomes four times as complicated if the technique of [AMR] is used. We need new techniques. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 21. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Hyperplane sections Hyperplane section y − z = 0. Lemma 2 Let φj (x, y, z) be as in (3). Then a) For n = 2k + 1 > 3, φj (x, y, y) = (x + y)2k −2; b) For n > 3 and n ≡ 3 (mod 4), x + y does not divide φj (x, y, y). c) For n > 5 and n ≡ 1 (mod 4), φj (x, y, y) = (x + y)2l −2S(x, y), such that x + y does not divide S(x, y), where j = 1 + 2l m, l ≥ 2, and m is an odd number greater than 1. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 22. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Main results: Gold case Theorem 11 For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x], where deg(h) ≡ 3 (mod 4) and deg(h) < 2k + 1. Then, φ(x, y, z) is absolutely irreducible. Theorem 12 For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x]. Let d = deg(h). If d≡ 1 (mod 4) and d < 2k + 1. If (φ2k +1, φd ) = 1, then φ(x, y, z) is absolutely irreducible. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 23. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Theorem 13 Let f (x) = xd + h(x) ∈ L[x], where d = 22k − 2k + 1, deg(h) ≤ 22k−1 − 2k−1 + 1. Suppose that there exist aj = 0 such that (φj , φn) = 1. Then φ(x, y, z) is absolutely irreducible. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 24. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) The affine transformation φ(x + 1, y + 1) We consider the affine transformation x ← x + 1, y ← y + 1 on the surface φ(x, y, z) = 0 (see (2)). Let φj (x, y) = φj (x + 1, y + 1, 1). Fact: For a Gold number n = 2k + 1 we have: φn(x, y) = α∈F2k −F2 (x + αy) (5) The fact follows directly from equation Next we will extend our results with two more theorems. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 25. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Theorem 14 Given a Gold number n, and a number m ≡ 5 (mod 8) and m > 5, then (φn, φm) = 1. As an application of Theorem 14: Theorem 15 For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x] where d = deg(h) ≡ 5 (mod 8). Then φ(x, y, z) is absolutely irreducible. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 26. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Kasami exponents Lemma 3 Let φj (x, y, z) be as before. For j = 2k + 1 > 3, φj (x, 1, 1) = (x + 1)2k −2. For j ≡ 3 (mod 4) > 3, φj (x, 1, 1) = (x2m + x2m−1 + ... + x + 1)2, where j = 3 + 4m for m ≥ 1. For j ≡ 1 (mod 4) > 5, φj (x, 1, 1) = (x + 1)2l −2(xm−1 + xm−2 + ... + x + 1)2l , where j = 1 + 2l m, l ≥ 2 and m > 1 is an odd number. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 27. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Theorem 16 For k ≥ 2, let f (x) = x22k −2k +1 + h(x) ∈ L[x] where d = deg(h) ≡ 3 (mod 4) < 22k − 2k+1 + 3. Then: If d ≤ 22k−1 − 2k−1 + 1 or if d > 22k−1 − 2k−1 + 1 and (2k − 1, d−1 2 ) = 1, then φ(x, y, z) is absolutely irreducible. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 28. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) PROOF: Suppose that φ factors as φ(x, y, z) = P(x, y, z)Q(x, y, z), where P and Q are polynomials defined on the algebraic closure of L. Then 22k −2k +1 j=3 aj φj (x, y, z) = (Ps +Ps−1 +...+P0)(Qt +Qt−1 +...+Q0) (6) where Pi , Qi the homogeneous components of P and Q. Assume that s ≥ t, then 22k − 2k + 1 > s ≥ 22k−1 − 2k−1 − 1 ≥ t > 0. Let e = 22k − 2k + 1 − d > 2k − 2. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 29. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) PsQt = Pα(x, y, z) (7) where α ∈ F2k , α = 0, 1. Since Pα are different absolutely irreducible factors, Ps and Qt are relatively prime. By the hypothesis on h(x), the homogeneous terms of degree r, for d − 3 < r < s + t, equal zero. Then the terms of degree s + t − 1 are PsQt−1 + Ps−1Qt = 0. Hence, Ps divides Ps−1Qt and this implies that Ps divides Ps−1. We conclude that Ps−1 = 0 as the degree of Ps−1 is less than the degree of Ps. In addition Qt−1 = 0 as Ps = 0. Similarly, Ps−2 = Qt−2 = 0, Ps−3 = Qt−3 = 0, ..., Ps−(e−1) = Qt−(e−1) = 0. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 30. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) The term of degree d − 3 is: PsQt−e + Ps−eQt = ad φd (x, y, z) (8) We show that Qt−e = 0. Suppose d ≤ 22k−1 − 2k−1 + 1. Then e = 22k − 2k + 1 − d ≥ 22k−1 − 2k−1. Since t ≤ 22k−1 − 2k−1 − 1, then t − e < 0. Thus Qt−e = 0. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 31. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) We assume that d > 22k−1 − 2k−1 + 1 and (2k − 1, d−1 2 ) = 1. Now consider y = z = 1. PsQt = (x + 1)2k −2 (x2k −2 + x2k −3 + ... + 1)2k (9) PsQt−e + Ps−eQt = ad (x2m + x2m−1 + ... + 1)2 (10) where Ps, Qt are functions of x of degree s, t respectively and d = 3 + 4m. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 32. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) x + 1 x2m + x2m−1 + ... + 1. The roots of x2k −2 + x2k −3 + ... + 1, x2m + x2m−1 + ... + 1 are also lth root of unity for l = 2k − 1 or l = d−1 2 . Then, by hypothesis, the left hand side of (9) and (10) are relatively prime, so (Ps, Qt) = 1 which implies that s = 2k(2k − 2) and t = 2k − 2. Thus, t − e < 0 implying that Qt−e = 0. Q.E.D. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 33. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Consider the particular subcase where deg(h) is a Gold number. Theorem 17 For k ≥ 2, let f (x) = x22k −2k +1 + h(x) ∈ L[x] where d = deg(h) = 2m + 1 < 22k − 2k + 1. Then φ(x, y, z) is absolutely irreducible. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 34. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) A new conjecture on φi Based on overwhelming evidence, we believe that (φ2k +1, φd ) = 1 is always the case, and therefore, theorem 12 is unconditionally true. From the evidence, we propose the following conjecture: Conjecture 1 If d≡ 1 (mod 4) and d is not a Gold or Kasami exponent, then, φ2k +1, φd are relatively prime for all k ≥ 1. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 35. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Theorem 18 Given a Gold number n and any odd number m (not a Gold number), then (φn, φm) = 1. Proof: Since (φn, φm) = 1 ⇔ (φn, φm) = 1, we will work with the functions φ. Let n = 2k + 1, m = 2i l + 1, where l > 1 is an odd integer. Let a ∈ F2k , a = 0, a = 1. By (6), we will prove the theorem by showing that no term (x + ay) divides φm. By definition of the functions φj , φj , in (3) and (6), (x + ay) divides φm(x, y) if and only if (x + ay) divides f (x, y) = φm(x, y)(x)(y)(x + y). Writing f (x, y) = (x + 1)m + (y + 1)m + 1 + (x + y + 1)m as a sum of homogeneous terms: Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 36. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) f (x, y) = Fm(x, y) + Fm−1(x, y) + ... + F2i +1(x, y) (11) Then (x + ay)|f (x, y) if and only if (x + ay) divides each homogeneous term Fr in (7). From (3), by direct computations, we have: Fm−1(x, y) = xm−1 + ym−1 + (x + y)m−1 (12) Fm(x, y) = xm + ym + (x + y)m (13) Then, for (x + ay) to divides (8) and (9), it should happen that Fm−1(ay, y) = (ay)m−1 + ym−1 + (ay + y)m−1 = 0 and Fm(ay, y) = (ay)m + ym + (ay + y)m = 0; which respectively implies that Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 37. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) (a + 1)l + al + 1 = 0 (14) (a + 1)l+1 + al+1 + 1 = 0 (15) Substituting (10) in (11) (al + 1)(a + 1) = al+1 + 1 al+1 + al + a + 1 = al+1 + 1 al−1 = 1 (16) ie, a is a (l − 1)-th root of unity. Furthermore, using this in (10) (a + 1)l−1 = 1 (17) ie, a + 1 is also a (l − 1)-th root of unity. Thus, if a does not satisfy either (12) or (13), the theorem is proved. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 38. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Suppose now that a satisfy (12) and (13). Then, let us consider the term Fm−(2i +1) in (7). Fm−(2i +1)(x, y)= m 2i +1 (xm−(2i +1) + ym−(2i +1) + (x + y)m−(2i +1)) First, using Lucas’s theorem, let us prove that m 2i +1 =1. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 39. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) This theorem uses the convention that a b = 0 if a < b. Let l = ar 2r + ar−12r−1 + ... + a12 + 1 be the base 2 expansion of l. Then, the expansion of m is m = 21l + 1 = ar 2i+r + ar−12i+r−1 + ... + a12i+1 + 2i + 1. Using the mentioned theorem we have that m 2i +1 =1. Now, let us prove that x + ay does not divide Fm−(2i +1). For x + ay to divide it, it should happen that Fm−(2i +1)(ay, y) = 0, however, Fm−(2i +1)(ay, y) = (ay)m−(2i +1) + ym−(2i +1) + (ay + y)m−(2i +1) = ((a + 1)m−(2i +1) + am−(2i +1) + 1)ym−(2i +1) = ((a + 1)2i (l−1) + a2i (l−1) + 1)ym−(2i +1) = ym−(2i +1), since a, a + 1 are (l − 1)-th roots of unity. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 40. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Using this results we can generalize theorems 7 and 8 in the following theorem. Theorem 19 For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x], where deg(h) is any odd number (nor a Gold number). Then φ(x, y, z) is absolutely irreducible. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 41. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Evidence for the new conjecture We define φj (x, y) = φj (x + 1, y + 1, 1). Then: For a Gold number j = 2k + 1 we have: φj (x, y) = α∈F2k −F2 (x + αy) (18) For any number j ≡ 5 (mod 8): φj (x, y) = x4 y + xy4 + 1+4l r=6 Fr (x, y) (19) where j = 1 + 4l, l > 1 is any odd number and Fr is zero or homogeneous of degree r. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 42. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Theorem 20 Given a Gold number n and a number m ≡ 5 (mod 8), then (φn, φm) = 1. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 43. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) PROOF: Since (φn, φm) = 1 ⇔ (φn, φm) = 1, we will work with the functions φ. Let n = 2k + 1, α ∈ Fk, α = 0, 1 and m = 4l + 1, where l > 1 is an odd integer. By eq. (18), (φn, φm) = 1 ⇔ (x + αy) φm(x, y). We will prove that no factor of the form (x + αy) divides φm. Suppose that (x + αy) divides φm(x, y). Then (x + αy) divides F(x, y) = φm(x, y)xy(x + y). Writing F(x, y) as a sum of homogeneous terms: F(x, y) = F5(x, y) + F6(x, y) + ... + F1+4l (x, y) Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 44. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Then, (x + αy)|F(x, y) if and only if (x + αy)|Fr (x, y) for all r. By eq. (19), (x − α)|x4y + xy4 implies α3 + 1 = 0, then α is a 3rd -root of unity. Consider the following three cases to prove the theorem: l ≡ 0 (mod 3) l ≡ 1 (mod 3) l ≡ 2 (mod 3) Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 45. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Case l ≡ 0 (mod 3) Let l ≡ 0 (mod 3), then l = 3q, m = 4(3q) + 1. If (x + αy) divides φm, by the Remainder theorem φm(αy, y) = 0. However, φm(αy, y) = (αy + 1)m + (y + 1)m + 1 + (α2 y + 1)m the term Fm−1 of degree m − 1 is: Fm−1 = m 1 [αm−1 + 1 + (α2)m−1]ym−1 = ym−1, contradicting that φm(αy, y) = 0. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 46. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Case l ≡ 1 (mod 3) Let l ≡ 1 (mod 3), then l = 1 + 3q, m = 4(1 + 3q) + 1 = 5 + 4(3q). If (x + αy) divides φm, φm(αy, y) = 0. However the term Fm−5 of degree m − 5 in φm(αy, y) is: Fm−5 = m 5 [αm−5 + 1 + (α2)m−5]ym−5 = ym−5. Contradiction. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 47. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Case l ≡ 2 (mod 3) If l ≡ 2 (mod 3), then l = 2 + 3q, m = 4(2 + 3q) + 1 = 9 + 4(3q). If (x + αy) divides φm, φm(αy, y) = 0; but computing the term Fm of degree m in φm(αy, y): Fm = m 0 [αm + 1 + (α2)m]ym = ym. Q.E.D. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 48. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions The Kasami-Welch Degree Case The case deg(h) ≡ 3 (mod 4). The case deg(h) ≡ 1 (mod 4). The affine transformation φ(x + 1, y + 1) Theorem 21 For k ≥ 2, let f (x) = x2k +1 + h(x) ∈ L[x] where d = deg(h) ≡ 5 (mod 8) < 2k + 1. Then φ(x, y, z) is absolutely irreducible. Theorem 22 All polynomials of the form f (x) = x65 + h(x) are not exceptional APN for all odd degree polynomials h. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 49. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Some Remarks In Theorem 12, (φ2k +1, φd ) = 1 is a necessary condition for f (x) not to be exceptional APN. This observation allows us to apply several key results. Janwa, McGuire and Wilson proved the absolute irreducibility of φd , when d ≡ 3 (mod 4). They also established absolute irreducibility for several other infinite cases with d ≡ 1 (mod 4). Moreover for d ≡ 5 (mod 8) > 13, if the maximal cyclic code of odd length m, Bm (where d = 2l m + 1, l > 0), has no codewords of weight 4 then, φd (x, y, z) is absolutely irreducible. For many values of m it is possible that Bl has no codewords of weight 4, for example, if m is a prime congruent to ±3 (mod 8). For more details and infinite classes of examples, see [1]. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 50. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Some Remarks 1 Until recently, it was thought that φd (x, y, z) was absolutely irreducible for the values of d ≡ 5 (mod 8). F. Hernando and G. McGuire [2], with the help of MAGMA, found that the polynomial φ205(x, y, z) factors in F2[x, y, z]. 2 Janwa and Wilson [3] proved, using different methods including Hensel’s lemma implemented on a computer, that φd (x, y, z) is absolutely irreducible for 3 < d < 100, provided that d is not a Gold or a Kasami-Welch number. 3 Subsequent results, (F´erard, Oyono, and Rodier [4] and Delgado and Janwa [5]) supplement our results for the Kasami degree polynomials. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 51. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Some Remarks Thank you for your attention. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 52. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Some Remarks Heeralal Janwa, Gary M. McGuire, and Richard M. Wilson. Double-error-correcting cyclic codes and absolutely irreducible polynomials over GF(2). J. Algebra, 178(2):665–676, 1995. Fernando Hernando and Gary McGuire. Proof of a conjecture on the sequence of exceptional numbers, classifying cyclic codes and APN functions. J. Algebra, 343:78–92, 2011. H. Janwa and R. M. Wilson. Hyperplane sections of Fermat varieties in P3 in char. 2 and some applications to cyclic codes. In Applied algebra, algebraic algorithms and error-correcting codes (San Juan, PR, 1993), volume 673 of Lecture Notes in Comput. Sci., pages 180–194. Springer, Berlin, 1993. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit
  • 53. Introduction Exceptional APN functions Lower Bounds Recent results Large Classes of Gold degree families that are not exceptional APN Main results Open Problems and Future Directions Some Remarks Eric F´erard, Roger Oyono, and Franccois Rodier. Some more functions that are not APN infinitely often. The case of Gold and Kasami exponents. In Arithmetic, geometry, cryptography and coding theory, volume 574 of Contemp. Math., pages 27–36. Amer. Math. Soc., Providence, RI, 2012. H. Janwa M. Delgado. On the conjecture on apn functions. 2012. Heeralal Janwa (Joint work with Moises Delgado) Progress on Exceptional APN Functions and Absolute Irreducibilit