The Green-Tao theorem and a relative Szemerédi theorem

Size: px

Start display at page:

Download "The Green-Tao theorem and a relative Szemerédi theorem"

Lucas Gallagher
5 years ago
Views:

1 The Green-Tao theorem and a relative Szemerédi theorem Yufei Zhao Massachusetts Institute of Technology Joint work with David Conlon (Oxford) and Jacob Fox (MIT) Simons Institute December 2013

2 Green Tao Theorem (arxiv 2004; Annals 2008) The primes contain arbitrarily long arithmetic progressions. Examples: 3, 5, 7 5, 11, 17, 23, 29 7, 37, 67, 97, 127, 157 Longest known: 26 terms

3 Green Tao Theorem (2008) The primes contain arbitrarily long arithmetic progressions (AP). Szemerédi s Theorem (1975) Every subset of N with positive density contains arbitrarily long APs. (upper) density of A N is lim sup [N] := {1, 2,..., N} P = prime numbers Prime number theorem: N P [N] N A [N] N 1 log N

4 Proof strategy of Green Tao theorem P = prime numbers, Q = almost primes P [N] P Q with relative positive density, i.e., Q [N] > δ Step 1: Relative Szemerédi theorem (informally) If S N satisfies certain pseudorandomness conditions, then every subset of S of positive density contains long APs. Step 2: Construct a superset of primes that satisfies the conditions.

5 Relative Szemerédi theorem Relative Szemerédi theorem (informally) If S N satisfies certain pseudorandomness conditions, then every subset of S of positive density contains long APs. What pseudorandomness conditions? Green Tao: 1 Linear forms condition 2 Correlation condition

6 Relative Szemerédi theorem Relative Szemerédi theorem (informally) If S N satisfies certain pseudorandomness conditions, then every subset of S of positive density contains long APs. What pseudorandomness conditions? Green Tao: 1 Linear forms condition 2 Correlation condition A natural question (e.g., asked by Green, Gowers,... ) Does relative Szemerédi theorem hold with weaker and more natural hypotheses?

7 Relative Szemerédi theorem Relative Szemerédi theorem (informally) If S N satisfies certain pseudorandomness conditions, then every subset of S of positive density contains long APs. What pseudorandomness conditions? Green Tao: 1 Linear forms condition 2 Correlation condition no longer needed A natural question (e.g., asked by Green, Gowers,... ) Does relative Szemerédi theorem hold with weaker and more natural hypotheses? Our main result Yes! A weak linear forms condition suffices.

8 Szemerédi s theorem Host set: N Relative Szemerédi theorem Host set: some sparse subset of integers Conclusion: relatively dense subsets contain long APs

9 Szemerédi s theorem Host set: N Relative Szemerédi theorem Host set: some sparse subset of integers Random host set Kohayakawa Luczak Rödl 96 3-AP, p N 1/2 Conlon Gowers 10+ Schacht 10+ k-ap, p N 1/(k 1) Pseudorandom host set Green Tao 08 Conlon Fox Z. 13+ linear forms + correlation linear forms Conclusion: relatively dense subsets contain long APs

10 Roth s theorem Roth s theorem (1952) If A [N] is 3-AP-free, then A = o(n). [N] := {1, 2,..., N} 3-AP = 3-term arithmetic progression It ll be easier (and equivalent) to work in Z N := Z/NZ.

11 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. G A X Y Z

12 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. G A x y iff 2x + y A X x Y y Z

13 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). X Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. G A x x z iff x z A Y z Z

14 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. G A X Y y y z iff y 2z A z Z

15 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). X Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. G A x y iff 2x + y A x x z iff x z A Y y y z iff y 2z A z Z

16 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. Triangle xyz in G A 2x + y, x z, y 2z A G A x y iff 2x + y A Y y X x y z iff y 2z A x z iff x z A z Z

17 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. Triangle xyz in G A 2x + y, x z, y 2z A It s a 3-AP with diff x y z G A x y iff 2x + y A Y y X x y z iff y 2z A x z iff x z A z Z

18 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. Triangle xyz in G A 2x + y, x z, y 2z A It s a 3-AP with diff x y z No triangles? Y G A x y iff 2x + y A y X x y z iff y 2z A x z iff x z A z Z

19 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. Triangle xyz in G A 2x + y, x z, y 2z A y It s a 3-AP with diff x y z Y y z iff y 2z A No triangles? Only triangles trivial 3-APs with diff 0. G A x y iff 2x + y A X x x z iff x z A z Z

20 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Given A, construct tripartite graph G A with vertex sets X = Y = Z = Z N. G A x y iff 2x + y A Triangle xyz in G A 2x + y, x z, y 2z A y z It s a 3-AP with diff x y z Y y z iff y 2z A No triangles? Only triangles trivial 3-APs with diff 0. Every edge of the graph is contained in exactly one triangle (the one with x + y + z = 0). X x x z iff x z A Z

21 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Constructed a graph with 3N vertices 3N A edges every edge in exactly one triangle

22 Proof of Roth s theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Constructed a graph with 3N vertices 3N A edges every edge in exactly one triangle Theorem (Ruzsa & Szemerédi 76) If every edge in a graph G = (V, E) is contained in exactly one triangle, then E = o( V 2 ). (a consequence of the triangle removal lemma) So 3N A = o(n 2 ). Thus A = o(n).

23 Relative Roth theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Relative Roth theorem (Conlon, Fox, Z.) If S Z N satisfies some pseudorandomness conditions, and A S is 3-AP-free, then A = o( S ).

24 Relative Roth theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Relative Roth theorem (Conlon, Fox, Z.) If S Z N satisfies some pseudorandomness conditions, and A S is 3-AP-free, then A = o( S ). G S x y iff 2x + y S Z N x x z iff x z S Z N y y z iff y 2z S z Z N

25 Relative Roth theorem Roth s theorem (1952) If A Z N is 3-AP-free, then A = o(n). Relative Roth theorem (Conlon, Fox, Z.) If S Z N satisfies some pseudorandomness conditions, and A S is 3-AP-free, then A = o( S ). G S x y iff 2x + y S Z N x x z iff x z S Pseudorandomness condition for S: G S has asymp. the expected number of embeddings of K 2,2,2 & its subgraphs (compared to random graph of same density) Z N y y z iff y 2z S z Z N K 2,2,2 & subgraphs, e.g.,

26 Analogy with quasirandom graphs Chung-Graham-Wilson 89 showed that in constant edge-density graphs, many quasirandomness conditions are equivalent, one of which is having the correct C 4 count 2-blow-up

27 Analogy with quasirandom graphs Chung-Graham-Wilson 89 showed that in constant edge-density graphs, many quasirandomness conditions are equivalent, one of which is having the correct C 4 count 2-blow-up In sparse graphs, the CGW equivalences do not hold.

28 Analogy with quasirandom graphs Chung-Graham-Wilson 89 showed that in constant edge-density graphs, many quasirandomness conditions are equivalent, one of which is having the correct C 4 count 2-blow-up In sparse graphs, the CGW equivalences do not hold. Our results can be viewed as saying that: Many extremal and Ramsey results about H (e.g., H = K 3 ) in sparse graphs hold if there is a host graph that behaves pseudorandomly with respect to counts of the 2-blow-up of H. 2-blow-up

29 Roth s theorem: from one 3-AP to many 3-APs Roth s theorem δ > 0. Every A Z N with A δn contains a 3-AP, provided N is sufficiently large.

30 Roth s theorem: from one 3-AP to many 3-APs Roth s theorem δ > 0. Every A Z N with A δn contains a 3-AP, provided N is sufficiently large. By an averaging argument (Varnavides), we get many 3-APs: Roth s theorem (counting version) δ > 0 c > 0 so that every A Z N with A δn contains at least cn 2 3-APs, provided N is sufficiently large.

31 Transference Start with (sparse) A S Z N, A δ S

32 Transference Start with (sparse) A S Z N, A δ S One can find a dense model Ã for A: (dense) Ã Z N, Ã N A S δ

33 Transference Start with (sparse) A S Z N, A δ S One can find a dense model Ã for A: (dense) Ã Z N, Ã N A S δ Counting lemma will tell us that ( ) N 3 {3-APs in A} {3-APs in S Ã}

34 Transference Start with (sparse) A S Z N, A δ S One can find a dense model Ã for A: (dense) Ã Z N, Ã N A S δ Counting lemma will tell us that ( ) N 3 {3-APs in A} {3-APs in S Ã} cn 2 [By Roth s Theorem] = relative Roth theorem

35 Roth s theorem: weighted version Roth s theorem (counting version) δ > 0 c > 0 so that every A Z N with A δn contains at least cn 2 3-APs, provided N is sufficiently large.

36 Roth s theorem: weighted version Roth s theorem (counting version) δ > 0 c > 0 so that every A Z N with A δn contains at least cn 2 3-APs, provided N is sufficiently large. Roth s theorem (weighted version) δ > 0 c > 0 so that every f : Z N [0, 1] with Ef δ satisfies E x,d ZN [f (x)f (x + d)f (x + 2d)] c provided N is sufficiently large.

37 Roth s theorem (weighted version) δ > 0 c > 0 so that every f : Z N [0, 1] with E[f ] δ satisfies E x,d ZN [f (x)f (x + d)f (x + 2d)] c provided N is sufficiently large. Sparse setting: some sparse host set S Z N. More generally, use a normalized measure: ν : Z N [0, ) with Eν = 1. E.g., ν = N S 1 S normalized indicator function.

38 Roth s theorem (weighted version) δ > 0 c > 0 so that every f : Z N [0, 1] with E[f ] δ satisfies E x,d ZN [f (x)f (x + d)f (x + 2d)] c provided N is sufficiently large. Sparse setting: some sparse host set S Z N. More generally, use a normalized measure: ν : Z N [0, ) with Eν = 1. E.g., ν = N S 1 S normalized indicator function. The subset A S with A δ S corresponds to f : Z N [0, ), Ef δ and f majorized by ν, meaning that f (x) ν(x) x Z N.

39 Dense setting Sparse setting Sets A Z N A δ A S Z N A δ S Functions f : Z N [0, 1] Ef δ f ν : Z N [0, ) Ef δ, Eν = 1 (sparse with ν 1 dense setting)

40 Relative Roth theorem Roth s theorem (weighted version) δ > 0 c > 0 so that every f : Z N [0, 1] with Ef δ satisfies E x,d ZN [f (x)f (x + d)f (x + 2d)] c provided N is sufficiently large. Relative Roth theorem (Conlon, Fox, Z.) δ > 0 c > 0 so that if ν : Z N [0, ) satisfies the 3-linear forms condition, and f : Z N [0, ) majorized by ν and Ef δ, then E x,d ZN [f (x)f (x + d)f (x + 2d)] c provided N is sufficiently large.

41 3-linear forms condition The density of K 2,2,2 in Z N x ν(2x + y) ν(x z) y z Z N ν( y 2z) Z N

42 Relative Roth theorem (Conlon, Fox, Z.) δ > 0 c > 0 so that if ν : Z N [0, ) satisfies the 3-linear forms condition, and f : Z N [0, ) majorized by ν and Ef δ, then E x,d ZN [f (x)f (x + d)f (x + 2d)] c provided N is sufficiently large. ν : Z N [0, ) satisfies the 3-linear forms condition if E[ν(2x + y)ν(2x + y)ν(2x + y )ν(2x + y ) ν(x z)ν(x z)ν(x z )ν(x z ) ν( y 2z)ν( y 2z)ν( y 2z )ν( y 2z )] = 1 + o(1) as well as if any subset of the 12 factors were deleted.

43 Relative Szemerédi theorem (Conlon, Fox, Z.) δ > 0, k N c(k, δ) > 0 so that if ν : Z N [0, ) satisfies the k-linear forms condition, and f : Z N [0, ) majorized by ν and Ef δ, then E x,d ZN [f (x)f (x + d)f (x + 2d) f (x + (k 1)d)] c(k, δ) provided N is sufficiently large. k = 4: build a weighted 4-partite 3-uniform hypergraph on W X Y : ν(3w + 2x + y ) on W X Z: ν(2w + x z) on W Y Z: ν(w y 2z) on X Y Z: ν( x 2y 3z) common diff: w x y z

44 Relative Szemerédi theorem (Conlon, Fox, Z.) δ > 0, k N c(k, δ) > 0 so that if ν : Z N [0, ) satisfies the k-linear forms condition, and f : Z N [0, ) majorized by ν and Ef δ, then E x,d ZN [f (x)f (x + d)f (x + 2d) f (x + (k 1)d)] c(k, δ) provided N is sufficiently large. k = 4: build a weighted 4-partite 3-uniform hypergraph on W X Y : ν(3w + 2x + y ) on W X Z: ν(2w + x z) on W Y Z: ν(w y 2z) on X Y Z: ν( x 2y 3z) common diff: w x y z 4-linear forms condition: correct count of the 2-blow-up of the simplex K (3) 4 (as well as its subgraphs)

45 Two approaches Conlon, Fox, Z. A relative Szemerédi theorem. 20pp Z. An arithmetic transference proof of a relative Sz. theorem. 6pp

46 Two approaches Conlon, Fox, Z. A relative Szemerédi theorem. 20pp Transfer hypergraph removal lemma Z. An arithmetic transference proof of a relative Sz. theorem. 6pp

47 Two approaches Conlon, Fox, Z. A relative Szemerédi theorem. 20pp Transfer hypergraph removal lemma Z. An arithmetic transference proof of a relative Sz. theorem. 6pp Transfer Szemerédi s theorem

48 Two approaches Conlon, Fox, Z. A relative Szemerédi theorem. 20pp Transfer hypergraph removal lemma More general Z. An arithmetic transference proof of a relative Sz. theorem. 6pp Transfer Szemerédi s theorem More direct

49 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ

50 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ Dense model theorem: one can approximate f (in cut norm) by (dense) f : ZN [0, 1] E f = Ef

51 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ Dense model theorem: one can approximate f (in cut norm) by (dense) f : ZN [0, 1] E f = Ef Counting lemma implies E x,d [f (x)f (x + d)f (x + 2d)] E x,d [ f (x) f (x + d) f (x + 2d)]

52 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ Dense model theorem: one can approximate f (in cut norm) by (dense) f : ZN [0, 1] E f = Ef Counting lemma implies E x,d [f (x)f (x + d)f (x + 2d)] E x,d [ f (x) f (x + d) f (x + 2d)] c [By Roth s Thm (weighted version)] = relative Roth theorem

53 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ Dense model theorem: one can approximate f (in cut norm) by (dense) f : ZN [0, 1] E f = Ef Counting lemma implies E x,d [f (x)f (x + d)f (x + 2d)] E x,d [ f (x) f (x + d) f (x + 2d)] c [By Roth s Thm (weighted version)] = relative Roth theorem

54 In what sense does 0 f 1 approximate 0 f ν?

55 In what sense does 0 f 1 approximate 0 f ν? Previous approach (Green Tao): Gowers uniformity norm Our approach: cut norm (aka discrepancy)

56 In what sense does 0 f 1 approximate 0 f ν? Previous approach (Green Tao): Gowers uniformity norm Our approach: cut norm (aka discrepancy) Using cut norm: Cheaper dense model theorem Trickier counting lemma

57 Cut norm X Weighted bipartite graphs g, g : X Y R Cut norm (Frieze-Kannan): g g A ɛ means that for all A X and B Y : E x X [(g(x, y) g(x, y))1 A (x)1 B (y)] ɛ y Y B Y

58 Cut norm X Weighted bipartite graphs g, g : X Y R Cut norm (Frieze-Kannan): g g A ɛ means that for all A X and B Y : E x X [(g(x, y) g(x, y))1 A (x)1 B (y)] ɛ y Y B Y For Z N : f, f : Z N R being ɛ-close in cut norm means: for all A, B Z N E x,y ZN [(f (2x + y) f (2x + y))1 A (x)1 B (y)] ɛ. (weaker than being close in Gowers uniformity norm)

59 Dense model theorem Theorem (Dense model) If ν : Z N [0, ) is close to 1 in cut norm then f : Z N [0, ) majorized by ν f : Z N [0, 1] s.t. f is close to f in cut norm and Ef = E f

60 Dense model theorem Theorem (Dense model) If ν : Z N [0, ) is close to 1 in cut norm then f : Z N [0, ) majorized by ν f : Z N [0, 1] s.t. f is close to f in cut norm and Ef = E f Proof approaches 1. Regularity-type energy-increment argument (Green Tao, Tao Ziegler) 2. Separating hyperplane theorem / LP duality + Weierstrass polynomial approximation theorem (Gowers & Reingold Trevisan Tulsiani Vadhan)

61 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ Dense model theorem: one can approximate f (in cut norm) by (dense) f : ZN [0, 1] E f = Ef Counting lemma implies E x,d [f (x)f (x + d)f (x + 2d)] E x,d [ f (x) f (x + d) f (x + 2d)] c [By Roth s Thm (weighted version)] = relative Roth theorem

62 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ Dense model theorem: one can approximate f (in cut norm) by (dense) f : ZN [0, 1] E f = Ef Counting lemma implies E x,d [f (x)f (x + d)f (x + 2d)] E x,d [ f (x) f (x + d) f (x + 2d)] c [By Roth s Thm (weighted version)] = relative Roth theorem

63 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z

64 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z E[(g(x, y) g(x, y))1 A (x)1 B (y)] ɛ A X, B Y

65 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z E[(g(x, y) g(x, y))a(x)b(y)] ɛ a: X [0, 1], b : Y [0, 1]

66 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z E[(g(x, y) g(x, y))a(x)b(y)] ɛ a: X [0, 1], b : Y [0, 1] E[g(x, y)g(x, z)g(y, z)]

67 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z E[(g(x, y) g(x, y))a(x)b(y)] ɛ a: X [0, 1], b : Y [0, 1] E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y)g(x, z)g(y, z)] + O(ɛ)

68 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z E[(g(x, y) g(x, y))a(x)b(y)] ɛ a: X [0, 1], b : Y [0, 1] E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y)g(x, z)g(y, z)] + O(ɛ) = E[ g(x, y) g(x, z)g(y, z)] + O(ɛ)

69 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z E[(g(x, y) g(x, y))a(x)b(y)] ɛ a: X [0, 1], b : Y [0, 1] E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y)g(x, z)g(y, z)] + O(ɛ) = E[ g(x, y) g(x, z)g(y, z)] + O(ɛ) = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ)

70 Counting lemma Weighted graphs g, g : (X Y ) (X Z) (Y Z) R Triangle counting lemma (dense setting) Assume 0 g, g 1. If g g ɛ, then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ). y x z E[(g(x, y) g(x, y))a(x)b(y)] ɛ a: X [0, 1], b : Y [0, 1] E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y)g(x, z)g(y, z)] + O(ɛ) = E[ g(x, y) g(x, z)g(y, z)] + O(ɛ) = E[ g(x, y) g(x, z) g(y, z)] + O(ɛ) This argument doesn t work in the sparse setting (g unbounded)

71 Sparse counting lemma Sparse triangle counting lemma (Conlon, Fox, Z.) Assume that ν satisfies the 3-linear forms condition. If 0 g ν, 0 g 1 and g g = o(1), then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + o(1).

72 Sparse counting lemma Sparse triangle counting lemma (Conlon, Fox, Z.) Assume that ν satisfies the 3-linear forms condition. If 0 g ν, 0 g 1 and g g = o(1), then E[g(x, y)g(x, z)g(y, z)] = E[ g(x, y) g(x, z) g(y, z)] + o(1). Proof ingredients 1 Cauchy-Schwarz 2 Densification 3 Apply cut norm/discrepancy (as in dense case)

73 Densification x E[g(x, z)g(y, z)g(x, z )g(y, z )] y z z

74 Densification x E[g(x, z)g(y, z)g(x, z )g(y, z )] y z z Set g (x, y) := E z [g(x, z )g(y, z )], i.e., codegrees g (x, y) 1 for almost all (x, y)

75 Densification x E[g(x, z)g(y, z)g(x, z )g(y, z )] = E[g (x, y)g(x, z)g(y, z)] y z Set g (x, y) := E z [g(x, z )g(y, z )], i.e., codegrees g (x, y) 1 for almost all (x, y)

76 Densification x E[g(x, z)g(y, z)g(x, z )g(y, z )] = E[g (x, y)g(x, z)g(y, z)] y z Set g (x, y) := E z [g(x, z )g(y, z )], i.e., codegrees g (x, y) 1 for almost all (x, y) Made X Y dense. Now repeat for X Z & Y Z. Reduce to dense setting.

77 Transference Start with f ν (sparse) f : Z N [0, ) Ef δ Dense model theorem: one can approximate f (in cut norm) by (dense) f : ZN [0, 1] E f = Ef Counting lemma implies E x,d [f (x)f (x + d)f (x + 2d)] E x,d [ f (x) f (x + d) f (x + 2d)] c [By Roth s Thm (weighted version)] = relative Roth theorem

78 Coming Soon The Green-Tao theorem: an exposition

79 Coming Soon The Green-Tao theorem: an exposition A gentle exposition giving a complete & self-contained proof of the Green-Tao theorem (other than a black-box application of Szemerédi s theorem) 25 pages

80 Relative Szemerédi theorem (Conlon, Fox, Z.) If ν : Z N [0, ) satisfies the k-linear forms condition, then any f with 0 f ν and E x,d ZN [f (x)f (x + d)f (x + 2d) f (x + (k 1)d)] = o(1) must satisfy Ef = o(1). 3-linear forms condition: (x, x, y, y, z, z Unif(Z N )) E[ν(2x + y)ν(2x + y)ν(2x + y )ν(2x + y ) ν(x z)ν(x z)ν(x z )ν(x z ) ν( y 2z)ν( y 2z)ν( y 2z )ν( y 2z )] = 1 + o(1) as well as if any subset of the 12 factors were deleted. 4-linear forms condition: E[ν(3w + 2x + y) ] = 1 + o(1) THANK YOU!

The Green-Tao theorem and a relative Szemerédi theorem

The Green-Tao theorem and a relative Szemerédi theorem Yufei Zhao Massachusetts Institute of Technology Based on joint work with David Conlon (Oxford) and Jacob Fox (MIT) Green Tao Theorem (arxiv 2004;