WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES

Size: px
Start display at page:

Download "WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES"

Transcription

1 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES BENJAMIN SCHWEINHART Abstract. We study the asyptotic behavior of rando variables of the for Eα i x,..., x n = d b α b,d PH ix,...,x n where { x j are i.i.d. saples fro a probability easure on a triangulable }j N etric space, and PH i x,..., x n denotes the i-diensional reduced persistent hoology of the Čech coplex of {x,..., x n }. These quantities are a higherdiensional generalization of the α-weighted su of a inial spanning tree; we seek to prove analogues of the theores of Steele [6] and Aldous and Steele [2] in this context. As a special case of our ain theore, we show that if { x j are distributed }j N independently and uniforly on the -diensional Euclidean sphere, α <, and 0 i < n, then there are real nubers γ and Γ so that α γ li n E α i x,..., x n Γ in probability. More generally, we prove results about the asyptotics of the expectation of Eα i for points sapled fro a locally bounded probability easure on a space that is the bi-lipschitz iage of an diensional Euclidean siplicial coplex.. Introduction We are interested in rando variables of the for Eα i x,..., x n = b,d PH i x,...,x n d b α where { x j are independent saples drawn fro a probability easure on a }j N triangulable etric space, and PH i x,..., x n denotes the i-diensional reduced persistent hoology of the Čech coplex of {x,..., x n }. The special case i = 0 is, Date: July 208.

2 2 BENJAMIN SCHWEINHART under a different guise, already the subject of an expansive literature in probabilistic cobinatorics; Eα 0 x gives the α-weight of the inial spanning tree on a finite subset of a etric space x, T x : Eα 0 x = 2 α e α In 988, Steele [6] showed the following: e T x Theore Steele. Let µ is a copactly supported probability distribution on R, and let {x n } n N be i.i.d. saples fro µ. If α <, α li n E 0 α x,..., x n c α, f x α/ R d with probability one, where f x is the probability density of the absolutely continuous part of µ, and c α, is a positive constant that depends only on α and. In 992, Aldous and Steele [2] showed that if {x i } i N sapled independently fro the unifor distribution on the unit cube in R, then li E α x,..., x n c d, d in the L 2 sense. Under the sae hypotheses, Kesten and Lee proved the following central liit theore in 996 [2]: Eα 0 X,..., X n E Eα 0 X,..., X n N 0, σ 2 n 2α α,d 2d in distribution, for any α > 0. Here, we take the first step toward a higher-diensional generalization of these celebrated results. Another special case of Eα i x α = gives the total lifetie persistence of x. Rando variables of the for E i x have been investigated by Hiraoka and Shirai [] in the context of Linial Meshula processes. They showed that if X is sapled fro the -Linial Meshula process then E E X O n which is a higher-diensional generalization of Frieze s ζ 3-theore for Erdós Rényi rando graphs [0]. Also, Adas et al. [] studied the behavior of the lifetie persistence of rando easures on Euclidean space, perforing coputational experients and conjecturing the existence of a liit function capturing finer properties of the persistent hoology.

3 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 3 The properties of E i α x for general i and n have until now, as far as we know, not been studied in a probabilistic context see the note at the end of the introduction. However, soe work has been done in the extreal context. In 200, Cohen-Steiner et al. [6] showed that if M is the bi-lipschitz iage of an -diensional siplicial coplex and α >, then E α i X is uniforly bounded for X M. We use their results to prove the upper bounds in Section 2. Furtherore, in our previous paper [3] we related the upper box diension of a subset X of a etric space to the behavior of E i α Y for extreal subsets Y X. We will say ore about the relation of this to the present work in Section.2... Our Results. The following are special cases of our ain theore: Theore 2. Let { x j }j N be be distributed independently and uniforly on the Sn. If α <, 0 i < n, and persistent hoology is taken with respect to the intrinsic etric on S n, γ li n α E α i x,..., x n Γ in probability, where γ and Γ are constants that depend on µ and α. Furtherore, there exists a D R so that in probability. li log n E i x,..., x n D Theore 3. Let { x j be be distributed independently and uniforly on an - }j N diensional Euclidean ball. If α <, 0 i < n, γ li n α E E α i x,..., x n Γ where γ and Γ are constants that depend on µ and α. In fact, the lower bound holds in probability. Furtherore, there exists a D R so that li log n E E i x,..., x n D We show a stronger result for copactly supported probability easures on R 2 that are locally bounded:

4 4 BENJAMIN SCHWEINHART Definition. A probability easure µ on R is locally bounded if there is a A R with positive volue and real nubers a a 0 > 0 so that for all Borel sets B A. a 0 vol B µ B a vol B Theore 4. Let µ is a copactly supported, locally bounded probability easure on R 2, and let {x n } n N be i.i.d. saples fro µ. If α <, γ li n α E α x,..., x n Γ in probability. In fact, the upper bound holds with probability one. Furtherore, there exists a constant D so that with probability one li log n E 2 x,..., x n D More generally, we prove results for locally bounded probability easures on spaces that are the bi-lipschitz iage of a copact, -diensional Euclidean siplicial coplex: Definition 2. Let M be the bi-lipschitz iage of a copact -diensional Euclidean siplicial coplex M under a ap φ M. A probability easure µ on M is locally bounded if there exists a subset A M with positive -diensional volue, and real nubers a a 0 > 0 so that a 0 for all Borel sets B A. vol B vol M µ φ vol B M B a vol M For exaple, a the unifor easure on a -diensional Rieannian anifold is locally bounded, as is any easure that is locally bounded with respect to the Rieannian volue. While there exist etric spaces M with point sets { x j }j N so that PH i x,..., x n O n this is thought to be soewhat pathological behavior [3].

5 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 5 Definition 3. A probability easure µ on a triangulable etric space has linear PH i expectation if E PH i {x,..., x n } O n Siilarly, µ has linear PH i variance if E PH i {x,..., x n } E PH i {x,..., x n } 2 O n For exaple, the unifor easure on a Euclidean ball [8] and any positive, continuous probability density on the Euclidean n-sphere [7] has linear PH i expectation. It is ore difficult to prove that a probability easure has linear PH i variance. As far as we are aware, this is only known for probability easures on R 2 and the unifor easure on the n-diensional Euclidean sphere [7] see Equation and Proposition 3. Theore 5. Let M be the bi-lipschitz iage of an -diensional Euclidean siplicial coplex, and 0 i <. If µ is a locally bounded probability easure on M, there are real nubers 0 < γ < Γ so that γn α E E i α x,..., x n Γ E PH i {x,..., x n } α for all sufficiently large n. In particular, if µ has linear PH i expectation, there is a real nuber Γ 0 so that γ li n α E E α i x,..., x n Γ 0 The lower bound holds in probability, and the upper bound does if µ has linear PH i variance. Furtherore, there exists a real nuber D so that E En i x,..., x n D log E PH i x,..., x n where analogously sharper stateents hold if µ has linear PH i expectation or variance. We prove the upper bound in Proposition 2 and the lower bound in Proposition 5. After copletion of this anuscript, we becae aware that Divol and Polonik [7] independently and concurrently proved a sharper result for the persistent hoology of points sapled fro bounded, absolutely continuous probability densities on [0, ]. We believe this anuscript is still useful in that the proofs are largely self-contained, and the ethods are applicable to other situations. In a later paper [4], we use the

6 6 BENJAMIN SCHWEINHART to study the behavior of E i α x,..., x n for i.i.d. points sapled fro a easure supported on a set of fractional diension..2. PH -diension. In [3], we defined a faily of persistent hoology diensions for a subset X of a etric space M in ters of the extreal behavior of Eα i Y for subsets x of X: { } di i PH X = inf α : Eα i x < C x X That is, E i α x is uniforly bounded for all α > di i PH X, but not for α < di i PH X. Note that the persistent hoology is taken with respect M. Our results were the first rigorously relating persistent hoology to a classically defined fractal diension, the upper box diension, but the definition is difficult to copute with in practice. Here, we define a siilar notion of fractal diension for easures on a etric space that ay be ore coputable in practice: Definition 4. The PH i -diension of a probability easure on a a triangulable etric space is { } di i PH µ = sup α : li sup E Eα i x,..., x n = Clearly, di i PH µ di i PH supp µ. As a corollary to our ain theore, we show: Theore 6. Let M be the bi-lipschitz iage of a copact -diensional Euclidean siplicial coplex, and 0 i <. If µ is a locally bounded probability easure on M, di i PH µ =.3. Persistent Hoology. If X is a bounded subset of a triangulable etric space M, let X ɛ denote the ɛ-neighborhood of X : X ɛ = {x M : d x, X < ɛ} Also, let H i X be the reduced hoology of X, with coefficients in a field k. The persistent hoology of X is the product ɛ>0 H i X ɛ, together with the inclusion aps i ɛ0,ɛ : H i X ɛ0 H i X ɛ for ɛ 0 < ɛ. The structure of persistent hoology is captured by a set of intervals, which we refer to as PH i X [8]. These intervals represent how the topology of X ɛ changes as ɛ increases. Under certain finiteness hypotheses which are satisfied if X is a finite point set PH i X is the unique

7 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 7 set of intervals so that the rank of i ɛ0,ɛ equals the nuber of intervals containing ɛ 0, ɛ [5]. If X is finite PH i X is the sae as the persistent hoology of the Čech coplex of X. Note that this depends on the abient etric space. Here, if µ is a probability easure on M and { x j }j N are sapled fro µ, then PH i x,..., x n is the persistent hoology with abient etric space M. All questions we study here would also be interesting in the context of the Vietoris Rips Coplex..4. Notation. In the following, an -space will be the bi-lipschitz iage of a copact -diensional Euclidean siplicial coplex. Also, if the easure µ is obvious fro the context, { x j will denote a collection of independent rando }j N variables with coon distribution µ. Also, x n will be shorthand for {x,..., x n } and x will denote a finite point set. 2. Upper Bounds Our strategy to prove an upper bound for the asyptotics of E i α {x,..., x n } will be to bound the nuber and length of the persistent hoology intervals in ters of the nuber of siplices in a triangulation of the abient etric space. The approach is siilar to that in our earlier paper [3]. 2.. Preliinaries. We require the following result, which is proven by bounding the nuber of persistent hoology intervals of a triangulable etric space of length greater than δ in ters of the nuber of siplices in a triangulation of esh δ: Proposition. Cohen-Steiner, Edelsbrunner, Harer, and Mileyko [6] Let M be an -space. There exists a real nuber C 0 so that for any 0 i <, X M, and δ > 0, {b, d PH i X : d b > δ} C 0 δ We use this result to bound E i α x in ters of the nuber of PH i intervals of x: Lea. Let M be an -space, α <, and i N. There exists a real nuber C > 0 so that Eα i X C PH i X α for all X M. Furtherore, there exists a real nuber D > 0 so that E i X D log PH i X

8 8 BENJAMIN SCHWEINHART for all X M. Proof. Dilating M by a factor r ultiplies E i α X by r α, so we ay assue without loss of generality that the diaeter of M is less than one. Let n = PH i X and I k = { b, d PH i X : Also, let C 0 be as in Proposition so I k C 0 2 k 2 < d b } k+ 2 k The largest C 0 intervals of PH i X each have length less than or equal to 2 0, the next largest C 0 2 intervals have length less than or equal to 2, and so on. It follows that if then log2 2n/C 0 l = n l C 0 2 k k=0 and E i α X l α C 0 2 k 2 k k=0 If α =, the previous inequality becoes as desired. E i α X C 0 l = O log n

9 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 9 Otherwise, if α <, E i α X l C 0 2 k α k=0 2 αl+ = C 0 2 α C 0 2 α 2 αl+ C 0 2 α 2 = C n α log 2 2n/C 0 +2 α where C = C 04 α 2 α The Upper Bound. The upper bound in our ain theore now follows iediately fro Jensen s inequality, as the function f x = x α is concave for 0 < α : Proposition 2. Let M be an -space, let i be a natural nuber less than, and let µ be a locally bounded probability easure on M. For all 0 < α < there exists a real nuber C > 0 so that E Eα i x,..., x n C E PH i x,..., x n α In particular, if µ has linear PH i expectation and linear PH i variance then there is a C > 0 so that α li n E i α x,..., x n C in probability. Furtherore, there exists a real nuber D so that E E i x,..., x n D log PH i x,..., x n In particular, if µ has linear PH i expectation and linear PH i variance then there is a D > 0 so that li log n Ei x,..., x n D in probability.

10 0 BENJAMIN SCHWEINHART Proof. Let x be a finite subset of B, and let C be as in Lea. If α <, E Eα i x E C PH i x α by Lea C E PH i x α by Jensen s inequality as desired. If µ has linear PH i expectation and linear PH i -variance, Chebyshev s Inequality iplies that li PH i x,..., x n /n C 2 in probability, for soe C 2 > 0, and the desired stateent follows fro Lea. The proof for the case α = is siilar Sharper Upper Bounds. Our sharper upper bounds in Theores 2 and 3 follow fro the fact that if {x,..., x n } is a finite subset of R of S in general position then PH i x,..., x n DT x,..., x n where DT x,..., x n is the nuber of siplices of the Delaunay triangulation on {x,..., x n }. In fact, the Alpha coplex is a filtration on the siplices of the Delaunay triangulation that is hootopy equivalent to the ɛ-neighborhood filtration of the points {x,..., x n } [9]. This construction is usually defined for points in Euclidean space, but easily extends to points on the -sphere, in which case the Delaunay triangulation is the spherical convex hull of the points. Proposition 3. If B be a bounded subset of R Eα i x,..., x n = O n + 2 α for any general position point set {x,..., x n } contained in B. Proof. The Upper Bound Theore [5] iplies that if X R then DT x,..., x n = O n + 2 The desired stateent follows iediately fro Lea and Equation

11 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 3. Lower Bounds Our strategy to prove lower bounds for the asyptotics of weighted PH -sus is to study collections of sets whose persistent hoology obeys a super-additivity property. We define certain cubical occupancy events giving rise to such collections, and prove that they occur with positive probability for sets of i.i.d. points drawn fro a locally bounded probability easure on an -space. We bootstrap these results by subdividing a subset of an -diensional cube into any sall sub-cubes. This bootstrapping arguent is siilar to the one we used to prove a lower bound for PH i diension in our previous paper [3]. In the following, fix 0 i <. 3.. Super-additivity for Persistent Hoology. Persistent hoology does not in general obey a super-additivity property, but we can define a subclass of sets whose persistent hoology does. If X and T are subsets of a triangulable etric space and b < d, let M X,T b, d be the rank of the hooorphis on hoology induced by the inclusion X b X d X d T d where X ɛ denotes the ɛ-neighborhood of X. Note that M X,C b, d N X b, d where N X b, d is the nuber of intervals of PH i X with birth ties less than b and death ties greater than d. We will show that if C is an -diensional cube and X C, then quantities of the for M X, C d, b obey a super-additivity property. Lea 2. Let {C,..., C n } be -diensional cubes in R so that C j C k C j If X j C j for j =,..., n for any 0 b < d. N j X j b, d M j X j, j C j b, d j, k {,..., n} : j k n M Xj, C j b, d Proof. Let k {,..., n}, S = k j= X j, T = n j= C j, X = X k, and C = C k. See Figure. j=

12 2 BENJAMIN SCHWEINHART Figure. The setup in the proof of Lea 2. We consider the cases i = and i < separately. If i =, Alexander Duality iplies that N S b, d is the nuber of bounded coponents of the copleent of S b that intersect non-trivially with the copleent of S d. Siilarly, M X,C b, d is the nuber of bounded coponents of the copleent of X b that intersect non-trivially with X d C d c. Note that all bounded coponents of Xb c are contained within the interior of C, because C is convex and separates R into two coponents. Let Y be a coponent of the copleent of X b that intersects non-trivially with Xd C d c, and let y Y Xd C d c. C separates R into two coponents so Therefore, d y, S d y, S T = d y, X C > d Y S d c Y S d T d c = Y X d C d c Applying the sae arguent to each X j and counting coponents of the copleent yields the desired inequalities. Otherwise, assue that i. We will show that M S X,T b, d M S,T b, d + M X, C b, d and the desired result will follow by induction. Note that X ɛ S ɛ X ɛ S ɛ T ɛ C ɛ for any ɛ > 0. Consider the following coutative diagra of inclusion hooorphiss and Mayer-Vietoris sequences:

13 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 3 α b +β b H i X b S b H i X b H i S b H i X b S b 0 = H i Cd φ ψ ζ α d +β d H i Xd C d Hi S d T d H i X d S d T d Observe that M X, C b, d = rank φ, M S,T b, d = rank ψ, and M X S,T b, d = rank ζ. It follows that M X S,T b, d = rank ζ rank α d + β d φ ψ = rank φ ψ because H i Cd = 0 = rank φ + rank ψ = M X, C b, d + M S,T b, d k M Xj, C j b, d j= by induction 3.2. Occupancy Events. If B is a subset of an -space, define the occupancy event { 0 x B = 0 δ B, x = x B > 0 Also, if {A i } r i= and { } s B j, are collections of subsets of M, let j= ξ x, {A i }, { B j } = { δ A i, x = 0 and δ B j, x = i, j 0 otherwise Lea 3. Let µ be a locally bounded probability easure on an -space M. There exists a real nuber V 0 > 0 so for any r, s N there there exists a real nuber γ 0 > 0 so that for any collections of disjoint, congruent cubes { } { } A k i and Bj k, for i {,..., r}, j {,..., s}, and k {,..., n} for a total of r + s n cubes with volue

14 4 BENJAMIN SCHWEINHART Figure 2. The setup in the proof of Lea 4. then for all sufficiently large n. vol A k i = vol Bj k = V 0 /n i, j, k n P ξ k= { } { x, A k i, } Bj k s γ 0 Proof. The proof is nearly identical to that of Lea 3 in [4]. Lea 4. Let 0 < b < d < /6, and V 0 > 0. There exists a λ 0 > 0 so that if C R is an -diensional cube of width R and λ > λ 0, there exist disjoint, congruent cubes { } A j and {Bk } of width R V 0 /λ so that ξ x, { } A j, {Bk } = = M x, C Rb, Rd > 0 Proof. We ay assue without loss of generality that R = and C is centered at the origin. Let S i R be an i diensional sphere of diaeter /3 centered at the origin; note that PH i S i consists of a single interval 0, /6. Let κ = in b, /6 d and 0 = κ/. li δ /δ = δ 0

15 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 5 so there is a real nuber > 0 so that δ /δ < κ for all δ <. Set λ 0 = V 0 in 0, Choose λ > λ 0, set δ = V 0 /λ, and let C be the cube of width δ /δ centered at the origin. Subdivide C into /δ sub-cubes of width δ. Call this collection of sub-cubes {C l } and let { } { } Aj = c {C l } : S i c = and {B k } = See Figure 2 for an illustration. If x C and the event ξ x, { } A j, {Bk } occurs, then d H x C, S i < κ { } c {C l } : S i c where d H is the Hausdorff distance and we used the fact that the diagonal of an -diensional cube of width δ is δ. The stability of the bottleneck distance [5] iplies that PH i x C includes an interval ˆb, ˆd so that ˆb < κ b < d /3 κ < ˆd In particular, N x C b, d > 0 By construction, 2 d x C, C \ C > 2 δ d C, C > 6 6 κ d so the ɛ-neighborhoods of x C and C \ C are disjoint for all ɛ d. It follows that the aps on hoology induced by the inclusions x C ɛ x ɛ and x ɛ x ɛ C ɛ are injective for all ɛ d. Therefore, M x, C b, d > 0, as desired Proof of the Lower Bound. In the reainder, let µ be a locally bounded probability easure on an -space M, let { x j be i.i.d. saples fro µ, and let }j N x n = {x,..., x n }. Also, let C be as in Lea 3, and rescale M if necessary so that C is a unit cube. Finally, let λ 0 be as in Lea 4.

16 6 BENJAMIN SCHWEINHART The Euclidean Case. For clarity, we first consider the special case where φ M is the identity ap, and µ is a locally bounded probability easure on a copact Euclidean siplicial coplex. The arguent for the general case contains any of the sae eleents. Lea 5. Let 0 < b 0 < d 0 < /6. If n 0 > λ 0, there is a γ > 0 so that li n N x n n0 n0 b0, d0 > γ n n in probability. Proof. Let V 0 be as in Definition 3, and let r = A i and s = Bj, where {Ai } and { Bj } are as in the previous lea. Assuing n > n 0, let ω = n 0 n. Subdivide R into cubes of width ω, and let {D l } Kn l= be the cubes that are fully contained in C. Note that K n := {D l } n/n 0 By the previous lea, there are collections of disjoint, congruent sub-cubes { A l,..., A l r and { B, l..., Bs} l of width ω V0 /n 0 contained inside each cube Dl so that { } { ξ x n, A l i, Bj} l = = M xn Dl, D ωb l 0, ωd 0 > 0 } Note that N xn ωb 0, ωd 0 K n l= K n l= M xn D l, D l ωb 0, ωd 0 by Lea 2 ξ { } { } x n, A l i, Bj l Let γ 0 be as in Lea 3 and γ < γ 0 /n 0. Set δ = + γ n 0 γ 0 2 and ɛ = δ δ

17 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 7 so /2 < δ < and 0 < ɛ <. Also, find a N so that K n > δn/n 0 for all n > N. Note that 2 γn = γ 0δn δ < ɛ γ 0 K n n 0 δ for all n > N. Therefore, if n > N, P N xn ωb 0, ωd 0 > γn which converges to as n. K n P ξ l= { } { x n, A l i, Bj} l > γn P B K n, γ 0 > γn by Lea 3 P B K n, γ 0 > ɛ γ 0 K n by Equation 2 We can now prove the lower bound in the Euclidean setting: Proposition 4. There is a γ > 0 so that in probability. α li n E i α x n γ Proof. Let 0 < b < d < /6, and let n 0 > λ 0 and γ be as before. Also, let ω = n 0 n /. We have that li in probability. α n E i α x n α li n ωd ωb α N xn ωb, ωd n α/ 0 = li d bα N xn ωb, ωd n n α/ 0 d b α γ by Lea 5 := γ

18 8 BENJAMIN SCHWEINHART The General Case. Before proving the lower bound in our ain theore, we require an interleaving result for the persistent hoology of iages of bi-lipschitz aps: Lea 6. Let M 0 and M be etric spaces and let ψ : M 0 M be L-bilipshitz. If X M 0 and 0 b 0 < d 0 N X b 0 /L, Ld 0 N ψx b 0, d 0 N X Lb 0, d 0 /L Proof. Fix i N, and let j ɛ0,ɛ : X ɛ0 X ɛ and k ɛ0,ɛ : φ X ɛ0 φ X ɛ denote the inclusion aps for ɛ 0 ɛ. By the definition of a bi-lipschitz ap L d M 0 x, y d M ψ x, ψ y Ld M0 x, y for all x, y M 0. In particular, we have the following inclusions: ψ X b0 /L ψ Xb0 ψ X d0 ψ X Ld0 It follows that the rank of ap on hoology induced by i b0 /L,Ld 0 is less than or equal to the rank of the ap induced by j b0,d 0 where we have used that a bi-lipschitz ap is a hoeoorphis. Therefore, N X b 0 /L, Ld 0 N ψx b 0, d 0 The arguent for the other inequality is very siilar. Proposition 5. Let µ be a locally bounded probability easure on an -space M and 0 i <. There is a γ > 0 so that in probability li α n E i α x,... x n > γ Proof. Let L be the bi-lipschitz constant of φ M, and choose b, d > 0 so that Set so L 2 b < d < /6 n 0 = ax d/l Lb, n 0 3 n 0 d/l Lb

19 WEIGHTED PERSISTENT HOMOLOGY SUMS OF RANDOM ČECH COMPLEXES 9 Let ω = n 0 Lea 5 to y n. First, n E i α x n and y n = φ x n. Our strategy is to bound E i α x n by applying ω d/l Lb α N xn ωlb, ωd/l n α/ N xn ωlb, ωd/l by Equation 3 n α/ N yn ωb, ωd by Lea 6 = n α/ N yn ωb, ωd Therefore, li α n E i α x n li n N y n ωb, ωd > γ in probability, where γ > 0 is as given in Lea 5. References [] Henry Adas, Manuchehr Ainian, Elin Farnell, Michael Kirby, Joshua Mirth, Rachel Neville, Chris Peterson, Patrick Shipan, and Clayton Shonkwiler. A fractal diension for easures via persistent hoology. Preprint, 208. [2] David Aldous and J. Michael Steele. Asyptotics for Euclidean inial spanning trees on rando points. Probability Theory and Related Fields, 992. [3] Ulrich Bauer and Florian Pausinger. Persistent Betti nubers of rando Čech coplexes. arxiv: [4] Oer Bobrowski, Matthew Kahle, and Prioz Skraba. Maxially persistent cycles in rando geoetric coplexes. The Annals of Applied Probability, 207. [5] David Cohen-Steiner, Herbert Edelsbrunner, and John Harer. Stability of persistence diagras. Discrete & Coputational Geoetry, 37:0320, [6] David Cohen-Steiner, Herbert Edelsbrunner, John Harer, and Yuriy Mileyko. Lipschitz functions have l p -stable persistence. Foundations of Coputational Matheatics, 200. [7] Vincent Divol and Wolfgang Polonik. On the choice of weight functions for linear representations of persistence diagras. arxiv: , July 208. [8] Rex Dwyer. Higher-diensional voronoi diagras in linear expected tie. Discrete and Coputational Geoetry, 99. [9] H. Edelsbrunner, D. Letscher, and A. Zoorodian. Topological persistence and siplificitaion. Discrete and Coputational Geoetry, [0] A. M. Frieze. On the value of a rando iniu spanning tree proble. Discrete Applied Matheatics, 985.

20 20 BENJAMIN SCHWEINHART [] Yasuaki Hiraoka and Tooyuki Shirai. Miniu spanning acycle and lifetie of persistent hoology in the Linial Meshula process. Rando Structures & Algoriths, 207. [2] Harry Kesten and Sungchul Lee. The central liit theore for weighted inial spanning trees on rando points. Annals of Applied Probability, 996. [3] B. Schweinhart. Persistent hoology and the upper box diension. arxiv: [4] Benjain Schweinhart. The persistent hoology of rando geoetric coplexes on fractals. arxiv: [5] Richard Stalney. The upper bound conjecture and Cohen-Macaulay rings. Studies in Applied Matheatics, 975. [6] J. Michael Steele. Growth rates of Euclidean inial spanning trees with power weighted edges. Annals of Probability, 988. [7] Johannes Steeseder. Rando polytopes with vertices on the sphere. PhD thesis, Paris-Lodron University of Salzburg, 204. [8] A. Zoorodian and G. Carlsson. Coputing persistent hoology. Discrete and Coputational Geoetry, 33: , 2005.

THE PERSISTENT HOMOLOGY OF RANDOM GEOMETRIC COMPLEXES ON FRACTALS

THE PERSISTENT HOMOLOGY OF RANDOM GEOMETRIC COMPLEXES ON FRACTALS THE PERSISTENT HOMOLOGY OF RANDOM GEOMETRIC COMPLEXES ON FRACTALS BENJAMIN SCHWEINHART Abstract. We study the asymptotic behavior of the persistent homology of i.i.d. samples from a d-ahlfors regular measure

More information

Supplement to: Subsampling Methods for Persistent Homology

Supplement to: Subsampling Methods for Persistent Homology Suppleent to: Subsapling Methods for Persistent Hoology A. Technical results In this section, we present soe technical results that will be used to prove the ain theores. First, we expand the notation

More information

arxiv: v5 [math.mg] 6 Sep 2018

arxiv: v5 [math.mg] 6 Sep 2018 PERSISTENT HOMOLOGY AND THE UPPER BOX DIMENSION BENJAMIN SCHWEINHART arxiv:1802.00533v5 [math.mg] 6 Sep 2018 Abstract. We introduce a fractal dimension for a metric space based on the persistent homology

More information

Computational and Statistical Learning Theory

Computational and Statistical Learning Theory Coputational and Statistical Learning Theory Proble sets 5 and 6 Due: Noveber th Please send your solutions to learning-subissions@ttic.edu Notations/Definitions Recall the definition of saple based Radeacher

More information

Research Article On the Isolated Vertices and Connectivity in Random Intersection Graphs

Research Article On the Isolated Vertices and Connectivity in Random Intersection Graphs International Cobinatorics Volue 2011, Article ID 872703, 9 pages doi:10.1155/2011/872703 Research Article On the Isolated Vertices and Connectivity in Rando Intersection Graphs Yilun Shang Institute for

More information

E0 370 Statistical Learning Theory Lecture 5 (Aug 25, 2011)

E0 370 Statistical Learning Theory Lecture 5 (Aug 25, 2011) E0 370 Statistical Learning Theory Lecture 5 Aug 5, 0 Covering Nubers, Pseudo-Diension, and Fat-Shattering Diension Lecturer: Shivani Agarwal Scribe: Shivani Agarwal Introduction So far we have seen how

More information

Prerequisites. We recall: Theorem 2 A subset of a countably innite set is countable.

Prerequisites. We recall: Theorem 2 A subset of a countably innite set is countable. Prerequisites 1 Set Theory We recall the basic facts about countable and uncountable sets, union and intersection of sets and iages and preiages of functions. 1.1 Countable and uncountable sets We can

More information

The isomorphism problem of Hausdorff measures and Hölder restrictions of functions

The isomorphism problem of Hausdorff measures and Hölder restrictions of functions The isoorphis proble of Hausdorff easures and Hölder restrictions of functions Doctoral thesis András Máthé PhD School of Matheatics Pure Matheatics Progra School Leader: Prof. Miklós Laczkovich Progra

More information

4 = (0.02) 3 13, = 0.25 because = 25. Simi-

4 = (0.02) 3 13, = 0.25 because = 25. Simi- Theore. Let b and be integers greater than. If = (. a a 2 a i ) b,then for any t N, in base (b + t), the fraction has the digital representation = (. a a 2 a i ) b+t, where a i = a i + tk i with k i =

More information

L p moments of random vectors via majorizing measures

L p moments of random vectors via majorizing measures L p oents of rando vectors via ajorizing easures Olivier Guédon, Mark Rudelson Abstract For a rando vector X in R n, we obtain bounds on the size of a saple, for which the epirical p-th oents of linear

More information

CENTRE FOR STOCHASTIC GEOMETRY AND ADVANCED BIOIMAGING. RESEARCH REPORT. Christophe A.N. Biscio and Jesper Møller

CENTRE FOR STOCHASTIC GEOMETRY AND ADVANCED BIOIMAGING.   RESEARCH REPORT. Christophe A.N. Biscio and Jesper Møller CENTRE FOR STOCHASTIC GEOMETRY AND ADVANCED BIOIMAGING www.csgb.dk RESEARCH REPORT 2016 Christophe A.N. Biscio and Jesper Møller The accuulated persistence function, a new useful functional suary statistic

More information

The degree of a typical vertex in generalized random intersection graph models

The degree of a typical vertex in generalized random intersection graph models Discrete Matheatics 306 006 15 165 www.elsevier.co/locate/disc The degree of a typical vertex in generalized rando intersection graph odels Jerzy Jaworski a, Michał Karoński a, Dudley Stark b a Departent

More information

arxiv: v1 [math.nt] 14 Sep 2014

arxiv: v1 [math.nt] 14 Sep 2014 ROTATION REMAINDERS P. JAMESON GRABER, WASHINGTON AND LEE UNIVERSITY 08 arxiv:1409.411v1 [ath.nt] 14 Sep 014 Abstract. We study properties of an array of nubers, called the triangle, in which each row

More information

Polytopes and arrangements: Diameter and curvature

Polytopes and arrangements: Diameter and curvature Operations Research Letters 36 2008 2 222 Operations Research Letters wwwelsevierco/locate/orl Polytopes and arrangeents: Diaeter and curvature Antoine Deza, Taás Terlaky, Yuriy Zinchenko McMaster University,

More information

Statistics and Probability Letters

Statistics and Probability Letters Statistics and Probability Letters 79 2009 223 233 Contents lists available at ScienceDirect Statistics and Probability Letters journal hoepage: www.elsevier.co/locate/stapro A CLT for a one-diensional

More information

M ath. Res. Lett. 15 (2008), no. 2, c International Press 2008 SUM-PRODUCT ESTIMATES VIA DIRECTED EXPANDERS. Van H. Vu. 1.

M ath. Res. Lett. 15 (2008), no. 2, c International Press 2008 SUM-PRODUCT ESTIMATES VIA DIRECTED EXPANDERS. Van H. Vu. 1. M ath. Res. Lett. 15 (2008), no. 2, 375 388 c International Press 2008 SUM-PRODUCT ESTIMATES VIA DIRECTED EXPANDERS Van H. Vu Abstract. Let F q be a finite field of order q and P be a polynoial in F q[x

More information

arxiv: v2 [math.kt] 3 Apr 2018

arxiv: v2 [math.kt] 3 Apr 2018 EQUIVARIANT K-HOMOLOGY FOR HYPERBOLIC REFLECTION GROUPS arxiv:170705133v [athkt] 3 Apr 018 JEAN-FRANÇOIS LAFONT, IVONNE J ORTIZ, ALEXANDER D RAHM, AND RUBÉN J SÁNCHEZ-GARCÍA Abstract We copute the equivariant

More information

PREPRINT 2006:17. Inequalities of the Brunn-Minkowski Type for Gaussian Measures CHRISTER BORELL

PREPRINT 2006:17. Inequalities of the Brunn-Minkowski Type for Gaussian Measures CHRISTER BORELL PREPRINT 2006:7 Inequalities of the Brunn-Minkowski Type for Gaussian Measures CHRISTER BORELL Departent of Matheatical Sciences Division of Matheatics CHALMERS UNIVERSITY OF TECHNOLOGY GÖTEBORG UNIVERSITY

More information

MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.436J/15.085J Fall 2008 Lecture 11 10/15/2008 ABSTRACT INTEGRATION I

MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.436J/15.085J Fall 2008 Lecture 11 10/15/2008 ABSTRACT INTEGRATION I MASSACHUSETTS INSTITUTE OF TECHNOLOGY 6.436J/15.085J Fall 2008 Lecture 11 10/15/2008 ABSTRACT INTEGRATION I Contents 1. Preliinaries 2. The ain result 3. The Rieann integral 4. The integral of a nonnegative

More information

Testing Properties of Collections of Distributions

Testing Properties of Collections of Distributions Testing Properties of Collections of Distributions Reut Levi Dana Ron Ronitt Rubinfeld April 9, 0 Abstract We propose a fraework for studying property testing of collections of distributions, where the

More information

E0 370 Statistical Learning Theory Lecture 6 (Aug 30, 2011) Margin Analysis

E0 370 Statistical Learning Theory Lecture 6 (Aug 30, 2011) Margin Analysis E0 370 tatistical Learning Theory Lecture 6 (Aug 30, 20) Margin Analysis Lecturer: hivani Agarwal cribe: Narasihan R Introduction In the last few lectures we have seen how to obtain high confidence bounds

More information

arxiv: v2 [math.co] 3 Dec 2008

arxiv: v2 [math.co] 3 Dec 2008 arxiv:0805.2814v2 [ath.co] 3 Dec 2008 Connectivity of the Unifor Rando Intersection Graph Sion R. Blacburn and Stefanie Gere Departent of Matheatics Royal Holloway, University of London Egha, Surrey TW20

More information

A := A i : {A i } S. is an algebra. The same object is obtained when the union in required to be disjoint.

A := A i : {A i } S. is an algebra. The same object is obtained when the union in required to be disjoint. 59 6. ABSTRACT MEASURE THEORY Having developed the Lebesgue integral with respect to the general easures, we now have a general concept with few specific exaples to actually test it on. Indeed, so far

More information

arxiv: v1 [cs.ds] 3 Feb 2014

arxiv: v1 [cs.ds] 3 Feb 2014 arxiv:40.043v [cs.ds] 3 Feb 04 A Bound on the Expected Optiality of Rando Feasible Solutions to Cobinatorial Optiization Probles Evan A. Sultani The Johns Hopins University APL evan@sultani.co http://www.sultani.co/

More information

Sharp Time Data Tradeoffs for Linear Inverse Problems

Sharp Time Data Tradeoffs for Linear Inverse Problems Sharp Tie Data Tradeoffs for Linear Inverse Probles Saet Oyak Benjain Recht Mahdi Soltanolkotabi January 016 Abstract In this paper we characterize sharp tie-data tradeoffs for optiization probles used

More information

Characterization of the Line Complexity of Cellular Automata Generated by Polynomial Transition Rules. Bertrand Stone

Characterization of the Line Complexity of Cellular Automata Generated by Polynomial Transition Rules. Bertrand Stone Characterization of the Line Coplexity of Cellular Autoata Generated by Polynoial Transition Rules Bertrand Stone Abstract Cellular autoata are discrete dynaical systes which consist of changing patterns

More information

A note on the multiplication of sparse matrices

A note on the multiplication of sparse matrices Cent. Eur. J. Cop. Sci. 41) 2014 1-11 DOI: 10.2478/s13537-014-0201-x Central European Journal of Coputer Science A note on the ultiplication of sparse atrices Research Article Keivan Borna 12, Sohrab Aboozarkhani

More information

Probability and Stochastic Processes: A Friendly Introduction for Electrical and Computer Engineers Roy D. Yates and David J.

Probability and Stochastic Processes: A Friendly Introduction for Electrical and Computer Engineers Roy D. Yates and David J. Probability and Stochastic Processes: A Friendly Introduction for Electrical and oputer Engineers Roy D. Yates and David J. Goodan Proble Solutions : Yates and Goodan,1..3 1.3.1 1.4.6 1.4.7 1.4.8 1..6

More information

Uniform Approximation and Bernstein Polynomials with Coefficients in the Unit Interval

Uniform Approximation and Bernstein Polynomials with Coefficients in the Unit Interval Unifor Approxiation and Bernstein Polynoials with Coefficients in the Unit Interval Weiang Qian and Marc D. Riedel Electrical and Coputer Engineering, University of Minnesota 200 Union St. S.E. Minneapolis,

More information

Tail estimates for norms of sums of log-concave random vectors

Tail estimates for norms of sums of log-concave random vectors Tail estiates for nors of sus of log-concave rando vectors Rados law Adaczak Rafa l Lata la Alexander E. Litvak Alain Pajor Nicole Toczak-Jaegerann Abstract We establish new tail estiates for order statistics

More information

NON-COMMUTATIVE GRÖBNER BASES FOR COMMUTATIVE ALGEBRAS

NON-COMMUTATIVE GRÖBNER BASES FOR COMMUTATIVE ALGEBRAS PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volue 126, Nuber 3, March 1998, Pages 687 691 S 0002-9939(98)04229-4 NON-COMMUTATIVE GRÖBNER BASES FOR COMMUTATIVE ALGEBRAS DAVID EISENBUD, IRENA PEEVA,

More information

Chapter 6 1-D Continuous Groups

Chapter 6 1-D Continuous Groups Chapter 6 1-D Continuous Groups Continuous groups consist of group eleents labelled by one or ore continuous variables, say a 1, a 2,, a r, where each variable has a well- defined range. This chapter explores:

More information

Math Reviews classifications (2000): Primary 54F05; Secondary 54D20, 54D65

Math Reviews classifications (2000): Primary 54F05; Secondary 54D20, 54D65 The Monotone Lindelöf Property and Separability in Ordered Spaces by H. Bennett, Texas Tech University, Lubbock, TX 79409 D. Lutzer, College of Willia and Mary, Williasburg, VA 23187-8795 M. Matveev, Irvine,

More information

Supplementary Material for Fast and Provable Algorithms for Spectrally Sparse Signal Reconstruction via Low-Rank Hankel Matrix Completion

Supplementary Material for Fast and Provable Algorithms for Spectrally Sparse Signal Reconstruction via Low-Rank Hankel Matrix Completion Suppleentary Material for Fast and Provable Algoriths for Spectrally Sparse Signal Reconstruction via Low-Ran Hanel Matrix Copletion Jian-Feng Cai Tianing Wang Ke Wei March 1, 017 Abstract We establish

More information

Computable Shell Decomposition Bounds

Computable Shell Decomposition Bounds Coputable Shell Decoposition Bounds John Langford TTI-Chicago jcl@cs.cu.edu David McAllester TTI-Chicago dac@autoreason.co Editor: Leslie Pack Kaelbling and David Cohn Abstract Haussler, Kearns, Seung

More information

Distance Optimal Target Assignment in Robotic Networks under Communication and Sensing Constraints

Distance Optimal Target Assignment in Robotic Networks under Communication and Sensing Constraints Distance Optial Target Assignent in Robotic Networks under Counication and Sensing Constraints Jingjin Yu CSAIL @ MIT/MechE @ BU Soon-Jo Chung Petros G. Voulgaris AE @ University of Illinois Supported

More information

CS Lecture 13. More Maximum Likelihood

CS Lecture 13. More Maximum Likelihood CS 6347 Lecture 13 More Maxiu Likelihood Recap Last tie: Introduction to axiu likelihood estiation MLE for Bayesian networks Optial CPTs correspond to epirical counts Today: MLE for CRFs 2 Maxiu Likelihood

More information

Note on generating all subsets of a finite set with disjoint unions

Note on generating all subsets of a finite set with disjoint unions Note on generating all subsets of a finite set with disjoint unions David Ellis e-ail: dce27@ca.ac.uk Subitted: Dec 2, 2008; Accepted: May 12, 2009; Published: May 20, 2009 Matheatics Subject Classification:

More information

Learnability of Gaussians with flexible variances

Learnability of Gaussians with flexible variances Learnability of Gaussians with flexible variances Ding-Xuan Zhou City University of Hong Kong E-ail: azhou@cityu.edu.hk Supported in part by Research Grants Council of Hong Kong Start October 20, 2007

More information

1 Rademacher Complexity Bounds

1 Rademacher Complexity Bounds COS 511: Theoretical Machine Learning Lecturer: Rob Schapire Lecture #10 Scribe: Max Goer March 07, 2013 1 Radeacher Coplexity Bounds Recall the following theore fro last lecture: Theore 1. With probability

More information

Multi-Dimensional Hegselmann-Krause Dynamics

Multi-Dimensional Hegselmann-Krause Dynamics Multi-Diensional Hegselann-Krause Dynaics A. Nedić Industrial and Enterprise Systes Engineering Dept. University of Illinois Urbana, IL 680 angelia@illinois.edu B. Touri Coordinated Science Laboratory

More information

VC Dimension and Sauer s Lemma

VC Dimension and Sauer s Lemma CMSC 35900 (Spring 2008) Learning Theory Lecture: VC Diension and Sauer s Lea Instructors: Sha Kakade and Abuj Tewari Radeacher Averages and Growth Function Theore Let F be a class of ±-valued functions

More information

1 Proof of learning bounds

1 Proof of learning bounds COS 511: Theoretical Machine Learning Lecturer: Rob Schapire Lecture #4 Scribe: Akshay Mittal February 13, 2013 1 Proof of learning bounds For intuition of the following theore, suppose there exists a

More information

Consistent Multiclass Algorithms for Complex Performance Measures. Supplementary Material

Consistent Multiclass Algorithms for Complex Performance Measures. Supplementary Material Consistent Multiclass Algoriths for Coplex Perforance Measures Suppleentary Material Notations. Let λ be the base easure over n given by the unifor rando variable (say U over n. Hence, for all easurable

More information

VARIATIONS ON THE THEME OF THE UNIFORM BOUNDARY CONDITION

VARIATIONS ON THE THEME OF THE UNIFORM BOUNDARY CONDITION VARIATIONS ON THE THEME OF THE UNIFORM BOUNDARY CONDITION DANIEL FAUSER AND CLARA LÖH ABSTRACT. The unifor boundary condition in a nored chain coplex ass for a unifor linear bound on fillings of null-hoologous

More information

DIOPHANTINE NUMBERS, DIMENSION AND DENJOY MAPS

DIOPHANTINE NUMBERS, DIMENSION AND DENJOY MAPS DIOPHANTINE NUMBERS, DIMENSION AND DENJOY MAPS BRYNA KRA AND JÖRG SCHMELING Abstract. We study the effect of the arithetic properties of the rotation nuber on the inial set of an aperiodic, orientation

More information

Some Classical Ergodic Theorems

Some Classical Ergodic Theorems Soe Classical Ergodic Theores Matt Rosenzweig Contents Classical Ergodic Theores. Mean Ergodic Theores........................................2 Maxial Ergodic Theore.....................................

More information

e-companion ONLY AVAILABLE IN ELECTRONIC FORM

e-companion ONLY AVAILABLE IN ELECTRONIC FORM OPERATIONS RESEARCH doi 10.1287/opre.1070.0427ec pp. ec1 ec5 e-copanion ONLY AVAILABLE IN ELECTRONIC FORM infors 07 INFORMS Electronic Copanion A Learning Approach for Interactive Marketing to a Custoer

More information

1 Bounding the Margin

1 Bounding the Margin COS 511: Theoretical Machine Learning Lecturer: Rob Schapire Lecture #12 Scribe: Jian Min Si March 14, 2013 1 Bounding the Margin We are continuing the proof of a bound on the generalization error of AdaBoost

More information

Handout 7. and Pr [M(x) = χ L (x) M(x) =? ] = 1.

Handout 7. and Pr [M(x) = χ L (x) M(x) =? ] = 1. Notes on Coplexity Theory Last updated: October, 2005 Jonathan Katz Handout 7 1 More on Randoized Coplexity Classes Reinder: so far we have seen RP,coRP, and BPP. We introduce two ore tie-bounded randoized

More information

The proofs of Theorem 1-3 are along the lines of Wied and Galeano (2013).

The proofs of Theorem 1-3 are along the lines of Wied and Galeano (2013). A Appendix: Proofs The proofs of Theore 1-3 are along the lines of Wied and Galeano (2013) Proof of Theore 1 Let D[d 1, d 2 ] be the space of càdlàg functions on the interval [d 1, d 2 ] equipped with

More information

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation Course Notes for EE7C (Spring 018: Convex Optiization and Approxiation Instructor: Moritz Hardt Eail: hardt+ee7c@berkeley.edu Graduate Instructor: Max Sichowitz Eail: sichow+ee7c@berkeley.edu October 15,

More information

LORENTZ SPACES AND REAL INTERPOLATION THE KEEL-TAO APPROACH

LORENTZ SPACES AND REAL INTERPOLATION THE KEEL-TAO APPROACH LORENTZ SPACES AND REAL INTERPOLATION THE KEEL-TAO APPROACH GUILLERMO REY. Introduction If an operator T is bounded on two Lebesgue spaces, the theory of coplex interpolation allows us to deduce the boundedness

More information

13.2 Fully Polynomial Randomized Approximation Scheme for Permanent of Random 0-1 Matrices

13.2 Fully Polynomial Randomized Approximation Scheme for Permanent of Random 0-1 Matrices CS71 Randoness & Coputation Spring 018 Instructor: Alistair Sinclair Lecture 13: February 7 Disclaier: These notes have not been subjected to the usual scrutiny accorded to foral publications. They ay

More information

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines

Intelligent Systems: Reasoning and Recognition. Perceptrons and Support Vector Machines Intelligent Systes: Reasoning and Recognition Jaes L. Crowley osig 1 Winter Seester 2018 Lesson 6 27 February 2018 Outline Perceptrons and Support Vector achines Notation...2 Linear odels...3 Lines, Planes

More information

Distributed Subgradient Methods for Multi-agent Optimization

Distributed Subgradient Methods for Multi-agent Optimization 1 Distributed Subgradient Methods for Multi-agent Optiization Angelia Nedić and Asuan Ozdaglar October 29, 2007 Abstract We study a distributed coputation odel for optiizing a su of convex objective functions

More information

Computable Shell Decomposition Bounds

Computable Shell Decomposition Bounds Journal of Machine Learning Research 5 (2004) 529-547 Subitted 1/03; Revised 8/03; Published 5/04 Coputable Shell Decoposition Bounds John Langford David McAllester Toyota Technology Institute at Chicago

More information

3.8 Three Types of Convergence

3.8 Three Types of Convergence 3.8 Three Types of Convergence 3.8 Three Types of Convergence 93 Suppose that we are given a sequence functions {f k } k N on a set X and another function f on X. What does it ean for f k to converge to

More information

G G G G G. Spec k G. G Spec k G G. G G m G. G Spec k. Spec k

G G G G G. Spec k G. G Spec k G G. G G m G. G Spec k. Spec k 12 VICTORIA HOSKINS 3. Algebraic group actions and quotients In this section we consider group actions on algebraic varieties and also describe what type of quotients we would like to have for such group

More information

Block designs and statistics

Block designs and statistics Bloc designs and statistics Notes for Math 447 May 3, 2011 The ain paraeters of a bloc design are nuber of varieties v, bloc size, nuber of blocs b. A design is built on a set of v eleents. Each eleent

More information

In this chapter, we consider several graph-theoretic and probabilistic models

In this chapter, we consider several graph-theoretic and probabilistic models THREE ONE GRAPH-THEORETIC AND STATISTICAL MODELS 3.1 INTRODUCTION In this chapter, we consider several graph-theoretic and probabilistic odels for a social network, which we do under different assuptions

More information

Theore A. Let n (n 4) be arcoplete inial iersed hypersurface in R n+. 4(n ) Then there exists a constant C (n) = (n 2)n S (n) > 0 such that if kak n d

Theore A. Let n (n 4) be arcoplete inial iersed hypersurface in R n+. 4(n ) Then there exists a constant C (n) = (n 2)n S (n) > 0 such that if kak n d Gap theores for inial subanifolds in R n+ Lei Ni Departent of atheatics Purdue University West Lafayette, IN 47907 99 atheatics Subject Classification 53C2 53C42 Introduction Let be a copact iersed inial

More information

A new type of lower bound for the largest eigenvalue of a symmetric matrix

A new type of lower bound for the largest eigenvalue of a symmetric matrix Linear Algebra and its Applications 47 7 9 9 www.elsevier.co/locate/laa A new type of lower bound for the largest eigenvalue of a syetric atrix Piet Van Mieghe Delft University of Technology, P.O. Box

More information

A Bernstein-Markov Theorem for Normed Spaces

A Bernstein-Markov Theorem for Normed Spaces A Bernstein-Markov Theore for Nored Spaces Lawrence A. Harris Departent of Matheatics, University of Kentucky Lexington, Kentucky 40506-0027 Abstract Let X and Y be real nored linear spaces and let φ :

More information

Metric Entropy of Convex Hulls

Metric Entropy of Convex Hulls Metric Entropy of Convex Hulls Fuchang Gao University of Idaho Abstract Let T be a precopact subset of a Hilbert space. The etric entropy of the convex hull of T is estiated in ters of the etric entropy

More information

Quantum algorithms (CO 781, Winter 2008) Prof. Andrew Childs, University of Waterloo LECTURE 15: Unstructured search and spatial search

Quantum algorithms (CO 781, Winter 2008) Prof. Andrew Childs, University of Waterloo LECTURE 15: Unstructured search and spatial search Quantu algoriths (CO 781, Winter 2008) Prof Andrew Childs, University of Waterloo LECTURE 15: Unstructured search and spatial search ow we begin to discuss applications of quantu walks to search algoriths

More information

Lecture 21. Interior Point Methods Setup and Algorithm

Lecture 21. Interior Point Methods Setup and Algorithm Lecture 21 Interior Point Methods In 1984, Kararkar introduced a new weakly polynoial tie algorith for solving LPs [Kar84a], [Kar84b]. His algorith was theoretically faster than the ellipsoid ethod and

More information

The linear sampling method and the MUSIC algorithm

The linear sampling method and the MUSIC algorithm INSTITUTE OF PHYSICS PUBLISHING INVERSE PROBLEMS Inverse Probles 17 (2001) 591 595 www.iop.org/journals/ip PII: S0266-5611(01)16989-3 The linear sapling ethod and the MUSIC algorith Margaret Cheney Departent

More information

ORIGAMI CONSTRUCTIONS OF RINGS OF INTEGERS OF IMAGINARY QUADRATIC FIELDS

ORIGAMI CONSTRUCTIONS OF RINGS OF INTEGERS OF IMAGINARY QUADRATIC FIELDS #A34 INTEGERS 17 (017) ORIGAMI CONSTRUCTIONS OF RINGS OF INTEGERS OF IMAGINARY QUADRATIC FIELDS Jürgen Kritschgau Departent of Matheatics, Iowa State University, Aes, Iowa jkritsch@iastateedu Adriana Salerno

More information

ON SEQUENCES OF NUMBERS IN GENERALIZED ARITHMETIC AND GEOMETRIC PROGRESSIONS

ON SEQUENCES OF NUMBERS IN GENERALIZED ARITHMETIC AND GEOMETRIC PROGRESSIONS Palestine Journal of Matheatics Vol 4) 05), 70 76 Palestine Polytechnic University-PPU 05 ON SEQUENCES OF NUMBERS IN GENERALIZED ARITHMETIC AND GEOMETRIC PROGRESSIONS Julius Fergy T Rabago Counicated by

More information

PERMANENT WEAK AMENABILITY OF GROUP ALGEBRAS OF FREE GROUPS

PERMANENT WEAK AMENABILITY OF GROUP ALGEBRAS OF FREE GROUPS PERMANENT WEAK AMENABILITY OF GROUP ALGEBRAS OF FREE GROUPS B. E. JOHNSON ABSTRACT We show that all derivations fro the group algebra (G) of a free group into its nth dual, where n is a positive even integer,

More information

arxiv: v1 [math.co] 19 Apr 2017

arxiv: v1 [math.co] 19 Apr 2017 PROOF OF CHAPOTON S CONJECTURE ON NEWTON POLYTOPES OF q-ehrhart POLYNOMIALS arxiv:1704.0561v1 [ath.co] 19 Apr 017 JANG SOO KIM AND U-KEUN SONG Abstract. Recently, Chapoton found a q-analog of Ehrhart polynoials,

More information

Curious Bounds for Floor Function Sums

Curious Bounds for Floor Function Sums 1 47 6 11 Journal of Integer Sequences, Vol. 1 (018), Article 18.1.8 Curious Bounds for Floor Function Sus Thotsaporn Thanatipanonda and Elaine Wong 1 Science Division Mahidol University International

More information

Fairness via priority scheduling

Fairness via priority scheduling Fairness via priority scheduling Veeraruna Kavitha, N Heachandra and Debayan Das IEOR, IIT Bobay, Mubai, 400076, India vavitha,nh,debayan}@iitbacin Abstract In the context of ulti-agent resource allocation

More information

Feature Extraction Techniques

Feature Extraction Techniques Feature Extraction Techniques Unsupervised Learning II Feature Extraction Unsupervised ethods can also be used to find features which can be useful for categorization. There are unsupervised ethods that

More information

An EGZ generalization for 5 colors

An EGZ generalization for 5 colors An EGZ generalization for 5 colors David Grynkiewicz and Andrew Schultz July 6, 00 Abstract Let g zs(, k) (g zs(, k + 1)) be the inial integer such that any coloring of the integers fro U U k 1,..., g

More information

Lower Bounds for Quantized Matrix Completion

Lower Bounds for Quantized Matrix Completion Lower Bounds for Quantized Matrix Copletion Mary Wootters and Yaniv Plan Departent of Matheatics University of Michigan Ann Arbor, MI Eail: wootters, yplan}@uich.edu Mark A. Davenport School of Elec. &

More information

An Algorithm for Quantization of Discrete Probability Distributions

An Algorithm for Quantization of Discrete Probability Distributions An Algorith for Quantization of Discrete Probability Distributions Yuriy A. Reznik Qualco Inc., San Diego, CA Eail: yreznik@ieee.org Abstract We study the proble of quantization of discrete probability

More information

1 Proving the Fundamental Theorem of Statistical Learning

1 Proving the Fundamental Theorem of Statistical Learning THEORETICAL MACHINE LEARNING COS 5 LECTURE #7 APRIL 5, 6 LECTURER: ELAD HAZAN NAME: FERMI MA ANDDANIEL SUO oving te Fundaental Teore of Statistical Learning In tis section, we prove te following: Teore.

More information

Shannon Sampling II. Connections to Learning Theory

Shannon Sampling II. Connections to Learning Theory Shannon Sapling II Connections to Learning heory Steve Sale oyota echnological Institute at Chicago 147 East 60th Street, Chicago, IL 60637, USA E-ail: sale@athberkeleyedu Ding-Xuan Zhou Departent of Matheatics,

More information

Question 1. Question 3. Question 4. Graduate Analysis I Exercise 4

Question 1. Question 3. Question 4. Graduate Analysis I Exercise 4 Graduate Analysis I Exercise 4 Question 1 If f is easurable and λ is any real nuber, f + λ and λf are easurable. Proof. Since {f > a λ} is easurable, {f + λ > a} = {f > a λ} is easurable, then f + λ is

More information

ADVANCES ON THE BESSIS- MOUSSA-VILLANI TRACE CONJECTURE

ADVANCES ON THE BESSIS- MOUSSA-VILLANI TRACE CONJECTURE ADVANCES ON THE BESSIS- MOUSSA-VILLANI TRACE CONJECTURE CHRISTOPHER J. HILLAR Abstract. A long-standing conjecture asserts that the polynoial p(t = Tr(A + tb ] has nonnegative coefficients whenever is

More information

A Low-Complexity Congestion Control and Scheduling Algorithm for Multihop Wireless Networks with Order-Optimal Per-Flow Delay

A Low-Complexity Congestion Control and Scheduling Algorithm for Multihop Wireless Networks with Order-Optimal Per-Flow Delay A Low-Coplexity Congestion Control and Scheduling Algorith for Multihop Wireless Networks with Order-Optial Per-Flow Delay Po-Kai Huang, Xiaojun Lin, and Chih-Chun Wang School of Electrical and Coputer

More information

Physics 215 Winter The Density Matrix

Physics 215 Winter The Density Matrix Physics 215 Winter 2018 The Density Matrix The quantu space of states is a Hilbert space H. Any state vector ψ H is a pure state. Since any linear cobination of eleents of H are also an eleent of H, it

More information

Bounds on the Minimax Rate for Estimating a Prior over a VC Class from Independent Learning Tasks

Bounds on the Minimax Rate for Estimating a Prior over a VC Class from Independent Learning Tasks Bounds on the Miniax Rate for Estiating a Prior over a VC Class fro Independent Learning Tasks Liu Yang Steve Hanneke Jaie Carbonell Deceber 01 CMU-ML-1-11 School of Coputer Science Carnegie Mellon University

More information

Acyclic Colorings of Directed Graphs

Acyclic Colorings of Directed Graphs Acyclic Colorings of Directed Graphs Noah Golowich Septeber 9, 014 arxiv:1409.7535v1 [ath.co] 6 Sep 014 Abstract The acyclic chroatic nuber of a directed graph D, denoted χ A (D), is the iniu positive

More information

A NEW CHARACTERIZATION OF THE CR SPHERE AND THE SHARP EIGENVALUE ESTIMATE FOR THE KOHN LAPLACIAN. 1. Introduction

A NEW CHARACTERIZATION OF THE CR SPHERE AND THE SHARP EIGENVALUE ESTIMATE FOR THE KOHN LAPLACIAN. 1. Introduction A NEW CHARACTERIZATION OF THE CR SPHERE AND THE SHARP EIGENVALUE ESTIATE FOR THE KOHN LAPLACIAN SONG-YING LI, DUONG NGOC SON, AND XIAODONG WANG 1. Introduction Let, θ be a strictly pseudoconvex pseudo-heritian

More information

Chicago, Chicago, Illinois, USA b Department of Mathematics NDSU, Fargo, North Dakota, USA

Chicago, Chicago, Illinois, USA b Department of Mathematics NDSU, Fargo, North Dakota, USA This article was downloaded by: [Sean Sather-Wagstaff] On: 25 August 2013, At: 09:06 Publisher: Taylor & Francis Infora Ltd Registered in England and Wales Registered Nuber: 1072954 Registered office:

More information

Rademacher Complexity Margin Bounds for Learning with a Large Number of Classes

Rademacher Complexity Margin Bounds for Learning with a Large Number of Classes Radeacher Coplexity Margin Bounds for Learning with a Large Nuber of Classes Vitaly Kuznetsov Courant Institute of Matheatical Sciences, 25 Mercer street, New York, NY, 002 Mehryar Mohri Courant Institute

More information

INDEPENDENT SETS IN HYPERGRAPHS

INDEPENDENT SETS IN HYPERGRAPHS INDEPENDENT SETS IN HYPERGRAPHS Abstract. Many iportant theores and conjectures in cobinatorics, such as the theore of Szeerédi on arithetic progressions and the Erdős-Stone Theore in extreal graph theory,

More information

NOGA ALON, JÓZSEF BALOGH, ROBERT MORRIS, AND WOJCIECH SAMOTIJ

NOGA ALON, JÓZSEF BALOGH, ROBERT MORRIS, AND WOJCIECH SAMOTIJ A REFINEMENT OF THE CAMERON-ERDŐS CONJECTURE NOGA ALON, JÓZSEF BALOGH, ROBERT MORRIS, AND WOJCIECH SAMOTIJ Abstract. In this paper we study su-free subsets of the set {1,..., n}, that is, subsets of the

More information

Fourier Series Summary (From Salivahanan et al, 2002)

Fourier Series Summary (From Salivahanan et al, 2002) Fourier Series Suary (Fro Salivahanan et al, ) A periodic continuous signal f(t), - < t

More information

RIGIDITY OF QUASI-EINSTEIN METRICS

RIGIDITY OF QUASI-EINSTEIN METRICS RIGIDITY OF QUASI-EINSTEIN METRICS JEFFREY CASE, YU-JEN SHU, AND GUOFANG WEI Abstract. We call a etric quasi-einstein if the -Bakry-Eery Ricci tensor is a constant ultiple of the etric tensor. This is

More information

Least Squares Fitting of Data

Least Squares Fitting of Data Least Squares Fitting of Data David Eberly, Geoetric Tools, Redond WA 98052 https://www.geoetrictools.co/ This work is licensed under the Creative Coons Attribution 4.0 International License. To view a

More information

Kernel Methods and Support Vector Machines

Kernel Methods and Support Vector Machines Intelligent Systes: Reasoning and Recognition Jaes L. Crowley ENSIAG 2 / osig 1 Second Seester 2012/2013 Lesson 20 2 ay 2013 Kernel ethods and Support Vector achines Contents Kernel Functions...2 Quadratic

More information

Lecture 21 Nov 18, 2015

Lecture 21 Nov 18, 2015 CS 388R: Randoized Algoriths Fall 05 Prof. Eric Price Lecture Nov 8, 05 Scribe: Chad Voegele, Arun Sai Overview In the last class, we defined the ters cut sparsifier and spectral sparsifier and introduced

More information

A Note on the Applied Use of MDL Approximations

A Note on the Applied Use of MDL Approximations A Note on the Applied Use of MDL Approxiations Daniel J. Navarro Departent of Psychology Ohio State University Abstract An applied proble is discussed in which two nested psychological odels of retention

More information

Descent polynomials. Mohamed Omar Department of Mathematics, Harvey Mudd College, 301 Platt Boulevard, Claremont, CA , USA,

Descent polynomials. Mohamed Omar Department of Mathematics, Harvey Mudd College, 301 Platt Boulevard, Claremont, CA , USA, Descent polynoials arxiv:1710.11033v2 [ath.co] 13 Nov 2017 Alexander Diaz-Lopez Departent of Matheatics and Statistics, Villanova University, 800 Lancaster Avenue, Villanova, PA 19085, USA, alexander.diaz-lopez@villanova.edu

More information

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation

Course Notes for EE227C (Spring 2018): Convex Optimization and Approximation Course Notes for EE227C (Spring 2018): Convex Optiization and Approxiation Instructor: Moritz Hardt Eail: hardt+ee227c@berkeley.edu Graduate Instructor: Max Sichowitz Eail: sichow+ee227c@berkeley.edu October

More information

Fixed-to-Variable Length Distribution Matching

Fixed-to-Variable Length Distribution Matching Fixed-to-Variable Length Distribution Matching Rana Ali Ajad and Georg Böcherer Institute for Counications Engineering Technische Universität München, Gerany Eail: raa2463@gail.co,georg.boecherer@tu.de

More information