A note on norms of signed sums of vectors

Size: px
Start display at page:

Download "A note on norms of signed sums of vectors"

Transcription

1 A ote o orms of siged sums of vectors Giorgos Chasapis ad Nikos Skarmogiais Abstract Our startig poit is a improved versio of a result of. Hajela related to a questio of Komlós: we show that if f) is a fuctio such that lim f) = ad f) = o), there exists 0 = 0f) such that for every 0 ad ay S { 1, 1} with cardiality S 2 /f) oe ca fid orthoormal vectors x 1,..., x R that satisfy 1x x c log f) for all 1,..., ) S. We obtai aalogous results i the case where x 1,..., x are idepedet radom poits uiformly distributed i the Euclidea uit ball B2 or ay symmetric covex body, ad the l -orm is replaced by a arbitrary orm o R. 1 Itroductio Give a pair K, of symmetric covex bodies i R, the parameter βk, ) is the smallest r > 0 such that for ay x 1,..., x K there exist sigs 1,..., { 1, 1} for which 1 x x r. A geeral lower boud for βk, ) was proved by Baaszczyk; i [2] he showed that if K ad are two symmetric covex bodies i R the 1.1) βk, ) c K / ) 1/ for a absolute costat c > 0, where K deotes the volume of K. I what follows we write B p for the uit ball of l p = R, p ), 1 p. A well-kow theorem of Specer [18] states that βb, B ) c, where c > 0 is a absolute costat the same result was proved idepedetly by Gluski i [9]): there exists a absolute costat c > 0 such that for ay 1 ad x 1,..., x R with x i 1, we may fid 1,..., { 1, 1} such that 1.2) 1 x x c. From 1.1) oe readily sees that this result is optimal up to absolute costats. A well-kow questio of Komlós see [19] ad [20]) asks if the sequece βb 2, B ) is bouded. Sice B B 2, a positive aswer to this questio immediately implies Specer s boud. The best kow estimate is due to Baaszczyk [3]: there exists a absolute costat c > 0 such that for ay 1 ad x 1,..., x R with x i 2 1 oe may fid 1,..., { 1, 1} such that 1.3) 1 x x c log. I fact, Baaszczyk proved a more geeral theorem: if K is a covex body i R with Gaussia measure γ K) 1/2 the βb 2, K) C, where C > 0 is a absolute costat; this implies 1.3) because γ rb ) 1/2 for all r c log. While the method of [3] is o-costructive, a algorithmic proof of the O log ) boud i the above statemet was recetly give by Basal, adush ad Garg [6]. The startig poit of this ote is a result of Hajela [11] i the directio of providig a egative aswer to the questio of Komlós. 1

2 Theorem 1.1 Hajela). Let f) be a fuctio such that lim f) = ad f) = o). For ay 0 < λ < 1 2 there exists 0 = 0 f, λ) such that for every 0 ad ay S { 1, 1} with cardiality S 2 /f) oe ca fid orthoormal vectors x 1,..., x R that satisfy ) λ log log f) 1 x x exp log log log f) for all 1,..., ) S. I fact, Hajela cojectures i [11] that the questio of Komlós has a egative aswer ad that the estimate 1.3) that was later obtaied by Baaszczyk should be optimal. Our first result is a improved versio of Theorem 1.1. Theorem 1.2. There exists a absolute costat c 0, 1) such that the followig holds true: For every 1 ad 1 < δ < 1, ad for ay S { 1, 1} with S 2 δ, there exist orthoormal vectors x 1,..., x i R such that i x i c log1/δ) for all 1,..., ) S. Theorem 1.2 implies a stroger versio of Hajela s theorem. Let f) be a fuctio such that lim f) = ad f) = o). Note that if we let δ = δf, ) = f) 1 i Theorem 1.2 the 1 < δ < 1 for large eough, resultig to the lower boud i x i c log f), which improves upo the estimate of Theorem 1.1. I the proof of Theorem 1.2, preseted i Sectio 2 below, iitially we follow the idea of Hajela: the vectors x 1,..., x are obtaied from a radom rotatio of the stadard basis e 1,..., e of R. The improvemet o the estimates is due to Lemma 2.1 which provides a stroger small-ball probability estimate for the -orm o the sphere. The above method to lower-boud the l -orm of a siged sum of vectors has obvious limitatios. Namely, oe should cosider a subset S { 1, 1} of cardiality 2 o) for Theorem 1.2 to obtai a lower boud of order greater tha the oe i 1.1). Thus, this lie of thikig by itself does ot seem adequate to provide a egative aswer to the questio of Komlós. Nevertheless, we fid the lik betwee small ball probability estimates ad the orm of siged sums of vectors iterestig i its ow right, so i Sectio 3 we further explore this pheomeo i several aspects. Iitially, the vectors x 1,... x are assumed to satisfy a lower boud of the form ix i 2 c for all choices of sigs i = ±1, ad the l -orm is replaced by a arbitrary orm o R. I particular, give a symmetric covex body i R, we will prove that for x 1,... x as above ad for ay big subset S { 1, 1}, the -orm of iux i is large for every 1,..., ) S, with overwhelmig probability with respect to U O). This largeess is determied by the Gaussia measure of dilates of ; to make this more precise, we give the followig defiitio. efiitio 1.3. Give a symmetric covex body i R, ad δ 0, 1), let t,δ := max{t > 0 : γ 2t m)) 2 δ e) }, where m) is the media of with respect to the stadard Gaussia measure γ o R. I the case that is the uit ball B p of some l p, p [1, ], we abbreviate t p,δ := t B p,δ. Note that, for ay symmetric covex body i R ad δ 0, 1), t,δ satisfies the bouds 1.4) c 1 1/ t,δ m) c 2 m), 2

3 for some absolute costats c 1, c 2 > 0 for completeess, we provide a short justificatio of 1.4) i Sectio 3.1). Although the defiitio of t,δ might seem somewhat artificial at first sight, we trust that the idea behid it will become clear i the sequel see the commets after Theorem 1.5). Usig the otatio itroduced above, our first result is the followig statemet. Theorem 1.4. Let be a symmetric covex body i R ad δ 0, 1). For ay τ > 0, ay -tuple of vectors x 1,..., x with mi ix i 2 τ ad ay S { 1, 1} with S 2 δ, there is some i=±1 U O) with ν U) 1 e such that for every U U, i Ux i ) τt,δ m) holds for all = 1,..., ) S. We the cosider the case where x 1,..., x are poits i a arbitrary covex body K i R. We observe that a alterative proof of 1.1) ca be deduced from a more geeral result of Gluski ad V. Milma i [10]: Let be a star body i R with 0 it) ad V 1,..., V m be measurable subsets of R with = V 1 = = V m. For ay λ 1,..., λ m R ad ay 0 < t < 1 oe has P x i ) m m, x i V i : λ i x i t m λ 2 i ) 1/2 ) ) te 1 t2 2, where the probability is with respect to the product of the uiform probability measures µ i A) = A Vi V i. The proof of this estimate i [10] is based o a rearragemet iequality of Brascamp, Lieb ad Luttiger [5]. As a ext step, usig Theorem 1.4 we obtai the followig variat of the result of Gluski ad Milma, i the case that V i = B 2 for all i. Theorem 1.5. Let δ 0, 1), be a symmetric covex body i R ad S { 1, 1} with S 2 δ. The ) P x i ) B2 : i x i 1 10 t,δm), for some S 2e. This theorem may be viewed as a extesio of Hajela s result: the l -orm is replaced by a arbitrary orm o R ad the statemet holds for a radom choice of vectors i the Euclidea uit ball. I this cotext, the role of the parameter t,δ i the statemet of our results becomes ow more trasparet: Sice, by 1.4), t,δ m) t,δ M) t,δ M)vrad) 1/, where M) = S 1 x dσx) ad vrad) = / B 2 ) 1/, ad sice M)vrad) 1 this is a simple cosequece of Hölder s iequality), we see that Theorem 1.5 provides iformatio stroger tha the oe from 1.1) provided that M)vrad) 1 ad/or t,δ 1. This is the case for the -orm ad it would be iterestig to provide further examples with sharp estimates. The fuctio t γ 2t m)) that appears i the defiitio of t,δ above has bee studied i [12], [13] ad [16] ad elsewhere. A sample of estimates is the followig: I [12] it is proved that for every 0 < t < 1 2 oe has γ {x : x t m)}) 1 2 td), where { )} m) d) = mi, log γ. 2 3

4 I [13] it is proved that if γ ) 1 2 the for every 0 < t < 1 2 oe has γ t) 2t) r2 )/4 γ ) where r) is the iradius of. I [16] it is proved that for every 0 < t < 1 2 oe has γ {x : x t m)}) 1 2 tc/β), where ad c > 0 is a absolute costat. β) = Var γ ) M 2 ) I this ote we discuss the estimates that ca be derived from the above i special istaces, as for example whe is a l p ball. Fially, the argumet i the proof of Theorem 1.5 ca be further geeralized to the case where x 1,..., x are idepedet radom poits chose uiformly from ay symmetric covex body. Theorem 1.6. There exists a absolute costat c > 0 such that the followig holds: Let δ 0, 1), be a symmetric covex body i R ad S { 1, 1} with S 2 δ. For ay symmetric covex body K i R with K = B2 we may fid U O) with ν U) > 1 2e /2 such that, for ay U U, ) P z i ) UK) UK) : i z i ct,δ m), for some S e /2. 2 A improved versio of Hajela s result I this sectio we preset the proof of Theorem 1.2. I what follows, e 1,..., e is the stadard basis of R. We deote by S 1 the Euclidea sphere i R, ad by σ the uique rotatioally ivariat probability measure o S 1. Recall that σ ca be defied via the Haar probability measure ν o the orthogoal group O) as follows: For every measurable A S 1 we have σa) = ν {U O) : Ue 1 ) A}). We eed a small ball probability estimate for the l -orm, which is give i the ext lemma. It provides a boud whose behaviour for small values of t will play a crucial role i the sequel; i the case of this type of behaviour has bee observed i various works see e.g. [21] ad [15]). We provide a direct ad short proof of the exact statemet that we eed. Lemma 2.1. There exists a absolute costat c 0, 1) such that, for ay 1 ad δ 1, 4) 1, { }) σ θ S 1 : θ c log1/δ) < 2 δ, where σ is the rotatioally ivariat probability measure o S 1. Proof. We use the fact see [12] for its simple proof) that if A is a symmetric covex body i R the σs 1 A) 2γ 2 A), to write { σ θ S 1 : θ s }) 2 2γ {x R : x s}) = Φs))) 2 exp 21 Φs))), 4

5 where, as usual, Φ stads for the cumulative distributio fuctio of the stadard ormal distributio, that is, Φx) = 2π) 1/2 x e t2 /2 dt. Usig the iequality t + s) 2 2t 2 + s 2 ), we have that, for ay s > 0, It follows that Fially, if δ 1, Φs) = 1 2π which proves the desired statemet. { σ θ S 1 : θ 0 2π e t+s)2 2 dt e s2 ) the choosig s = 1 2 log ) 1 δ we get 0 e t2 dt = e s s }) 2 2 exp ) 2 e s2. 2 exp 2 ) e s2 = 2 exp 2 ) δ 1/4 < 2 δ, Proof of Theorem 1.2. The idea of the proof is the same as i Hajela s paper. We cosider orthoormal systems obtaied as radom rotatios of the stadard basis e 1,..., e of R. For ay = 1,..., ) { 1, 1}, there is some V O) such that e i = V ie i ). Thus, by the rotatioal ivariace of the Haar measure, we have, for ay α > 0, ) }) { ν {U O) : U ) }) i e i α = ν U O) : U e i α. Let δ 1, 1/4). Sice 1/2 e i S 1, by the defiitio of the measure σ o S 1 it follows, takig α = c log1/δ) where c is the costat i Lemma 2.1, that ) ν {U O) : U i e i c }) log1/δ) = ν {U O) : U e i { = σ }) θ S 1 : θ c log1/δ). ) }) c log1/δ) Applyig Lemma 2.1 we get σ { θ S 1 : θ c log1/δ) )) < 2 δ. This shows that ) 2.1) ν {U O) : U i e i c }) log1/δ) < 2 δ. 5

6 Now let S { 1, 1} with S 2 δ. The, by the uio boud ad 2.1), ν {U O) : i Ue i ) c }) log1/δ), for every S = = 1 ν {U O) : i Ue i ) c }) log1/δ), for some S 1 ν {U O) : i Ue i ) c }) log1/δ) S > 1 S 2 δ 0. We thus deduce that there exists some U 0 O) such that, if we set x i = U 0 e i ), i = 1,...,, the i x i c log1/δ), for ay = 1,..., ) S, give that δ < 1/4. Fially, ote that i the case δ 1/4, 1) the wated result holds trivially, sice for ay U O) ad every { 1, 1}, i Ue i 1 c log 1/δ is valid, for a suitable absolute costat c > 0. 3 Siged sums of radom vectors from a covex body The argumet used for the proof of Theorem 1.2 motivates us to cosider a more geeral settig. As a first step, we let the l orm be replaced by a arbitrary orm i R ad relax our assumptios o the choice of the vectors x 1,..., x B 2. As a secod step, we will cosider a further geeralizatio, lettig x 1,..., x be chose uiformly ad idepedetly from B 2, or ay covex body K. 3.1 Norms of siged sums of vectors Let be a symmetric covex body i R. We cosider the media m) of Z where Z N0, I ) is a stadard Gaussia radom vector i R. Repeatig the argumet of the previous sectio we ca show the followig. Propositio 3.1. Let be a symmetric covex body i R ad τ > 0. Assume that the vectors x 1,..., x R satisfy ζ i x i τ 2, for all ζ i = ±1. The, for ay S { 1, 1} with S γ 2t τ m) ) < 1, we may fid U O) with ν U) > 1 S γ 2t τ m)) such that ay U U satisfies i Ux i ) t m) for all = 1,..., ) S. 6

7 Proof. Let x 1,..., x R be as i the statemet above, ad fix 1,...,. Normalizig by ix i 2 ad usig the fact that the latter is greater tha τ, we get }) ν {U O) : i Ux i ) t m) Therefore, ν {U O) : U ) ix i ix i 2 { = σ θ S 1 : θ t m) }) τ 2t ) γ τ m) = P Z 2t ) τ m). }) t m) τ }) ν {U O) : i Ux i ) t m), for every S 1 }) ν {U O) : i Ux i ) t m) S ) 2t > 1 S γ τ m). This proves the claim of the propositio. Theorem 1.4 ow is a direct corollary of Propositio 3.1. Proof of Theorem 1.4. Let t = τt,δ ad S { 1, 1} with S 2 δ. By the defiitio of t,δ we have that S γ 2t τ m)) < e. We ca thus apply Propositio 3.1 to get U O) such that ν U) 1 e ad i Ux i t m) = τt,δ m) for every U U ad all 1,..., ) S, which is the statemet of the theorem. Before we proceed, let us briefly explai at this poit the geeral bouds o the parameter t,δ, stated i the itroductio. Proof of 1.4). For the upper boud, it is straightforward by the defiitio of the media that γ m)) = γ Z m)) 1 2 > 2δ e), hece t,δ 1 2. For the lower boud, sice, itegratig i polar coordiates, x S 1 Markov s iequality implies that σ x e η B 2 ) 1/ ) e η dσx) = B 2, holds for every η > 0. We ca relate the latter to the Gaussia measure; usig the same argumet as i the proof of [12, Lemma 2.1], oe ca prove that for ay a 1, ) 1 γ e η B 1/ ) ) 2 B ) σ x e η 1/ ) ) γ B 1 a a 2 e η + γ B a 2. 7

8 We ca boud the secod term usig, e.g. [4, Propositio 2.2], to get γ 1 ) a B 2 a exp a 2 1) 2a ). 2 Note that, for η = log4e) we have 2e η = e) 2 δ e) for ay δ 0, 1), so it suffices to have a a 2 ) 1) exp 2a 2 4e) = e η, which is satisfied if a is large eough a = 20 will do). Now, sice γ every δ 0, 1), the defiitio of t,δ implies that for some absolute costat c 1 > 0. t,δ 1 160e m) 3.2 Radom poits from covex bodies 1 80e ) B 1/ 2 1 c 1 1/ m), ) ) B 1/ 2 δ e) for We will see ow that, suitably ormalized, the hypothesis o the orm of the sum ζ 1 x ζ x i the statemet of Propositio 3.1 is satisfied with overwhelmig probability whe the x i s are chose uiformly ad idepedetly at radom from the iterior of a geeral symmetric covex body K. The required lower boud, which actually holds for ay orm i R i place of the Euclidea orm, will be deduced as a corollary of the followig result of Gluski ad V. Milma [10]: Propositio 3.2 Gluski-V. Milma). Let be a star body i R with 0 it) ad V 1,..., V m be measurable subsets of R with = V 1 = = V m. For ay λ 1,..., λ m R ad ay 0 < t < 1 oe has ) P x i ) m m, x i V i : λ i x i t m ) 1/2 ) λ 2 i te 1 t2 2. Below, we will make use of the followig special istace of Propositio 3.2, which also provides a alterative proof of Baaszczyk s geeral lower boud 1.1) for βk, ). Corollary 3.3. Let K ad be symmetric covex bodies i R, ad let x 1,..., x be radom poits chose uiformly from K. The iequality i x i 1 ) 1/ K 10 is the valid for ay choice of 1,..., { 1, 1}, with probability greater tha 1 e with respect to x 1,..., x. Proof. Let t > 0 ad assume first that K =. We have ) P x i ) 1 K : i x i t, for some 1,..., { 1, 1} ) 2 P x i ) 1 K : i x i t Now if we apply Propositio 3.2 for m :=, V i := K ad λ i := 1 i for every i, ad t such that 2te 1 t2 )/2 < e 1 say t = 1/10) we get ) P x i ) 1 K : i x i t, for some 1,..., { 1, 1} 2 te 1 t2 2 ) < e. 8 2

9 We deduce that, with probability greater tha 1 e with respect to x 1,..., x ), we have i x i > 10 1 for every choice of 1,..., { 1, 1}. For a geeral pair of symmetric covex bodies K,, set a = / K ) 1/ ad K = ak. The we ca apply the above for the pair K, to deduce that for every choice of 1,..., { 1, 1}, i ax i > 1, 10 or, equivaletly, i x i > 1 10 ) 1/ K holds with probability greater tha 1 e with respect to x 1,..., x ). Remark 3.4. Note that Corollary 3.3 immediately implies Baaszczyk s lower boud o βk, ): βk, ) c K / ) 1/ Poits from the ball First we deal with the situatio where the vectors x 1,..., x are chose idepedetly ad uiformly at radom from B 2. The statemet of Theorem 1.5 follows immediately from the defiitio of t,δ ad the ext cosequece of Propositio 3.1. This ca be viewed as a geeralizatio of Hajela s result, i the spirit of Propositio 3.2. Theorem 3.5. Let be a symmetric covex body i R ad S { 1, 1}. The ) P x i ) B2 : i x i t m), for some S S γ 20t m)) + e. Proof. Let A B2 ) be the set { A := x i ) B2 : i x 2 i 1 }, for every 10 1,..., { 1, 1}. By Corollary 3.3, applied for K = = B2, we have that PA c ) < e. This fact, combied with Propositio 3.1 gives ) P z i ) B2 : i z i t m), for some S ) = P x i ), U) B2 ) O) : i Ux i ) t m), for some S ) P x i ), U) A O) : i Ux i ) t m), for some S + e { })] [ν U O) : i Ux i ) t m), for some S dµx ) dµx 1 ) + e as claimed. A < S γ 20t m)) + e, 9

10 Applicatio: the case of l p. As a applicatio let us cosider the case = Bp, 1 p. We remark that, although the estimate βb2, B2 ) seems to be well-kow a proof of this ca be foud i [7]), we could ot locate i the literature ay upper boud o the parameter βb2, Bp ), for p 2. However, for 1 p < 2, oe ca use the previously metioed estimate for βb2, B2 ): if we kow that for every x 1,..., x B2 there exists i ) { 1, 1} such that 1x x 2, the oe ca write 1 x x p 1 p x x 2 1/p. The above estimate is actually sharp, by the lower boud 1.1) ad the fact that B 2 / B p ) 1/ the order of 1 p 1 2. O the other had, for the case p > 2 oe could use the estimate of Baasczcyk βb2, B ) c log i a similar fashio, to deduce that βb2, Bp ) c 1/p log. Recall the defiitio of the parameter β i [17]: β) = Var γ ) M 2. ) For = B p, β) has bee computed i [17]: oe has βb p ) 2p p 2 if 1 p c log ad βb p ) log ) 2 if C log p, for some absolute costats c, C > 0 for a more detailed aalysis, oe may cosult [14]). Moreover, from results of Guédo ad Lata la see [8, Sectio 2.4]) it is kow that, i geeral, m) E Z for ay symmetric covex body i R, ad i particular while mb p ) E Z p 1/p p, 1 p log, mb p ) E Z p log, log p. Therefore, choosig t = 1 4 i Theorem 3.5 ad usig the estimate γ {x : x p t mb p )}) 1 2 tc/βb p ), we get: Corollary 3.6. For ay p 1 there exists a costat c p > 0 such that for ay 0 p) ad ay S { 1, 1} with S 2 cp we have that a radom -tuple of poits i B2 satisfies, with probability greater tha 1 e, i x i c p p 1/p for all = 1,..., ) S. Remark 3.7. It is useful to compare the last result with 1.1). For every p log, sice B 2 / B p ) 1/ is of the order of 1 p 1 2 we see that for ay S { 1, 1} with S 2 cp a radom -tuple of poits i B2 satisfies, with probability greater tha 1 e, i x i c p p ) B 1/ 2 Bp for all = 1,..., ) S. O the other had, i the case p > log oe ca use the fact that the orms p ad are equivalet to deduce from Theorem 1.2 that for ay 0 < ϱ < 1 ad S { 1, 1} with S 2 1 ϱ, a radom -tuple of poits i B2 satisfies, with probability greater tha 1 e, i x i cϱ) p log ) B 1/ 2 Bp for all = 1,..., ) S. I fact, the results of this sectio poit out a geeral way to derive variats of 1.1) i this spirit. A meaigful lower boud o ix i ca be obtaied with high probability for the radom -tuple x 1,..., x, if oe is willig to drop the requiremet that this holds for ay = 1,..., ) { 1, 1}, but rather restrict themselves to a big subset S { 1, 1}. is of 10

11 3.2.2 Poits from a symmetric covex body Fially, we study the case where x 1,..., x are chose uiformly ad idepedetly from a arbitrary symmetric covex body K i R. We shall prove the followig geeralizatio of Theorem 3.5, which i tur, for t := ct,δ will give the claim i Theorem 1.6. Theorem 3.8. There exists a absolute costat c > 0 such the followig holds: Let be a symmetric covex body i R ad S { 1, 1}. For ay symmetric covex body K i R let t > 0 be such that S γ ct B 2 / K ) 1/ m)) < e. The, we may fid U O) with ν U) > 1 2e /2 such that, for all U U, ) P z i ) UK) UK) : i z i t m), for some S e /2. Proof. Let { A = x i ) K : i x i > ) 1/ K, for every B2 1,..., { 1, 1}}. By Corollary 3.3, applied for = B2, we have that PA c ) < e. Usig Propositio 3.1 we write ) P z i ) UK) : O) i z i t m), for some S ) = P x i ), U) K O) : i Ux i t m), for some S ) P x i ), U) A O) : i Ux i t m), for some S + e { })] [ν U O) : i Ux i t m), for some S dµx ) dµx 1 ) + e A < S γ 20t B 2 / K ) 1/ m)) + e. Assume that t is chose so that S γ 20t B 2 / K ) 1/ m)) < e. Applyig Markov s iequality we may fid U O) with ν U) > 1 2e /2 such that ) P z i ) UK) : i z i t m), for some S e /2. This proves the theorem. Remark 3.9. Theorem 3.8 illustrates oce more the mai idea behid the improvemet upo Hajela s result, Theorem 1.2, as well as the rest of the results i this ote; small ball probability estimates for the Gaussia measure of covex bodies ca be liked to the deductio of lower bouds for variats of the quatity βk, ), where the -orm of the siged sum of a radom -tuple of vectors ca be lower bouded for every choice of sigs i ay appropriately large subset of { 1, 1}. 11

12 Ackowledgemets. We would like to thak Apostolos Giaopoulos for useful discussios. The research of the first amed author is supported by the Natioal Scholarship Foudatio IKY), sposored by the act Scholarship grats for secod-degree graduate studies, from resources of the operatioal program Mapower evelopmet, Educatio ad Life-log Learig, , co-fuded by the Europea Social Fud ESF) ad the Greek state. The secod amed author is supported by a Ph Scholarship from the Helleic Foudatio for Research ad Iovatio ELIEK); research umber 70/3/ Refereces [1] S. Artstei-Avida, A. Giaopoulos ad V.. Milma, Asymptotic Geometric Aalysis, Part I, Amer. Math. Soc., Mathematical Surveys ad Moographs ). [2] W. Baaszczyk, Balacig vectors ad covex bodies, Studia Math ), [3] W. Baaszczyk, Balacig vectors ad Gaussia measures of -dimesioal covex bodies, Radom Structures Algorithms ), o. 4, [4] A. Barviok, Measure Cocetratio, lecture otes, Uiversity of Michiga 2005), available at math.lsa.umich.edu/~barviok/total710.pdf. [5] H. J. Brascamp, E. H. Lieb ad J. M. Luttiger, A geeral rearragemet iequality for multiple itegrals, J. Fuct. Aal ), [6] N. Basal,. adush, ad S. Garg, A algorithm for Komlós cojecture matchig Baaszczyks boud, I 57th Aual IEEE Symposium o Foudatios of Computer Sciece, FOCS 2016, New Bruswick, NJ, USA, October , [7] I. Báráy, O the power of liear depedecies, Buildig Bridges, Bolyai Society Mathematical Studies, Vol. 19, Spriger, Berli 2008), pp [8] S. Brazitikos, A. Giaopoulos, P. Valettas ad B-H. Vritsiou, Geometry of isotropic covex bodies, Amer. Math. Society, Mathematical Surveys ad Moographs ). [9] E.. Gluski, Extremal properties of orthogoal parallelepipeds ad their applicatios to the geometry of Baach spaces, Math. USSR Sborik ), [10] E. Gluski ad V. Milma, Geometric probability ad radom cotype 2, Geometric Aspects of Fuctioal Aalysis, volume 1850 of Lecture Notes i Math., pages Spriger, Berli, [11]. Hajela, O a Cojecture of Komlós about Siged Sums of Vectors iside the Sphere, Eur. J. Comb. 91): 33 37, 1988). [12] B. Klartag ad R. Vershyi, Small ball probability ad voretzky Theorem, Israel J. Math ), o. 1, [13] R. Lata la ad K. Oleszkiewicz, Small ball probability estimates i terms of widths, Studia Math ), [14] A. Lytova ad K. E. Tikhomirov, The variace of the l p -orm of the gaussia vector, ad voretzkys theorem, preprit 2017). [15] V.. Milma ad G. Schechtma, A isomorphic versio of voretzky s theorem, C.R. Acad. Sci. Paris ), [16] G. Paouris ad P. Valettas, A Gaussia small deviatio iequality for covex fuctios, Aals of Probability ), [17] G. Paouris, P. Valettas ad J. Zi, Radom versio of voretzky s theorem i l p, Stochastic Process. Appl ), o. 10, [18] J. Specer, Six stadard deviatios suffice, Tras. Amer. Math. Soc ), [19] J. Specer, Balacig vectors i the max orm, Combiatorica 61) 1986), [20] J. Specer, Te Lectures o the Probabilistic Method, Society for Idustrial ad Applied Mathematics, Philadelphia, Pe [21] A. Szakowski, O voretzky s theorem o almost spherical sectios of covex bodies, Israel J. Math ),

13 Keywords: covex bodies, balacig vectors, siged sums, Komlós cojecture, radom rotatios, gaussia measure, small ball probability MSC: Primary 52A23; Secodary 46B06, 52A40, Giorgos Chasapis: epartmet of Mathematics, Uiversity of Athes, Paepistimioupolis , Athes, Greece. Nikos Skarmogiais: epartmet of Mathematics, Uiversity of Athes, Paepistimioupolis , Athes, Greece. 13

The random version of Dvoretzky s theorem in l n

The random version of Dvoretzky s theorem in l n The radom versio of Dvoretzky s theorem i l Gideo Schechtma Abstract We show that with high probability a sectio of the l ball of dimesio k cε log c > 0 a uiversal costat) is ε close to a multiple of the

More information

Self-normalized deviation inequalities with application to t-statistic

Self-normalized deviation inequalities with application to t-statistic Self-ormalized deviatio iequalities with applicatio to t-statistic Xiequa Fa Ceter for Applied Mathematics, Tiaji Uiversity, 30007 Tiaji, Chia Abstract Let ξ i i 1 be a sequece of idepedet ad symmetric

More information

A remark on p-summing norms of operators

A remark on p-summing norms of operators A remark o p-summig orms of operators Artem Zvavitch Abstract. I this paper we improve a result of W. B. Johso ad G. Schechtma by provig that the p-summig orm of ay operator with -dimesioal domai ca be

More information

Minimal surface area position of a convex body is not always an M-position

Minimal surface area position of a convex body is not always an M-position Miimal surface area positio of a covex body is ot always a M-positio Christos Saroglou Abstract Milma proved that there exists a absolute costat C > 0 such that, for every covex body i R there exists a

More information

An Introduction to Randomized Algorithms

An Introduction to Randomized Algorithms A Itroductio to Radomized Algorithms The focus of this lecture is to study a radomized algorithm for quick sort, aalyze it usig probabilistic recurrece relatios, ad also provide more geeral tools for aalysis

More information

arxiv: v1 [math.pr] 13 Oct 2011

arxiv: v1 [math.pr] 13 Oct 2011 A tail iequality for quadratic forms of subgaussia radom vectors Daiel Hsu, Sham M. Kakade,, ad Tog Zhag 3 arxiv:0.84v math.pr] 3 Oct 0 Microsoft Research New Eglad Departmet of Statistics, Wharto School,

More information

Several properties of new ellipsoids

Several properties of new ellipsoids Appl. Math. Mech. -Egl. Ed. 008 9(7):967 973 DOI 10.1007/s10483-008-0716-y c Shaghai Uiversity ad Spriger-Verlag 008 Applied Mathematics ad Mechaics (Eglish Editio) Several properties of ew ellipsoids

More information

Rademacher Complexity

Rademacher Complexity EECS 598: Statistical Learig Theory, Witer 204 Topic 0 Rademacher Complexity Lecturer: Clayto Scott Scribe: Ya Deg, Kevi Moo Disclaimer: These otes have ot bee subjected to the usual scrutiy reserved for

More information

Convergence of random variables. (telegram style notes) P.J.C. Spreij

Convergence of random variables. (telegram style notes) P.J.C. Spreij Covergece of radom variables (telegram style otes).j.c. Spreij this versio: September 6, 2005 Itroductio As we kow, radom variables are by defiitio measurable fuctios o some uderlyig measurable space

More information

Boundaries and the James theorem

Boundaries and the James theorem Boudaries ad the James theorem L. Vesely 1. Itroductio The followig theorem is importat ad well kow. All spaces cosidered here are real ormed or Baach spaces. Give a ormed space X, we deote by B X ad S

More information

An Extremal Property of the Regular Simplex

An Extremal Property of the Regular Simplex Covex Geometric Aalysis MSRI Publicatios Volume 34, 1998 A Extremal Property of the Regular Simplex MICHAEL SCHMUCKENSCHLÄGER Abstract. If C is a covex body i R such that the ellipsoid of miimal volume

More information

OFF-DIAGONAL MULTILINEAR INTERPOLATION BETWEEN ADJOINT OPERATORS

OFF-DIAGONAL MULTILINEAR INTERPOLATION BETWEEN ADJOINT OPERATORS OFF-DIAGONAL MULTILINEAR INTERPOLATION BETWEEN ADJOINT OPERATORS LOUKAS GRAFAKOS AND RICHARD G. LYNCH 2 Abstract. We exted a theorem by Grafakos ad Tao [5] o multiliear iterpolatio betwee adjoit operators

More information

On Random Line Segments in the Unit Square

On Random Line Segments in the Unit Square O Radom Lie Segmets i the Uit Square Thomas A. Courtade Departmet of Electrical Egieerig Uiversity of Califoria Los Ageles, Califoria 90095 Email: tacourta@ee.ucla.edu I. INTRODUCTION Let Q = [0, 1] [0,

More information

Lecture 2. The Lovász Local Lemma

Lecture 2. The Lovász Local Lemma Staford Uiversity Sprig 208 Math 233A: No-costructive methods i combiatorics Istructor: Ja Vodrák Lecture date: Jauary 0, 208 Origial scribe: Apoorva Khare Lecture 2. The Lovász Local Lemma 2. Itroductio

More information

Dimension-free PAC-Bayesian bounds for the estimation of the mean of a random vector

Dimension-free PAC-Bayesian bounds for the estimation of the mean of a random vector Dimesio-free PAC-Bayesia bouds for the estimatio of the mea of a radom vector Olivier Catoi CREST CNRS UMR 9194 Uiversité Paris Saclay olivier.catoi@esae.fr Ilaria Giulii Laboratoire de Probabilités et

More information

Product measures, Tonelli s and Fubini s theorems For use in MAT3400/4400, autumn 2014 Nadia S. Larsen. Version of 13 October 2014.

Product measures, Tonelli s and Fubini s theorems For use in MAT3400/4400, autumn 2014 Nadia S. Larsen. Version of 13 October 2014. Product measures, Toelli s ad Fubii s theorems For use i MAT3400/4400, autum 2014 Nadia S. Larse Versio of 13 October 2014. 1. Costructio of the product measure The purpose of these otes is to preset the

More information

AN EXTENSION OF SIMONS INEQUALITY AND APPLICATIONS. Robert DEVILLE and Catherine FINET

AN EXTENSION OF SIMONS INEQUALITY AND APPLICATIONS. Robert DEVILLE and Catherine FINET 2001 vol. XIV, um. 1, 95-104 ISSN 1139-1138 AN EXTENSION OF SIMONS INEQUALITY AND APPLICATIONS Robert DEVILLE ad Catherie FINET Abstract This article is devoted to a extesio of Simos iequality. As a cosequece,

More information

Lecture 27. Capacity of additive Gaussian noise channel and the sphere packing bound

Lecture 27. Capacity of additive Gaussian noise channel and the sphere packing bound Lecture 7 Ageda for the lecture Gaussia chael with average power costraits Capacity of additive Gaussia oise chael ad the sphere packig boud 7. Additive Gaussia oise chael Up to this poit, we have bee

More information

Asymptotic distribution of products of sums of independent random variables

Asymptotic distribution of products of sums of independent random variables Proc. Idia Acad. Sci. Math. Sci. Vol. 3, No., May 03, pp. 83 9. c Idia Academy of Scieces Asymptotic distributio of products of sums of idepedet radom variables YANLING WANG, SUXIA YAO ad HONGXIA DU ollege

More information

Optimally Sparse SVMs

Optimally Sparse SVMs A. Proof of Lemma 3. We here prove a lower boud o the umber of support vectors to achieve geeralizatio bouds of the form which we cosider. Importatly, this result holds ot oly for liear classifiers, but

More information

The log-behavior of n p(n) and n p(n)/n

The log-behavior of n p(n) and n p(n)/n Ramauja J. 44 017, 81-99 The log-behavior of p ad p/ William Y.C. Che 1 ad Ke Y. Zheg 1 Ceter for Applied Mathematics Tiaji Uiversity Tiaji 0007, P. R. Chia Ceter for Combiatorics, LPMC Nakai Uivercity

More information

Definition 4.2. (a) A sequence {x n } in a Banach space X is a basis for X if. unique scalars a n (x) such that x = n. a n (x) x n. (4.

Definition 4.2. (a) A sequence {x n } in a Banach space X is a basis for X if. unique scalars a n (x) such that x = n. a n (x) x n. (4. 4. BASES I BAACH SPACES 39 4. BASES I BAACH SPACES Sice a Baach space X is a vector space, it must possess a Hamel, or vector space, basis, i.e., a subset {x γ } γ Γ whose fiite liear spa is all of X ad

More information

An almost sure invariance principle for trimmed sums of random vectors

An almost sure invariance principle for trimmed sums of random vectors Proc. Idia Acad. Sci. Math. Sci. Vol. 20, No. 5, November 200, pp. 6 68. Idia Academy of Scieces A almost sure ivariace priciple for trimmed sums of radom vectors KE-ANG FU School of Statistics ad Mathematics,

More information

Sequences and Series of Functions

Sequences and Series of Functions Chapter 6 Sequeces ad Series of Fuctios 6.1. Covergece of a Sequece of Fuctios Poitwise Covergece. Defiitio 6.1. Let, for each N, fuctio f : A R be defied. If, for each x A, the sequece (f (x)) coverges

More information

If a subset E of R contains no open interval, is it of zero measure? For instance, is the set of irrationals in [0, 1] is of measure zero?

If a subset E of R contains no open interval, is it of zero measure? For instance, is the set of irrationals in [0, 1] is of measure zero? 2 Lebesgue Measure I Chapter 1 we defied the cocept of a set of measure zero, ad we have observed that every coutable set is of measure zero. Here are some atural questios: If a subset E of R cotais a

More information

Chapter 7 Isoperimetric problem

Chapter 7 Isoperimetric problem Chapter 7 Isoperimetric problem Recall that the isoperimetric problem (see the itroductio its coectio with ido s proble) is oe of the most classical problem of a shape optimizatio. It ca be formulated

More information

A REMARK ON A PROBLEM OF KLEE

A REMARK ON A PROBLEM OF KLEE C O L L O Q U I U M M A T H E M A T I C U M VOL. 71 1996 NO. 1 A REMARK ON A PROBLEM OF KLEE BY N. J. K A L T O N (COLUMBIA, MISSOURI) AND N. T. P E C K (URBANA, ILLINOIS) This paper treats a property

More information

2 Banach spaces and Hilbert spaces

2 Banach spaces and Hilbert spaces 2 Baach spaces ad Hilbert spaces Tryig to do aalysis i the ratioal umbers is difficult for example cosider the set {x Q : x 2 2}. This set is o-empty ad bouded above but does ot have a least upper boud

More information

Chapter 3. Strong convergence. 3.1 Definition of almost sure convergence

Chapter 3. Strong convergence. 3.1 Definition of almost sure convergence Chapter 3 Strog covergece As poited out i the Chapter 2, there are multiple ways to defie the otio of covergece of a sequece of radom variables. That chapter defied covergece i probability, covergece i

More information

7.1 Convergence of sequences of random variables

7.1 Convergence of sequences of random variables Chapter 7 Limit Theorems Throughout this sectio we will assume a probability space (, F, P), i which is defied a ifiite sequece of radom variables (X ) ad a radom variable X. The fact that for every ifiite

More information

This section is optional.

This section is optional. 4 Momet Geeratig Fuctios* This sectio is optioal. The momet geeratig fuctio g : R R of a radom variable X is defied as g(t) = E[e tx ]. Propositio 1. We have g () (0) = E[X ] for = 1, 2,... Proof. Therefore

More information

Lecture 19. sup y 1,..., yn B d n

Lecture 19. sup y 1,..., yn B d n STAT 06A: Polyomials of adom Variables Lecture date: Nov Lecture 19 Grothedieck s Iequality Scribe: Be Hough The scribes are based o a guest lecture by ya O Doell. I this lecture we prove Grothedieck s

More information

A HYPERPLANE INEQUALITY FOR MEASURES OF CONVEX BODIES IN R n, n 4

A HYPERPLANE INEQUALITY FOR MEASURES OF CONVEX BODIES IN R n, n 4 A HYPERPANE INEQUAITY FOR MEASURES OF CONVEX BODIES IN R, 4 AEXANDER ODOBSY Abstract. et 4. We show that for a arbitrary measure µ with eve cotiuous desity i R ad ay origi-symmetric covex body i R, µ()

More information

ON SOME INEQUALITIES IN NORMED LINEAR SPACES

ON SOME INEQUALITIES IN NORMED LINEAR SPACES ON SOME INEQUALITIES IN NORMED LINEAR SPACES S.S. DRAGOMIR Abstract. Upper ad lower bouds for the orm of a liear combiatio of vectors are give. Applicatios i obtaiig various iequalities for the quatities

More information

ON THE CENTRAL LIMIT PROPERTY OF CONVEX BODIES

ON THE CENTRAL LIMIT PROPERTY OF CONVEX BODIES ON THE CENTRAL LIMIT PROPERTY OF CONVEX BODIES S. G. Bobkov ad A. oldobsky February 1, 00 Abstract For isotropic covex bodies i R with isotropic costat L, we study the rate of covergece, as goes to ifiity,

More information

7.1 Convergence of sequences of random variables

7.1 Convergence of sequences of random variables Chapter 7 Limit theorems Throughout this sectio we will assume a probability space (Ω, F, P), i which is defied a ifiite sequece of radom variables (X ) ad a radom variable X. The fact that for every ifiite

More information

BIRKHOFF ERGODIC THEOREM

BIRKHOFF ERGODIC THEOREM BIRKHOFF ERGODIC THEOREM Abstract. We will give a proof of the poitwise ergodic theorem, which was first proved by Birkhoff. May improvemets have bee made sice Birkhoff s orgial proof. The versio we give

More information

ACO Comprehensive Exam 9 October 2007 Student code A. 1. Graph Theory

ACO Comprehensive Exam 9 October 2007 Student code A. 1. Graph Theory 1. Graph Theory Prove that there exist o simple plaar triagulatio T ad two distict adjacet vertices x, y V (T ) such that x ad y are the oly vertices of T of odd degree. Do ot use the Four-Color Theorem.

More information

Disjoint Systems. Abstract

Disjoint Systems. Abstract Disjoit Systems Noga Alo ad Bey Sudaov Departmet of Mathematics Raymod ad Beverly Sacler Faculty of Exact Scieces Tel Aviv Uiversity, Tel Aviv, Israel Abstract A disjoit system of type (,,, ) is a collectio

More information

Notes 19 : Martingale CLT

Notes 19 : Martingale CLT Notes 9 : Martigale CLT Math 733-734: Theory of Probability Lecturer: Sebastie Roch Refereces: [Bil95, Chapter 35], [Roc, Chapter 3]. Sice we have ot ecoutered weak covergece i some time, we first recall

More information

Lecture 19: Convergence

Lecture 19: Convergence Lecture 19: Covergece Asymptotic approach I statistical aalysis or iferece, a key to the success of fidig a good procedure is beig able to fid some momets ad/or distributios of various statistics. I may

More information

On equivalent strictly G-convex renormings of Banach spaces

On equivalent strictly G-convex renormings of Banach spaces Cet. Eur. J. Math. 8(5) 200 87-877 DOI: 0.2478/s533-00-0050-3 Cetral Europea Joural of Mathematics O equivalet strictly G-covex reormigs of Baach spaces Research Article Nataliia V. Boyko Departmet of

More information

MDIV. Multiple divisor functions

MDIV. Multiple divisor functions MDIV. Multiple divisor fuctios The fuctios τ k For k, defie τ k ( to be the umber of (ordered factorisatios of ito k factors, i other words, the umber of ordered k-tuples (j, j 2,..., j k with j j 2...

More information

Riesz-Fischer Sequences and Lower Frame Bounds

Riesz-Fischer Sequences and Lower Frame Bounds Zeitschrift für Aalysis ud ihre Aweduge Joural for Aalysis ad its Applicatios Volume 1 (00), No., 305 314 Riesz-Fischer Sequeces ad Lower Frame Bouds P. Casazza, O. Christese, S. Li ad A. Lider Abstract.

More information

ECE 330:541, Stochastic Signals and Systems Lecture Notes on Limit Theorems from Probability Fall 2002

ECE 330:541, Stochastic Signals and Systems Lecture Notes on Limit Theorems from Probability Fall 2002 ECE 330:541, Stochastic Sigals ad Systems Lecture Notes o Limit Theorems from robability Fall 00 I practice, there are two ways we ca costruct a ew sequece of radom variables from a old sequece of radom

More information

Law of the sum of Bernoulli random variables

Law of the sum of Bernoulli random variables Law of the sum of Beroulli radom variables Nicolas Chevallier Uiversité de Haute Alsace, 4, rue des frères Lumière 68093 Mulhouse icolas.chevallier@uha.fr December 006 Abstract Let be the set of all possible

More information

STAT Homework 1 - Solutions

STAT Homework 1 - Solutions STAT-36700 Homework 1 - Solutios Fall 018 September 11, 018 This cotais solutios for Homework 1. Please ote that we have icluded several additioal commets ad approaches to the problems to give you better

More information

Math Solutions to homework 6

Math Solutions to homework 6 Math 175 - Solutios to homework 6 Cédric De Groote November 16, 2017 Problem 1 (8.11 i the book): Let K be a compact Hermitia operator o a Hilbert space H ad let the kerel of K be {0}. Show that there

More information

Lecture 2 Long paths in random graphs

Lecture 2 Long paths in random graphs Lecture Log paths i radom graphs 1 Itroductio I this lecture we treat the appearace of log paths ad cycles i sparse radom graphs. will wor with the probability space G(, p) of biomial radom graphs, aalogous

More information

Seunghee Ye Ma 8: Week 5 Oct 28

Seunghee Ye Ma 8: Week 5 Oct 28 Week 5 Summary I Sectio, we go over the Mea Value Theorem ad its applicatios. I Sectio 2, we will recap what we have covered so far this term. Topics Page Mea Value Theorem. Applicatios of the Mea Value

More information

Lecture 7: October 18, 2017

Lecture 7: October 18, 2017 Iformatio ad Codig Theory Autum 207 Lecturer: Madhur Tulsiai Lecture 7: October 8, 207 Biary hypothesis testig I this lecture, we apply the tools developed i the past few lectures to uderstad the problem

More information

Solution. 1 Solutions of Homework 1. Sangchul Lee. October 27, Problem 1.1

Solution. 1 Solutions of Homework 1. Sangchul Lee. October 27, Problem 1.1 Solutio Sagchul Lee October 7, 017 1 Solutios of Homework 1 Problem 1.1 Let Ω,F,P) be a probability space. Show that if {A : N} F such that A := lim A exists, the PA) = lim PA ). Proof. Usig the cotiuity

More information

Distribution of Random Samples & Limit theorems

Distribution of Random Samples & Limit theorems STAT/MATH 395 A - PROBABILITY II UW Witer Quarter 2017 Néhémy Lim Distributio of Radom Samples & Limit theorems 1 Distributio of i.i.d. Samples Motivatig example. Assume that the goal of a study is to

More information

Gamma Distribution and Gamma Approximation

Gamma Distribution and Gamma Approximation Gamma Distributio ad Gamma Approimatio Xiaomig Zeg a Fuhua (Frak Cheg b a Xiame Uiversity, Xiame 365, Chia mzeg@jigia.mu.edu.c b Uiversity of Ketucky, Leigto, Ketucky 456-46, USA cheg@cs.uky.edu Abstract

More information

Appendix to Quicksort Asymptotics

Appendix to Quicksort Asymptotics Appedix to Quicksort Asymptotics James Alle Fill Departmet of Mathematical Scieces The Johs Hopkis Uiversity jimfill@jhu.edu ad http://www.mts.jhu.edu/~fill/ ad Svate Jaso Departmet of Mathematics Uppsala

More information

A Note on Sums of Independent Random Variables

A Note on Sums of Independent Random Variables Cotemorary Mathematics Volume 00 XXXX A Note o Sums of Ideedet Radom Variables Pawe l Hitczeko ad Stehe Motgomery-Smith Abstract I this ote a two sided boud o the tail robability of sums of ideedet ad

More information

Learning Theory: Lecture Notes

Learning Theory: Lecture Notes Learig Theory: Lecture Notes Kamalika Chaudhuri October 4, 0 Cocetratio of Averages Cocetratio of measure is very useful i showig bouds o the errors of machie-learig algorithms. We will begi with a basic

More information

5.1 Review of Singular Value Decomposition (SVD)

5.1 Review of Singular Value Decomposition (SVD) MGMT 69000: Topics i High-dimesioal Data Aalysis Falll 06 Lecture 5: Spectral Clusterig: Overview (cotd) ad Aalysis Lecturer: Jiamig Xu Scribe: Adarsh Barik, Taotao He, September 3, 06 Outlie Review of

More information

Topic 9: Sampling Distributions of Estimators

Topic 9: Sampling Distributions of Estimators Topic 9: Samplig Distributios of Estimators Course 003, 2018 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be

More information

Approximation by Superpositions of a Sigmoidal Function

Approximation by Superpositions of a Sigmoidal Function Zeitschrift für Aalysis ud ihre Aweduge Joural for Aalysis ad its Applicatios Volume 22 (2003, No. 2, 463 470 Approximatio by Superpositios of a Sigmoidal Fuctio G. Lewicki ad G. Mario Abstract. We geeralize

More information

Lecture 3: August 31

Lecture 3: August 31 36-705: Itermediate Statistics Fall 018 Lecturer: Siva Balakrisha Lecture 3: August 31 This lecture will be mostly a summary of other useful expoetial tail bouds We will ot prove ay of these i lecture,

More information

Some remarks for codes and lattices over imaginary quadratic

Some remarks for codes and lattices over imaginary quadratic Some remarks for codes ad lattices over imagiary quadratic fields Toy Shaska Oaklad Uiversity, Rochester, MI, USA. Caleb Shor Wester New Eglad Uiversity, Sprigfield, MA, USA. shaska@oaklad.edu Abstract

More information

A Note on the Kolmogorov-Feller Weak Law of Large Numbers

A Note on the Kolmogorov-Feller Weak Law of Large Numbers Joural of Mathematical Research with Applicatios Mar., 015, Vol. 35, No., pp. 3 8 DOI:10.3770/j.iss:095-651.015.0.013 Http://jmre.dlut.edu.c A Note o the Kolmogorov-Feller Weak Law of Large Numbers Yachu

More information

Notes for Lecture 11

Notes for Lecture 11 U.C. Berkeley CS78: Computatioal Complexity Hadout N Professor Luca Trevisa 3/4/008 Notes for Lecture Eigevalues, Expasio, ad Radom Walks As usual by ow, let G = (V, E) be a udirected d-regular graph with

More information

Let us give one more example of MLE. Example 3. The uniform distribution U[0, θ] on the interval [0, θ] has p.d.f.

Let us give one more example of MLE. Example 3. The uniform distribution U[0, θ] on the interval [0, θ] has p.d.f. Lecture 5 Let us give oe more example of MLE. Example 3. The uiform distributio U[0, ] o the iterval [0, ] has p.d.f. { 1 f(x =, 0 x, 0, otherwise The likelihood fuctio ϕ( = f(x i = 1 I(X 1,..., X [0,

More information

MAT1026 Calculus II Basic Convergence Tests for Series

MAT1026 Calculus II Basic Convergence Tests for Series MAT026 Calculus II Basic Covergece Tests for Series Egi MERMUT 202.03.08 Dokuz Eylül Uiversity Faculty of Sciece Departmet of Mathematics İzmir/TURKEY Cotets Mootoe Covergece Theorem 2 2 Series of Real

More information

(VII.A) Review of Orthogonality

(VII.A) Review of Orthogonality VII.A Review of Orthogoality At the begiig of our study of liear trasformatios i we briefly discussed projectios, rotatios ad projectios. I III.A, projectios were treated i the abstract ad without regard

More information

Supplementary Material for Fast Stochastic AUC Maximization with O(1/n)-Convergence Rate

Supplementary Material for Fast Stochastic AUC Maximization with O(1/n)-Convergence Rate Supplemetary Material for Fast Stochastic AUC Maximizatio with O/-Covergece Rate Migrui Liu Xiaoxua Zhag Zaiyi Che Xiaoyu Wag 3 iabao Yag echical Lemmas ized versio of Hoeffdig s iequality, ote that We

More information

Frequentist Inference

Frequentist Inference Frequetist Iferece The topics of the ext three sectios are useful applicatios of the Cetral Limit Theorem. Without kowig aythig about the uderlyig distributio of a sequece of radom variables {X i }, for

More information

Empirical Process Theory and Oracle Inequalities

Empirical Process Theory and Oracle Inequalities Stat 928: Statistical Learig Theory Lecture: 10 Empirical Process Theory ad Oracle Iequalities Istructor: Sham Kakade 1 Risk vs Risk See Lecture 0 for a discussio o termiology. 2 The Uio Boud / Boferoi

More information

Remarks on the Geometry of Coordinate Projections in R n

Remarks on the Geometry of Coordinate Projections in R n Remarks o the Geometry of Coordiate Projectios i R S. Medelso R. Vershyi Abstract We study geometric properties of coordiate projectios. Amog other results, we show that if a body K R has a almost extremal

More information

Local Approximation Properties for certain King type Operators

Local Approximation Properties for certain King type Operators Filomat 27:1 (2013, 173 181 DOI 102298/FIL1301173O Published by Faculty of Scieces ad athematics, Uiversity of Niš, Serbia Available at: http://wwwpmfiacrs/filomat Local Approimatio Properties for certai

More information

BETWEEN QUASICONVEX AND CONVEX SET-VALUED MAPPINGS. 1. Introduction. Throughout the paper we denote by X a linear space and by Y a topological linear

BETWEEN QUASICONVEX AND CONVEX SET-VALUED MAPPINGS. 1. Introduction. Throughout the paper we denote by X a linear space and by Y a topological linear BETWEEN QUASICONVEX AND CONVEX SET-VALUED MAPPINGS Abstract. The aim of this paper is to give sufficiet coditios for a quasicovex setvalued mappig to be covex. I particular, we recover several kow characterizatios

More information

SHARP INEQUALITIES INVOLVING THE CONSTANT e AND THE SEQUENCE (1 + 1/n) n

SHARP INEQUALITIES INVOLVING THE CONSTANT e AND THE SEQUENCE (1 + 1/n) n SHARP INEQUALITIES INVOLVING THE CONSTANT e AND THE SEQUENCE + / NECDET BATIR Abstract. Several ew ad sharp iequalities ivolvig the costat e ad the sequece + / are proved.. INTRODUCTION The costat e or

More information

CANTOR SETS WHICH ARE MINIMAL FOR QUASISYMMETRIC MAPS

CANTOR SETS WHICH ARE MINIMAL FOR QUASISYMMETRIC MAPS CANTOR SETS WHICH ARE MINIMAL FOR QUASISYMMETRIC MAPS HRANT A. HAKOBYAN Abstract. We show that middle iterval Cator sets of Hausdorff dimesio are miimal for quasisymmetric maps of a lie. Combiig this with

More information

Equivalent Banach Operator Ideal Norms 1

Equivalent Banach Operator Ideal Norms 1 It. Joural of Math. Aalysis, Vol. 6, 2012, o. 1, 19-27 Equivalet Baach Operator Ideal Norms 1 Musudi Sammy Chuka Uiversity College P.O. Box 109-60400, Keya sammusudi@yahoo.com Shem Aywa Maside Muliro Uiversity

More information

Berry-Esseen bounds for self-normalized martingales

Berry-Esseen bounds for self-normalized martingales Berry-Essee bouds for self-ormalized martigales Xiequa Fa a, Qi-Ma Shao b a Ceter for Applied Mathematics, Tiaji Uiversity, Tiaji 30007, Chia b Departmet of Statistics, The Chiese Uiversity of Hog Kog,

More information

THE ASYMPTOTIC COMPLEXITY OF MATRIX REDUCTION OVER FINITE FIELDS

THE ASYMPTOTIC COMPLEXITY OF MATRIX REDUCTION OVER FINITE FIELDS THE ASYMPTOTIC COMPLEXITY OF MATRIX REDUCTION OVER FINITE FIELDS DEMETRES CHRISTOFIDES Abstract. Cosider a ivertible matrix over some field. The Gauss-Jorda elimiatio reduces this matrix to the idetity

More information

Rational Bounds for the Logarithm Function with Applications

Rational Bounds for the Logarithm Function with Applications Ratioal Bouds for the Logarithm Fuctio with Applicatios Robert Bosch Abstract We fid ratioal bouds for the logarithm fuctio ad we show applicatios to problem-solvig. Itroductio Let a = + solvig the problem

More information

Topic 9: Sampling Distributions of Estimators

Topic 9: Sampling Distributions of Estimators Topic 9: Samplig Distributios of Estimators Course 003, 2018 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be

More information

Moment-entropy inequalities for a random vector

Moment-entropy inequalities for a random vector 1 Momet-etropy iequalities for a radom vector Erwi Lutwak, Deae ag, ad Gaoyog Zhag Abstract The p-th momet matrix is defied for a real radom vector, geeralizig the classical covariace matrix. Sharp iequalities

More information

A NOTE ON INVARIANT SETS OF ITERATED FUNCTION SYSTEMS

A NOTE ON INVARIANT SETS OF ITERATED FUNCTION SYSTEMS Acta Math. Hugar., 2007 DOI: 10.1007/s10474-007-7013-6 A NOTE ON INVARIANT SETS OF ITERATED FUNCTION SYSTEMS L. L. STACHÓ ad L. I. SZABÓ Bolyai Istitute, Uiversity of Szeged, Aradi vértaúk tere 1, H-6720

More information

Geometry of random sections of isotropic convex bodies

Geometry of random sections of isotropic convex bodies Geometry of radom sectios of isotropic covex bodies Apostolos Giaopoulos, Labrii Hioi ad Atois Tsolomitis Abstract Let K be a isotropic symmetric covex body i R. We show that a subspace F G, k codimesio

More information

A constructive analysis of convex-valued demand correspondence for weakly uniformly rotund and monotonic preference

A constructive analysis of convex-valued demand correspondence for weakly uniformly rotund and monotonic preference MPRA Muich Persoal RePEc Archive A costructive aalysis of covex-valued demad correspodece for weakly uiformly rotud ad mootoic preferece Yasuhito Taaka ad Atsuhiro Satoh. May 04 Olie at http://mpra.ub.ui-mueche.de/55889/

More information

Concentration on the l n p ball

Concentration on the l n p ball Cocetratio o the l p ball Gideo chechtma Joel Zi Abstract We prove a cocetratio iequality for fuctios, Lipschitz with respect to the Euclidea metric, o the ball of l p,1 p

More information

The value of Banach limits on a certain sequence of all rational numbers in the interval (0,1) Bao Qi Feng

The value of Banach limits on a certain sequence of all rational numbers in the interval (0,1) Bao Qi Feng The value of Baach limits o a certai sequece of all ratioal umbers i the iterval 0, Bao Qi Feg Departmet of Mathematical Scieces, Ket State Uiversity, Tuscarawas, 330 Uiversity Dr. NE, New Philadelphia,

More information

SOME GENERALIZATIONS OF OLIVIER S THEOREM

SOME GENERALIZATIONS OF OLIVIER S THEOREM SOME GENERALIZATIONS OF OLIVIER S THEOREM Alai Faisat, Sait-Étiee, Georges Grekos, Sait-Étiee, Ladislav Mišík Ostrava (Received Jauary 27, 2006) Abstract. Let a be a coverget series of positive real umbers.

More information

On forward improvement iteration for stopping problems

On forward improvement iteration for stopping problems O forward improvemet iteratio for stoppig problems Mathematical Istitute, Uiversity of Kiel, Ludewig-Mey-Str. 4, D-24098 Kiel, Germay irle@math.ui-iel.de Albrecht Irle Abstract. We cosider the optimal

More information

Mi-Hwa Ko and Tae-Sung Kim

Mi-Hwa Ko and Tae-Sung Kim J. Korea Math. Soc. 42 2005), No. 5, pp. 949 957 ALMOST SURE CONVERGENCE FOR WEIGHTED SUMS OF NEGATIVELY ORTHANT DEPENDENT RANDOM VARIABLES Mi-Hwa Ko ad Tae-Sug Kim Abstract. For weighted sum of a sequece

More information

Available online at J. Math. Comput. Sci. 2 (2012), No. 3, ISSN:

Available online at   J. Math. Comput. Sci. 2 (2012), No. 3, ISSN: Available olie at http://scik.org J. Math. Comput. Sci. 2 (202, No. 3, 656-672 ISSN: 927-5307 ON PARAMETER DEPENDENT REFINEMENT OF DISCRETE JENSEN S INEQUALITY FOR OPERATOR CONVEX FUNCTIONS L. HORVÁTH,

More information

On determinants and the volume of random polytopes in isotropic convex bodies

On determinants and the volume of random polytopes in isotropic convex bodies O determiats ad the volume of radom polytopes i isotropic covex bodies Accepted for publicatio i Geom. Dedicata Peter Pivovarov September 29, 200 Abstract I this paper we cosider radom polytopes geerated

More information

A Hadamard-type lower bound for symmetric diagonally dominant positive matrices

A Hadamard-type lower bound for symmetric diagonally dominant positive matrices A Hadamard-type lower boud for symmetric diagoally domiat positive matrices Christopher J. Hillar, Adre Wibisoo Uiversity of Califoria, Berkeley Jauary 7, 205 Abstract We prove a ew lower-boud form of

More information

Linear Regression Demystified

Linear Regression Demystified Liear Regressio Demystified Liear regressio is a importat subject i statistics. I elemetary statistics courses, formulae related to liear regressio are ofte stated without derivatio. This ote iteds to

More information

Fastest mixing Markov chain on a path

Fastest mixing Markov chain on a path Fastest mixig Markov chai o a path Stephe Boyd Persi Diacois Ju Su Li Xiao Revised July 2004 Abstract We ider the problem of assigig trasitio probabilities to the edges of a path, so the resultig Markov

More information

Chapter 2 The Monte Carlo Method

Chapter 2 The Monte Carlo Method Chapter 2 The Mote Carlo Method The Mote Carlo Method stads for a broad class of computatioal algorithms that rely o radom sampligs. It is ofte used i physical ad mathematical problems ad is most useful

More information

Probabilistic and Average Linear Widths in L -Norm with Respect to r-fold Wiener Measure

Probabilistic and Average Linear Widths in L -Norm with Respect to r-fold Wiener Measure joural of approximatio theory 84, 3140 (1996) Article No. 0003 Probabilistic ad Average Liear Widths i L -Norm with Respect to r-fold Wieer Measure V. E. Maiorov Departmet of Mathematics, Techio, Haifa,

More information

Econ 325/327 Notes on Sample Mean, Sample Proportion, Central Limit Theorem, Chi-square Distribution, Student s t distribution 1.

Econ 325/327 Notes on Sample Mean, Sample Proportion, Central Limit Theorem, Chi-square Distribution, Student s t distribution 1. Eco 325/327 Notes o Sample Mea, Sample Proportio, Cetral Limit Theorem, Chi-square Distributio, Studet s t distributio 1 Sample Mea By Hiro Kasahara We cosider a radom sample from a populatio. Defiitio

More information

A Note On The Exponential Of A Matrix Whose Elements Are All 1

A Note On The Exponential Of A Matrix Whose Elements Are All 1 Applied Mathematics E-Notes, 8(208), 92-99 c ISSN 607-250 Available free at mirror sites of http://wwwmaththuedutw/ ame/ A Note O The Expoetial Of A Matrix Whose Elemets Are All Reza Farhadia Received

More information

TR/46 OCTOBER THE ZEROS OF PARTIAL SUMS OF A MACLAURIN EXPANSION A. TALBOT

TR/46 OCTOBER THE ZEROS OF PARTIAL SUMS OF A MACLAURIN EXPANSION A. TALBOT TR/46 OCTOBER 974 THE ZEROS OF PARTIAL SUMS OF A MACLAURIN EXPANSION by A. TALBOT .. Itroductio. A problem i approximatio theory o which I have recetly worked [] required for its solutio a proof that the

More information

A Proof of Birkhoff s Ergodic Theorem

A Proof of Birkhoff s Ergodic Theorem A Proof of Birkhoff s Ergodic Theorem Joseph Hora September 2, 205 Itroductio I Fall 203, I was learig the basics of ergodic theory, ad I came across this theorem. Oe of my supervisors, Athoy Quas, showed

More information