Correction of Samplable Additive Errors

Size: px
Start display at page:

Download "Correction of Samplable Additive Errors"

Transcription

1 Correctio of Samplable Additive Errors Keji Yasuaga Kaazawa Uiversity Kakuma-machi, Kaazawa, , Japa Abstract We study the correctability of efficietly samplable errors. Specifically, we cosider samplable additive-error chaels, where ubouded-weight errors are sampled by a polyomial-time algorithm, ad added to the chael iput i a oblivious way. Assumig the existece of oe-way fuctios, there are samplable distributios Z over {0, 1} with etropy ϵ for 0 < ϵ < 1 that are ot correctable by efficiet codig schemes. Next, we show that there is a oracle relative to which there is a samplable Z with etropy ω(log ) that is ot correctable by efficiet sydrome decodig. For flat distributios Z with etropy m, we show that if Z forms a liear subspace, there is a liear code that corrects Z with rate R 1 m/. For geeral flat distributios Z, there is a liear code that corrects Z with error ϵ for rate R 1 (m + O(log(1/ϵ)))/, ad o codig scheme ca correct Z with error ϵ for rate R > 1 (m+log(1 ϵ))/. Fially, we observe that small-biased distributios are ot correctable by high-rate codes, ad hece there is a small-biased Z with etropy m that is ot correctable for rate R > 1 m/+(2 log +O(1))/. To derive these results, we use relatios betwee error-correctig codes ad other otios such as data compressio ad radomess codesers. I. INTRODUCTION I the theory of error-correctig codes, two of the wellstudied chael models are probabilistic chaels ad worstcase chaels. I probabilistic chaels, errors are itroduced through stochastic processes, ad the most well-kow oe is the biary symmetric chael. I worst-case (or adversarial) chaels, we cosider the worst-case error for a give errorcorrectig code ad a chael iput uder the restrictio o the weight of the error vector. If we view the itroductio of errors as computatio of the chael, probabilistic chaels perform low-cost computatio with little kowledge about the code ad the iput, while worst-case chaels perform highcost computatio with the full-kowledge. As itermediate chaels betwee these two chaels, Lipto [13] cosidered computatioally-bouded chaels, where the computatio of chaels is bouded by polyomial time. Guruswami ad Smith [8] preseted explicit optimal-rate codig schemes for several computatioally-bouded chaels, icludig additive-error chaels ad time/space bouded chaels. For a survey of previous work o itermediate chaels, see Sectio 2 of [9]. I this work, we also focus o computatioally-bouded chaels. I particular, we cosider additive-error chaels, i which errors are geerated idepedetly (or obliviously) of the code ad the chael iput, ad itroduced by addig to the iput. The biary symmetric chael is a example of additive-error chaels. We cosider a computatioallybouded aalogue of additive-error chaels, called samplable additive-error chaels. I these chaels, a error vector is sampled by efficiet computatio ad added to the codeword i a oblivious way. Namely, the samplig algorithm a its radom cois caot deped o the choice of the code or the particular codeword. This is stroger tha the stadard otio of obliviousess, where a oblivious chael ca deped o the code, but ot the codeword (cf. [12]). Furthermore, we do ot boud the (Hammig) weights of the error vectors. Although most of the existig work cosiders bouded-weight errors, this restrictio may ot be ecessary for modelig errors geerated by ature as a result of polyomial-time computatio. We cosider the settig i which codig schemes ca be desiged with the kowledge o the chael. More precisely, the code ca deped o the samplig algorithm of the samplable additive-error chael, but ot o its radom cois. This settig is icomparable to previous otios of error correctio agaist computatioally-bouded chaels. Our model is stroger because we do ot limit the umber of errors itroduced by the chael, but is weaker because the error caot deped o the code or the codeword. Our Cotributios We would like to characterize samplable additive-error chaels regardig the existece of efficiet reliable codig schemes. We use the Shao etropy of samplable distributios as a criterio. The reaso is that, if the etropy is zero, it is easy to achieve reliable commuicatio sice the error is a fixed strig ad this iformatio ca be used for desigig the codig scheme. O the other had, if the samplable distributio has the full etropy, we could ot achieve reliable commuicatio sice the truly radom error is added to the trasmitted codeword. Thus, there seems to be bouds o the existece of efficiet reliable codig schemes depedig o the etropy of the uderlyig samplable distributio. Whe reliable codig schemes exist, a importat quatity of the scheme is the iformatio rate (or simply rate), which is the ratio of the message legth to the codeword legth. We ivestigate the bouds o the rate whe reliable commuicatio is achievable. a) Our Results: Let Z be a distributio over {0, 1} associated with a samplable additive-error chael, ad H(Z) the Shao etropy of Z. First, we observe that Z ca simulate biary symmetric chaels. Therefore, it follows from the coverse of Shao s

2 TABLE I CORRECTABILITY OF SAMPLABLE ADDITIVE-ERROR Z H(Z) Correctabilities Assumptios Refereces 0 Efficietly correctable No Trivial 1 determiistic code, Z ot correctable by the code No Propositio 2 O(log ) code of R > Ω( log ) with efficiet sydrome decodig, Z ot correctable by the code Oracle access Theorem 3 ω(log ) Z ot correctable by efficiet sydrome decodig for R > ω( log ) Oracle access Corollary 1 ϵ for 0 < ϵ < 1 Z ot efficietly correctable Oe-way fuctio Theorem 1 H 2 (p) for 0 < p < 1 Z ot correctable for R > 1 H 2 (p) No Capacity of BSC 0 m liear subspace Z of dimesio m is correctable for R 1 m. No Theorem 4 flat distributio Z of mi-etropy m is 0 m (1) correctable with error ϵ for R 1 m 4 log(1/ϵ) No Theorem 5 (2) ot correctable with error ϵ for R > 1 m + log(1/(1 ϵ)) No Theorem 6 δ-biased distributio is ot correctable for rate R > 1 2 log(1/δ)+1 No Theorem 7 0 m small-biased Z ot correctable for R > 1 m 2 log +O(1) + No Corollary 2 Not correctable No Trivial theorem that there exists Z with H(Z) = H 2 (p) for 0 < p < 1 that is ot correctable by codes with rate R > 1 H 2 (p), where H 2 (p) is the biary etropy fuctio defied as H 2 (p) = p log 2 p (1 p) log 2 (1 p). Regardig efficiet codig schemes, we observe that if the additive error is pseudoradom (i the cryptographic sese), the o efficiet codig scheme ca correct the errors. This implies that assumig the existece of oe-way fuctios, there exist Z with H(Z) = ϵ for 0 < ϵ < 1 that are ot efficietly correctable. Next, we focus o liear codig schemes, where the code forms a liear space. We show that there is a oracle relative to which there exists Z with H(Z) = ω(log ) that is ot correctable by liear codig schemes that employ efficiet sydrome decodig. This result also implies that there is o black-box reductio from sydrome decodig algorithms for correctig Z to samplig algorithms for Z. For this result, we use the relatio betwee liear codes correctig additive errors ad liear data compressio. After that, we cosider errors from flat distributios Z. Whe Z forms a liear subspace, we preset a efficiet codig scheme that corrects Z by sydrome decodig with rate R 1 m/, where H(Z) = m. For geeral flat distributios Z, we show that Z are correctable with error ϵ for rate R 1 m/ 4 log(1/ϵ)/. This result is derived by usig the relatio betwee a liear code esemble ad a liear lossless codesers, established by Cheraghchi [3]. Coversely, we ca show that ay flat distributio Z is ot correctable with error ϵ for rate R > 1 m/+log(1/(1 ϵ))/. We also observe that o determiistic code ca correct the family of flat distributios with the same etropy. Specifically, we show that for ay determiistic codig scheme, there is a flat distributio Z with H(Z) = 1 that is ot corrected by the scheme. Fially, we cosider errors from small-biased distributios. We show that δ-biased distributios are ot correctable for rate R > 1 (2 log(1/δ) + 1)/, ad that there is a small-biased distributio Z with H(Z) = m that is ot correctable with error ϵ for rate R > 1 m/ + (2 log + O(1))/. To derive these results, we use the fact that small-biased distributios ca be used for keys of the oe-time pad for high-etropy messages, which is derived by Dodis ad Smith [6]. The results are summarized i Table I. b) Related Work: The otio of computatioallybouded chael was itroduced by Lipto [13]. He showed that if the seder ad the receiver ca share secret radomess, the the Shao capacity ca be achieved for this chael. Micali et al. [14] cosidered a similar chael model i a public-key settig, ad gave a codig scheme based o list-decodable codes ad digital sigature. Guruswami ad Smith [8] gave costructios of capacity achievig codes for worst-case additive-error chael ad time/space-bouded chaels. I their settig of additive-error chael, the weights of errors are bouded, ad the errors are oly idepedet of the ecoder s radom cois. They also gave strog impossibility results for bit-fixig chaels, but their results ca be applied to chaels that use the iformatio o the code ad the trasmitted codewords. I this work, we give impossibility results eve for chaels that do o use such iformatio. Samplable distributios were also studied i the cotext of data compressio [7], [16], [18], radomess extractor [15], [17], [4], ad radomess codeser [5]. II. PRELIMINARIES For a distributio X, we write x X to idicate that x is chose accordig to X. We may use X also as a radom variable distributed accordig to X. The support of X is Supp(X) = {x : Pr X (x) 0}, where Pr X (x) is the probability that X assigs to x. The Shao etropy of X is H(X) = E x X [ log Pr X (x)]. The mi-etropy of X is give by mi x Supp(X) { log Pr X (x)}. For two distributios X ad Y defied o the same fiite space S, the statistical distace betwee X ad Y is give by SD(X, Y ) = 1 2 s S Pr X(s) Pr Y (s). We say X ad Y is ϵ-close if SD(X, Y ) ϵ. A flat distributio is a distributio that is uiform over its support. For N, we write U as the uiform distributio over {0, 1}.

3 We defie the otio of additive-error correctig codes. Defiitio 1: (Additive-error correctig codes) For two fuctios Ec : F k F ad Dec : F F k, ad a distributio Z over F, where F is a fiite field, we say (Ec, Dec) corrects (additive error) Z with error ϵ if for ay x F k, we have that Pr z Z [Dec(Ec(x) + z) x] ϵ. The rate of (Ec, Dec) is k/. Defiitio 2: A distributio Z is said to be correctable with rate R ad error ϵ if there is a pair of fuctios (Ec, Dec) of rate R that corrects Z with error ϵ. We call a pair (Ec, Dec) a codig scheme or simply code. The codig scheme is called efficiet if Ec ad Dec ca be computed i polyomial-time i. The code is called liear if Ec is a liear mappig, that is, for ay x, y F ad a, b F, Ec(ax + by) = aec(x) + bec(y). If F = 2, we may use {0, 1} istead of F. Next, we defie sydrome decodig for liear codes. Defiitio 3: For a liear code (Ec, Dec), Dec is said to be a sydrome decoder if there is a fuctio Rec such that Dec(y) = (y Rec(y H T )) G 1, where G F R satisfies that Ec(x) = x G for x F R, ad H F R is a dual matrix for G (i.e., GH T = 0). We cosider a computatioally-bouded aalogue of additive-error chaels. We itroduce the otio of samplable distributios. Defiitio 4: A distributio family Z = {Z } N is said to be samplable if there is a probabilistic polyomial-time algorithm S such that S(1 ) is distributed accordig to Z for every N. We cosider the settig i which codig schemes ca deped o the samplig algorithm of Z, but ot o its radom cois, ad Z does ot use ay iformatio o the codig scheme or trasmitted codewords. I this settig, the radomizatio of codig schemes does ot help much. Propositio 1: Let (Ec, Dec) be a radomized codig scheme that corrects a distributio Z with error ϵ. The, there is a determiistic codig scheme that corrects Z with error ϵ. Proof: Assume that Ec uses at most l-bit radomess. Sice (Ec, Dec) corrects Z with error ϵ, we have that for every x F k, Pr z Z,r Ul [Dec(Ec(x; r) + z) x] ϵ. By the averagig argumet, for every x F k, there exists r x {0, 1} l such that Pr z Z [Dec(Ec(x; r x )) x] ϵ. Thus, by defiig Ec (x) = Ec(x; r x ), the determiistic codig scheme (Ec, Dec) corrects Z with error ϵ. The fact that the radomizatio does ot help much is cotrast to the settig of Guruswami ad Smith [8], where the chaels ca use the iformatio o the codig scheme ad trasmitted codewords, but ot the radom cois for ecodig. They preset a radomized codig scheme with optimal rate 1 H 2 (p) for worst-case additive-error chaels, for which determiistic codig schemes are oly kow to achieve rate 1 H 2 (2p), where p is the error rate of the chaels. Next, we defie the otio of data compressio. Defiitio 5: For two fuctios Com : F F ad Decom : F F, ad a distributio Z, we say (Com, Decom) compresses Z to legth m if 1) For ay z Supp(Z), Decom(Com(z)) = z, ad 2) E[ Com(Z) ] m. Defiitio 6: We say a distributio Z is compressible to legth m, if there are two fuctios Com ad Decom such that (Com, Decom) compresses Z to legth m. If Com is a liear mappig, (Com, Decom) is called a liear compressio. Fially, we defie the otio of lossless codesers. Defiitio 7: A fuctio f : F {0, 1} d F r is said to be a (m, ϵ)-lossless codeser if for ay distributio X of mi-etropy m, the distributio (f(x, Y ), Y ) is ϵ-close to a distributio (Z, U d ) with mi-etropy at least m + d, where Y is the uiform distributio over {0, 1} d. A codeser f is liear if for ay fixed z {0, 1} d, ay x, y F ad a, b F, f(ax + by, z) = af(x, z) + bf(y, z). III. CORRECTABILITY OF ADDITIVE ERRORS A. Errors from Pseudoradom Distributios We show that o efficiet codig scheme ca correct pseudoradom distributios. Theorem 1: Assume that a oe-way fuctio exists. The, for ay 0 < ϵ < 1, there is a samplable distributio Z over {0, 1} such that H(Z) ϵ ad o polyomial-time algorithms (Ec, Dec) ca correct Z. Proof: If a oe-way fuctio exists, there is a pseudoradom geerator G : {0, 1} ϵ {0, 1} secure for ay polyomial-time algorithm [10]. The, a distributio Z = G(U ϵ) is ot correctable by polyomial-time algorithms. If so, we ca costruct a polyomial-time distiguisher for pseudoradom geerator, ad the cotradictio. B. Ucorrectable Errors from Low-Etropy Distributios We cosider errors from low-etropy distributios that are ot correctable by efficiet codig schemes. We use the relatio betwee error correctio by liear code ad data compressio by liear compressio. The relatio was explicitly preseted by Caire et al. [2]. Theorem 2 ([2]): For ay distributio Z over F, Z is correctable with rate R by sydrome decodig if ad oly if Z is compressible by liear compressio to legth (1 R). Wee [18] showed that there is a oracle relative to which there is a samplable distributio over {0, 1} of etropy O(log ) that caot be compressed to legth less tha Ω(log ) by ay efficiet compressio. Lemma 1 ([18]): For ay k satisfyig 6 log s+o(1) < k <, there are a fuctio f : {0, 1} k {0, 1} ad a oracle O f such that give oracle access to O f, 1) f(u k ) is samplable, ad has the Shao etropy k. 2) f(u k ) caot be compressed to legth less tha 2 log s log O(1) by oracle circuits of size s. By combiig Lemma 1 ad Theorem 2, we obtai the followig theorem. Theorem 3: For ay k satisfyig 6 log s + 2 log + O(1) < k <, there are a fuctio f : {0, 1} k {0, 1} ad a oracle O f such that give oracle access to O f,

4 1) f(u k ) is samplable, ad has the Shao etropy k. 2) f(u k ) is ot correctable with rate R > (2 log s 3 log )/ O(1/) by ay liear code (Ec, Dec) implemeted by a oracle circuit of size s, where Dec is a sydrome decoder. Proof: Item 1 is the same as Lemma 1. We prove Item 2 i the rest. For cotradictio, assume that there is a oracle circuit of size s such that the circuit implemets a liear code (Ec, Dec) i which Dec is a sydrome decoder, ad (Ec, Dec) corrects f(u k ) with rate R > (2 log s 3 log )/ + O(1/). By Theorem 2, we ca costruct (Com, Decom) that ca compress f(u k ) to legth (1 R) < 2 log s 3 log O(1), ad is implemeted by a oracle circuit of size s + 2, where the additio term of 2 is due to the computatio of Com, which is defied as Com(z) = z H T. This cotradicts Lemma 1. The followig corollary immediately follows. Corollary 1: For ay k satisfyig ω(log ) < k <, there is a oracle relative to which there is a samplable distributio Z with H(Z) = k that is ot correctable by liear codes of rate R > ω((log )/) with efficiet sydrome decodig. C. Errors from Liear Subspaces Let Z = {z 1, z 2,..., z l } F be a set of liearly idepedet vectors. We ca costruct a liear code that corrects additive errors from the liear spa of Z. Theorem 4: There is a liear code of rate 1 l/ that corrects the liear spa of Z by sydrome decodig. Proof: Cosider l vectors w l+1,..., w F such that the set {z 1, z 2,..., z l, w l+1,..., w } forms a basis of F. The, there is a liear trasformatio T : F F l such that T (z i ) = e i ad T (w i ) = 0, where e i is the vector with 1 i the i-th positio ad 0 elsewhere. Let H be the matrix i F l such that xh T = T (x), ad cosider a code with parity check matrix H. Let z = l i=1 a iz i be a vector i the liear spa of Z, where a i F. Sice z H T = ( l i=1 a iz i ) H T = l i=1 a ie i = (a 1,..., a l ), the code ca correct the error z by sydrome decodig. Sice H F l is the parity check matrix, the rate of the code is ( l)/. D. Errors from Flat Distributios Cheraghchi [3] showed a relatio betwee lossless codesers ad liear codes correctig additive errors. He gave the equivalece betwee a liear lossless codeser for a flat distributio Z ad a liear code esemble i which most of them correct additive errors from Z. Based o his result, we ca show that, for ay flat distributio, there is a liear code that corrects errors from the distributio. Theorem 5: For ay ϵ > 0 ad flat distributio Z over {0, 1} with mi-etropy m, there is a liear code of rate 1 m/ 4 log(1/ϵ)/ that corrects Z with error ϵ by sydrome decodig. Proof: Let f : {0, 1} {0, 1} d {0, 1} r be a liear (m, ϵ)-lossless codeser. Defie a code esemble {C u } u {0,1} d such that C u is a liear code for which a parity check matrix H u satisfies that for each x {0, 1}, x H T u = f(x, u). Cheraghchi [3] proved the followig lemma. Lemma 2 (Lemma 15 of [3]): For ay flat distributio Z with mi-etropy m, at least a (1 2 ϵ) fractio of the choices of u {0, 1} d, the code C u corrects Z with error ϵ. We use a liear lossless codeser that ca be costructed from a uiversal hash family cosistig of liear fuctios. This is a geeralizatio of the Leftover Hash Lemma ad the proof is give i [3]. Lemma 3 (Lemma 7 of [3]): For every iteger r, m, ad ϵ > 0 with r m + 2 log(1/ϵ), there is a explicit (m, ϵ) liear lossless codeser f : {0, 1} {0, 1} {0, 1} r. The statemet of the theorem immediately follows by combiig Lemmas 2 ad 3. Coversely, we ca show that the rate achieved i Theorem 5 is almost optimal. Theorem 6: Let Z be ay flat distributio over {0, 1} with mi-etropy m. If a code of rate R corrects Z with error ϵ, the R 1 m/ + log(1/(1 ϵ))/. Proof: Let (Ec, Dec) be a code that corrects Z with error ϵ. For x {0, 1} R, defie D x = {y {0, 1} : Dec(y) = x}. That is, D x is the set of iputs that are decoded to x by Dec. Sice the code corrects the flat distributio Z with error ϵ, D x (1 ϵ)2 m for every x {0, 1} R. Sice each D x is disjoit, x {0,1} R D x 2. Therefore, we have that (1 ϵ)2 m 2 R 2, which implies the statemet. By Lemma 2, oe may hope to costruct a sigle code that corrects errors from ay flat distributio with the same etropy, as costructed i [3] for the case of biary symmetric chaels by usig Justese s costructio [11]. However, it is difficult to achieve sice we ca show that it is impossible by determiistic codig schemes, ad, as preseted i Propositio 1, the radomizatio of codig schemes does ot help i our settig. Propositio 2: For ay determiistic code, there is a flat distributio of mi-etropy 1 that is ot corrected by the code with error ϵ < 1/2. Proof: Defie a flat distributio to be a uiform distributio over two differet codewords c 1 ad c 2. If the iput to the decoder is c 1 + c 2, the decoder caot distiguish the two cases where the trasmitted codewords are c 1 ad c 2. Thus, the decoder outputs the wrog aswer with probability at least 1/2 for at least oe of the two cases. E. Errors from Small-Biased Distributios A sample space S {0, 1} is said to be δ-biased if for ay o-zero α {0, 1}, E s US [( 1) α s ] δ, where U S is the uiform distributio over S. Dodis ad Smith [6] proved that small-biased distributios ca be used as sources of keys of the oe-time pad for high-etropy messages. This result implies that high-rate codes caot correct errors from small-biased distributios. Theorem 7: Let S be a δ-biased sample space over {0, 1}. If a code of rate R corrects U S with error ϵ < 1/2, the R 1 (2 log(1/δ) + 1)/.

5 Proof: Assume for cotradictio that (Ec, Dec) corrects Z = U S with rate R ad error ϵ. Dodis ad Smith [6] give the oe-time pad lemma for high-etropy messages. Lemma 4: For a δ-biased sample space S, there is a distributio G such that for every distributio M over {0, 1} with mi-etropy at least t, SD(M U S, G) γ for γ = δ2 ( t 2)/2, where is the bit-wise exclusive-or. For b {0, 1}, let M b {0, 1} R be the uiform distributio over the set of strigs i which the first bit is b. Note that the mi-etropy of M b is R 1. Defie D b = Dec(Ec(M b ) U S ). By Lemma 4, SD(D 0, D 1 ) SD(Ec(M 0 ) U S, Ec(M 1 ) U S ) SD(Ec(M 0 ) U S, G) + SD(G, Ec(M 1 ) U S ) 2γ for γ = δ2 ( R 1)/2. Sice the code corrects U S with error ϵ, we have that SD(D 0, D 1 ) 1 2ϵ. Thus, we have that 2γ > 1 2ϵ > 0, ad hece R 1 (2 log(1/δ) + 1)/. Alo et al. [1] give a costructio of δ-biased sample spaces of size O( 2 /δ 2 ). This leads to the followig corollary. Corollary 2: There is a small-biased distributio of mietropy at most m that is ot corrected by codes with rate R > 1 m/ + (2 log + O(1))/ ad error ϵ < 1/2. IV. FUTURE WORK We preset some possible future work of this study. a) Further study o the correctability: I Theorem 3, we have show the impossibility result with some restrictios: allowig some oracle access ad the decodig is cofied to a powerful sydrome decodig. Thus, provig the impossibility results without these restrictios is possible future work. I this work, we have mostly discussed impossibility results. Thus, showig o-trivial possibility results is iterestig, i particular, for the case that the umber of errors itroduced by the chael is ubouded. To study the correctability of other samplable additive chaels, such as log-space chaels ad chaels computed by costat-depth circuits, may be iterestig future work. b) Geeralizig the results of Cheraghchi [3]: I Sectio III-D, we have used the results of Cheraghchi [3] who showed that a liear lossless codeser for a flat distributio Z is equivalet to a liear code correctig additive error Z. Oe possible future work is to geeralize this result to more geeral distributios tha flat distributios. Note that the complexity of decodig is ot cosidered i the above equivalece. Namely, lossless codesers do ot eed to recover the iput from the codesed output, ad thus the decoder for the correspodig liear codes may ot be doe efficietly. However, as preseted by Cheraghchi [3] for biary-symmetric chaels, it is possible to costruct a efficiet codig scheme usig iefficiet decoders based o Justese s cocateated costructio [11]. It may be iterestig to explore other distributios (or characterize distributios) that ca be efficietly correctable by Justese s costructio. c) Characterizig correctability: We have ivestigated the correctability of samplable additive errors usig the Shao etropy as a criterio. There may be aother better criterio for characterizig the correctability of these errors, which might be related to efficiet computability, to which samplability is directly related. ACKNOWLEDGMENT This research was supported i part by JSPS Grat-i-Aid for Scietific Research Numbers , , ad We thak aoymous reviewers for their helpful commets. REFERENCES [1] N. Alo, O. Goldreich, J. Håstad, ad R. Peralta. Simple costructio of almost k-wise idepedet radom variables. Radom Struct. Algorithms, 3(3): , [2] G. Caire, S. Shamai, ad S. Verdu. Noiseless data compressio with low desity parity check codes. I P. Gupta, G. Kramer, ad A. J. va Wijgaarde, editors, DIMACS Series i Discrete Mathematics ad Theoretical Computer Sciece, volume 66 of DIMACS Series i Discrete Mathematics ad Theoretical Computer Sciece, pages America Mathematical Society, [3] M. Cheraghchi. Capacity achievig codes from radomess coductors. I ISIT, pages IEEE, A exteded versio is available at [4] A. De ad T. Watso. Extractors ad lower bouds for locally samplable sources. TOCT, 4(1):3, [5] Y. Dodis, T. Ristepart, ad S. P. Vadha. Radomess codesers for efficietly samplable, seed-depedet sources. I R. Cramer, editor, TCC, volume 7194 of Lecture Notes i Computer Sciece, pages Spriger, [6] Y. Dodis ad A. Smith. Etropic security ad the ecryptio of high etropy messages. I J. Kilia, editor, TCC, volume 3378 of Lecture Notes i Computer Sciece, pages Spriger, [7] A. V. Goldberg ad M. Sipser. Compressio ad rakig. SIAM J. Comput., 20(3): , [8] V. Guruswami ad A. Smith. Codes for computatioally simple chaels: Explicit costructios with optimal rate. I FOCS, pages IEEE Computer Society, [9] V. Guruswami ad A. Smith. Optimal-rate code costructios for computatioally simple chaels. CoRR, abs/ , This is a exteded versio of [8]. [10] J. Håstad, R. Impagliazzo, L. A. Levi, ad M. Luby. A pseudoradom geerator from ay oe-way fuctio. SIAM J. Comput., 28(4): , [11] J. Justese. Class of costructive asymptotically good algebraic codes. IEEE Trasactios o Iformatio Theory, 18(5): , [12] M. Lagberg. Oblivious commuicatio chaels ad their capacity. IEEE Trasactios o Iformatio Theory, 54(1): , [13] R. J. Lipto. A ew approach to iformatio theory. I P. Ejalbert, E. W. Mayr, ad K. W. Wager, editors, STACS, volume 775 of Lecture Notes i Computer Sciece, pages Spriger, [14] S. Micali, C. Peikert, M. Suda, ad D. A. Wilso. Optimal error correctio for computatioally bouded oise. IEEE Trasactios o Iformatio Theory, 56(11): , [15] L. Trevisa ad S. P. Vadha. Extractig radomess from samplable distributios. I FOCS, pages IEEE Computer Society, [16] L. Trevisa, S. P. Vadha, ad D. Zuckerma. Compressio of samplable sources. Computatioal Complexity, 14(3): , [17] E. Viola. Extractors for circuit sources. I R. Ostrovsky, editor, FOCS, pages IEEE, [18] H. Wee. O pseudoetropy versus compressibility. I IEEE Coferece o Computatioal Complexity, pages IEEE Computer Society, 2004.

Lecture 11: Pseudorandom functions

Lecture 11: Pseudorandom functions COM S 6830 Cryptography Oct 1, 2009 Istructor: Rafael Pass 1 Recap Lecture 11: Pseudoradom fuctios Scribe: Stefao Ermo Defiitio 1 (Ge, Ec, Dec) is a sigle message secure ecryptio scheme if for all uppt

More information

Lecture 11: Channel Coding Theorem: Converse Part

Lecture 11: Channel Coding Theorem: Converse Part EE376A/STATS376A Iformatio Theory Lecture - 02/3/208 Lecture : Chael Codig Theorem: Coverse Part Lecturer: Tsachy Weissma Scribe: Erdem Bıyık I this lecture, we will cotiue our discussio o chael codig

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

UC Berkeley CS 170: Efficient Algorithms and Intractable Problems Handout 17 Lecturer: David Wagner April 3, Notes 17 for CS 170

UC Berkeley CS 170: Efficient Algorithms and Intractable Problems Handout 17 Lecturer: David Wagner April 3, Notes 17 for CS 170 UC Berkeley CS 170: Efficiet Algorithms ad Itractable Problems Hadout 17 Lecturer: David Wager April 3, 2003 Notes 17 for CS 170 1 The Lempel-Ziv algorithm There is a sese i which the Huffma codig was

More information

Lecture 1: Basic problems of coding theory

Lecture 1: Basic problems of coding theory Lecture 1: Basic problems of codig theory Error-Correctig Codes (Sprig 016) Rutgers Uiversity Swastik Kopparty Scribes: Abhishek Bhrushudi & Aditya Potukuchi Admiistrivia was discussed at the begiig of

More information

A Note on Matrix Rigidity

A Note on Matrix Rigidity A Note o Matrix Rigidity Joel Friedma Departmet of Computer Sciece Priceto Uiversity Priceto, NJ 08544 Jue 25, 1990 Revised October 25, 1991 Abstract I this paper we give a explicit costructio of matrices

More information

The multiplicative structure of finite field and a construction of LRC

The multiplicative structure of finite field and a construction of LRC IERG6120 Codig for Distributed Storage Systems Lecture 8-06/10/2016 The multiplicative structure of fiite field ad a costructio of LRC Lecturer: Keeth Shum Scribe: Zhouyi Hu Notatios: We use the otatio

More information

Information Theory Tutorial Communication over Channels with memory. Chi Zhang Department of Electrical Engineering University of Notre Dame

Information Theory Tutorial Communication over Channels with memory. Chi Zhang Department of Electrical Engineering University of Notre Dame Iformatio Theory Tutorial Commuicatio over Chaels with memory Chi Zhag Departmet of Electrical Egieerig Uiversity of Notre Dame Abstract A geeral capacity formula C = sup I(; Y ), which is correct for

More information

Basics of Probability Theory (for Theory of Computation courses)

Basics of Probability Theory (for Theory of Computation courses) Basics of Probability Theory (for Theory of Computatio courses) Oded Goldreich Departmet of Computer Sciece Weizma Istitute of Sciece Rehovot, Israel. oded.goldreich@weizma.ac.il November 24, 2008 Preface.

More information

Entropies & Information Theory

Entropies & Information Theory Etropies & Iformatio Theory LECTURE I Nilajaa Datta Uiversity of Cambridge,U.K. For more details: see lecture otes (Lecture 1- Lecture 5) o http://www.qi.damtp.cam.ac.uk/ode/223 Quatum Iformatio Theory

More information

ECE 564/645 - Digital Communication Systems (Spring 2014) Final Exam Friday, May 2nd, 8:00-10:00am, Marston 220

ECE 564/645 - Digital Communication Systems (Spring 2014) Final Exam Friday, May 2nd, 8:00-10:00am, Marston 220 ECE 564/645 - Digital Commuicatio Systems (Sprig 014) Fial Exam Friday, May d, 8:00-10:00am, Marsto 0 Overview The exam cosists of four (or five) problems for 100 (or 10) poits. The poits for each part

More information

Lecture 7: Channel coding theorem for discrete-time continuous memoryless channel

Lecture 7: Channel coding theorem for discrete-time continuous memoryless channel Lecture 7: Chael codig theorem for discrete-time cotiuous memoryless chael Lectured by Dr. Saif K. Mohammed Scribed by Mirsad Čirkić Iformatio Theory for Wireless Commuicatio ITWC Sprig 202 Let us first

More information

Lecture 9: Pseudo-random generators against space bounded computation,

Lecture 9: Pseudo-random generators against space bounded computation, Lecture 9: Pseudo-radom geerators agaist space bouded computatio, Primality Testig Topics i Pseudoradomess ad Complexity (Sprig 2018) Rutgers Uiversity Swastik Kopparty Scribes: Harsha Tirumala, Jiyu Zhag

More information

Lecture 9: Expanders Part 2, Extractors

Lecture 9: Expanders Part 2, Extractors Lecture 9: Expaders Part, Extractors Topics i Complexity Theory ad Pseudoradomess Sprig 013 Rutgers Uiversity Swastik Kopparty Scribes: Jaso Perry, Joh Kim I this lecture, we will discuss further the pseudoradomess

More information

A RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS

A RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS J. Japa Statist. Soc. Vol. 41 No. 1 2011 67 73 A RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS Yoichi Nishiyama* We cosider k-sample ad chage poit problems for idepedet data i a

More information

Lecture 14: Graph Entropy

Lecture 14: Graph Entropy 15-859: Iformatio Theory ad Applicatios i TCS Sprig 2013 Lecture 14: Graph Etropy March 19, 2013 Lecturer: Mahdi Cheraghchi Scribe: Euiwoog Lee 1 Recap Bergma s boud o the permaet Shearer s Lemma Number

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

Channel coding, linear block codes, Hamming and cyclic codes Lecture - 8

Channel coding, linear block codes, Hamming and cyclic codes Lecture - 8 Digital Commuicatio Chael codig, liear block codes, Hammig ad cyclic codes Lecture - 8 Ir. Muhamad Asial, MSc., PhD Ceter for Iformatio ad Commuicatio Egieerig Research (CICER) Electrical Egieerig Departmet

More information

Binary codes from graphs on triples and permutation decoding

Binary codes from graphs on triples and permutation decoding Biary codes from graphs o triples ad permutatio decodig J. D. Key Departmet of Mathematical Scieces Clemso Uiversity Clemso SC 29634 U.S.A. J. Moori ad B. G. Rodrigues School of Mathematics Statistics

More information

Asymptotic Coupling and Its Applications in Information Theory

Asymptotic Coupling and Its Applications in Information Theory Asymptotic Couplig ad Its Applicatios i Iformatio Theory Vicet Y. F. Ta Joit Work with Lei Yu Departmet of Electrical ad Computer Egieerig, Departmet of Mathematics, Natioal Uiversity of Sigapore IMS-APRM

More information

6.895 Essential Coding Theory October 20, Lecture 11. This lecture is focused in comparisons of the following properties/parameters of a code:

6.895 Essential Coding Theory October 20, Lecture 11. This lecture is focused in comparisons of the following properties/parameters of a code: 6.895 Essetial Codig Theory October 0, 004 Lecture 11 Lecturer: Madhu Suda Scribe: Aastasios Sidiropoulos 1 Overview This lecture is focused i comparisos of the followig properties/parameters of a code:

More information

Math 61CM - Solutions to homework 3

Math 61CM - Solutions to homework 3 Math 6CM - Solutios to homework 3 Cédric De Groote October 2 th, 208 Problem : Let F be a field, m 0 a fixed oegative iteger ad let V = {a 0 + a x + + a m x m a 0,, a m F} be the vector space cosistig

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

The Maximum-Likelihood Decoding Performance of Error-Correcting Codes

The Maximum-Likelihood Decoding Performance of Error-Correcting Codes The Maximum-Lielihood Decodig Performace of Error-Correctig Codes Hery D. Pfister ECE Departmet Texas A&M Uiversity August 27th, 2007 (rev. 0) November 2st, 203 (rev. ) Performace of Codes. Notatio X,

More information

Oblivious Transfer using Elliptic Curves

Oblivious Transfer using Elliptic Curves Oblivious Trasfer usig Elliptic Curves bhishek Parakh Louisiaa State Uiversity, ato Rouge, L May 4, 006 bstract: This paper proposes a algorithm for oblivious trasfer usig elliptic curves lso, we preset

More information

1 Review and Overview

1 Review and Overview DRAFT a fial versio will be posted shortly CS229T/STATS231: Statistical Learig Theory Lecturer: Tegyu Ma Lecture #3 Scribe: Migda Qiao October 1, 2013 1 Review ad Overview I the first half of this course,

More information

Increasing timing capacity using packet coloring

Increasing timing capacity using packet coloring 003 Coferece o Iformatio Scieces ad Systems, The Johs Hopkis Uiversity, March 4, 003 Icreasig timig capacity usig packet colorig Xi Liu ad R Srikat[] Coordiated Sciece Laboratory Uiversity of Illiois e-mail:

More information

Nonuniform Codes for Correcting Asymmetric Errors

Nonuniform Codes for Correcting Asymmetric Errors 2011 IEEE Iteratioal Symposium o Iformatio Theory Proceedigs Nouiform Codes for Correctig Asymmetric Errors Hogchao Zhou Electrical Egieerig Departmet Califoria Istitute of Techology Pasadea, CA 91125

More information

Problem Set 2 Solutions

Problem Set 2 Solutions CS271 Radomess & Computatio, Sprig 2018 Problem Set 2 Solutios Poit totals are i the margi; the maximum total umber of poits was 52. 1. Probabilistic method for domiatig sets 6pts Pick a radom subset S

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

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

Lecture 16: Monotone Formula Lower Bounds via Graph Entropy. 2 Monotone Formula Lower Bounds via Graph Entropy

Lecture 16: Monotone Formula Lower Bounds via Graph Entropy. 2 Monotone Formula Lower Bounds via Graph Entropy 15-859: Iformatio Theory ad Applicatios i TCS CMU: Sprig 2013 Lecture 16: Mootoe Formula Lower Bouds via Graph Etropy March 26, 2013 Lecturer: Mahdi Cheraghchi Scribe: Shashak Sigh 1 Recap Graph Etropy:

More information

Intro to Learning Theory

Intro to Learning Theory Lecture 1, October 18, 2016 Itro to Learig Theory Ruth Urer 1 Machie Learig ad Learig Theory Comig soo 2 Formal Framework 21 Basic otios I our formal model for machie learig, the istaces to be classified

More information

THE interference channel problem describes a setup where multiple pairs of transmitters and receivers share a communication

THE interference channel problem describes a setup where multiple pairs of transmitters and receivers share a communication Trade-off betwee Commuicatio ad Cooperatio i the Iterferece Chael Farhad Shirai EECS Departmet Uiversity of Michiga A Arbor,USA Email: fshirai@umich.edu S. Sadeep Pradha EECS Departmet Uiversity of Michiga

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

Machine Learning Theory Tübingen University, WS 2016/2017 Lecture 12

Machine Learning Theory Tübingen University, WS 2016/2017 Lecture 12 Machie Learig Theory Tübige Uiversity, WS 06/07 Lecture Tolstikhi Ilya Abstract I this lecture we derive risk bouds for kerel methods. We will start by showig that Soft Margi kerel SVM correspods to miimizig

More information

Lecture 11: Hash Functions and Random Oracle Model

Lecture 11: Hash Functions and Random Oracle Model CS 7810 Foudatios of Cryptography October 16, 017 Lecture 11: Hash Fuctios ad Radom Oracle Model Lecturer: Daiel Wichs Scribe: Akshar Varma 1 Topic Covered Defiitio of Hash Fuctios Merkle-Damgaård Theorem

More information

ROTATION-EQUIVALENCE CLASSES OF BINARY VECTORS. 1. Introduction

ROTATION-EQUIVALENCE CLASSES OF BINARY VECTORS. 1. Introduction t m Mathematical Publicatios DOI: 10.1515/tmmp-2016-0033 Tatra Mt. Math. Publ. 67 (2016, 93 98 ROTATION-EQUIVALENCE CLASSES OF BINARY VECTORS Otokar Grošek Viliam Hromada ABSTRACT. I this paper we study

More information

Lecture 10: Universal coding and prediction

Lecture 10: Universal coding and prediction 0-704: Iformatio Processig ad Learig Sprig 0 Lecture 0: Uiversal codig ad predictio Lecturer: Aarti Sigh Scribes: Georg M. Goerg Disclaimer: These otes have ot bee subjected to the usual scrutiy reserved

More information

Last time, we talked about how Equation (1) can simulate Equation (2). We asserted that Equation (2) can also simulate Equation (1).

Last time, we talked about how Equation (1) can simulate Equation (2). We asserted that Equation (2) can also simulate Equation (1). 6896 Quatum Complexity Theory Sept 23, 2008 Lecturer: Scott Aaroso Lecture 6 Last Time: Quatum Error-Correctio Quatum Query Model Deutsch-Jozsa Algorithm (Computes x y i oe query) Today: Berstei-Vazirii

More information

Large holes in quasi-random graphs

Large holes in quasi-random graphs Large holes i quasi-radom graphs Joaa Polcy Departmet of Discrete Mathematics Adam Mickiewicz Uiversity Pozań, Polad joaska@amuedupl Submitted: Nov 23, 2006; Accepted: Apr 10, 2008; Published: Apr 18,

More information

Accuracy Assessment for High-Dimensional Linear Regression

Accuracy Assessment for High-Dimensional Linear Regression Uiversity of Pesylvaia ScholarlyCommos Statistics Papers Wharto Faculty Research -016 Accuracy Assessmet for High-Dimesioal Liear Regressio Toy Cai Uiversity of Pesylvaia Zijia Guo Uiversity of Pesylvaia

More information

On the Capacity of Memoryless Adversary

On the Capacity of Memoryless Adversary O the Capacity of Memoryless Adversary Arya Mazumdar Departmet of ECE Uiversity of Miesota Twi Cities Mieapolis, MN 55455 email: arya@um.edu Abstract I this paper, we study a model of commuicatio uder

More information

Here, e(a, B) is defined as the number of edges between A and B in the n dimensional boolean hypercube.

Here, e(a, B) is defined as the number of edges between A and B in the n dimensional boolean hypercube. Lecture 2 Topics i Complexity Theory ad Pseudoradomess (Sprig 2013) Rutgers Uiversity Swastik Kopparty Scribes: Amey Bhagale, Mrial Kumar 1 Overview I this lecture, we will complete the proof of formula

More information

Random Walks on Discrete and Continuous Circles. by Jeffrey S. Rosenthal School of Mathematics, University of Minnesota, Minneapolis, MN, U.S.A.

Random Walks on Discrete and Continuous Circles. by Jeffrey S. Rosenthal School of Mathematics, University of Minnesota, Minneapolis, MN, U.S.A. Radom Walks o Discrete ad Cotiuous Circles by Jeffrey S. Rosethal School of Mathematics, Uiversity of Miesota, Mieapolis, MN, U.S.A. 55455 (Appeared i Joural of Applied Probability 30 (1993), 780 789.)

More information

1 Hash tables. 1.1 Implementation

1 Hash tables. 1.1 Implementation Lecture 8 Hash Tables, Uiversal Hash Fuctios, Balls ad Bis Scribes: Luke Johsto, Moses Charikar, G. Valiat Date: Oct 18, 2017 Adapted From Virgiia Williams lecture otes 1 Hash tables A hash table is a

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

Fixed-Threshold Polar Codes

Fixed-Threshold Polar Codes Fixed-Threshold Polar Codes Jig Guo Uiversity of Cambridge jg582@cam.ac.uk Albert Guillé i Fàbregas ICREA & Uiversitat Pompeu Fabra Uiversity of Cambridge guille@ieee.org Jossy Sayir Uiversity of Cambridge

More information

REGULARIZATION OF CERTAIN DIVERGENT SERIES OF POLYNOMIALS

REGULARIZATION OF CERTAIN DIVERGENT SERIES OF POLYNOMIALS REGULARIZATION OF CERTAIN DIVERGENT SERIES OF POLYNOMIALS LIVIU I. NICOLAESCU ABSTRACT. We ivestigate the geeralized covergece ad sums of series of the form P at P (x, where P R[x], a R,, ad T : R[x] R[x]

More information

Lecture 6: Source coding, Typicality, and Noisy channels and capacity

Lecture 6: Source coding, Typicality, and Noisy channels and capacity 15-859: Iformatio Theory ad Applicatios i TCS CMU: Sprig 2013 Lecture 6: Source codig, Typicality, ad Noisy chaels ad capacity Jauary 31, 2013 Lecturer: Mahdi Cheraghchi Scribe: Togbo Huag 1 Recap Uiversal

More information

Cooperative Communication Fundamentals & Coding Techniques

Cooperative Communication Fundamentals & Coding Techniques 3 th ICACT Tutorial Cooperative commuicatio fudametals & codig techiques Cooperative Commuicatio Fudametals & Codig Techiques 0..4 Electroics ad Telecommuicatio Research Istitute Kiug Jug 3 th ICACT Tutorial

More information

Entropy and Ergodic Theory Lecture 5: Joint typicality and conditional AEP

Entropy and Ergodic Theory Lecture 5: Joint typicality and conditional AEP Etropy ad Ergodic Theory Lecture 5: Joit typicality ad coditioal AEP 1 Notatio: from RVs back to distributios Let (Ω, F, P) be a probability space, ad let X ad Y be A- ad B-valued discrete RVs, respectively.

More information

Commutativity in Permutation Groups

Commutativity in Permutation Groups Commutativity i Permutatio Groups Richard Wito, PhD Abstract I the group Sym(S) of permutatios o a oempty set S, fixed poits ad trasiet poits are defied Prelimiary results o fixed ad trasiet poits are

More information

New Definition of Density on Knapsack Cryptosystems

New Definition of Density on Knapsack Cryptosystems Africacryt008@Casablaca 008.06.1 New Defiitio of Desity o Kasac Crytosystems Noboru Kuihiro The Uiversity of Toyo, Jaa 1/31 Kasac Scheme rough idea Public Key: asac: a={a 1, a,, a } Ecrytio: message m=m

More information

Machine Learning Theory (CS 6783)

Machine Learning Theory (CS 6783) Machie Learig Theory (CS 6783) Lecture 2 : Learig Frameworks, Examples Settig up learig problems. X : istace space or iput space Examples: Computer Visio: Raw M N image vectorized X = 0, 255 M N, SIFT

More information

Efficient GMM LECTURE 12 GMM II

Efficient GMM LECTURE 12 GMM II DECEMBER 1 010 LECTURE 1 II Efficiet The estimator depeds o the choice of the weight matrix A. The efficiet estimator is the oe that has the smallest asymptotic variace amog all estimators defied by differet

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

On the Linear Complexity of Feedback Registers

On the Linear Complexity of Feedback Registers O the Liear Complexity of Feedback Registers A. H. Cha M. Goresky A. Klapper Northeaster Uiversity Abstract I this paper, we study sequeces geerated by arbitrary feedback registers (ot ecessarily feedback

More information

1. Universal v.s. non-universal: know the source distribution or not.

1. Universal v.s. non-universal: know the source distribution or not. 28. Radom umber geerators Let s play the followig game: Give a stream of Ber( p) bits, with ukow p, we wat to tur them ito pure radom bits, i.e., idepedet fair coi flips Ber( / 2 ). Our goal is to fid

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

Lecture 3 The Lebesgue Integral

Lecture 3 The Lebesgue Integral Lecture 3: The Lebesgue Itegral 1 of 14 Course: Theory of Probability I Term: Fall 2013 Istructor: Gorda Zitkovic Lecture 3 The Lebesgue Itegral The costructio of the itegral Uless expressly specified

More information

62. Power series Definition 16. (Power series) Given a sequence {c n }, the series. c n x n = c 0 + c 1 x + c 2 x 2 + c 3 x 3 +

62. Power series Definition 16. (Power series) Given a sequence {c n }, the series. c n x n = c 0 + c 1 x + c 2 x 2 + c 3 x 3 + 62. Power series Defiitio 16. (Power series) Give a sequece {c }, the series c x = c 0 + c 1 x + c 2 x 2 + c 3 x 3 + is called a power series i the variable x. The umbers c are called the coefficiets of

More information

Information Theory and Statistics Lecture 4: Lempel-Ziv code

Information Theory and Statistics Lecture 4: Lempel-Ziv code Iformatio Theory ad Statistics Lecture 4: Lempel-Ziv code Łukasz Dębowski ldebowsk@ipipa.waw.pl Ph. D. Programme 203/204 Etropy rate is the limitig compressio rate Theorem For a statioary process (X i)

More information

Review of Elementary Cryptography. For more material, see my notes of CSE 5351, available on my webpage

Review of Elementary Cryptography. For more material, see my notes of CSE 5351, available on my webpage Review of Elemetary Cryptography For more material, see my otes of CSE 5351, available o my webpage Outlie Security (CPA, CCA, sematic security, idistiguishability) RSA ElGamal Homomorphic ecryptio 2 Two

More information

Lecture 15: Strong, Conditional, & Joint Typicality

Lecture 15: Strong, Conditional, & Joint Typicality EE376A/STATS376A Iformatio Theory Lecture 15-02/27/2018 Lecture 15: Strog, Coditioal, & Joit Typicality Lecturer: Tsachy Weissma Scribe: Nimit Sohoi, William McCloskey, Halwest Mohammad I this lecture,

More information

Discrete Mathematics for CS Spring 2008 David Wagner Note 22

Discrete Mathematics for CS Spring 2008 David Wagner Note 22 CS 70 Discrete Mathematics for CS Sprig 2008 David Wager Note 22 I.I.D. Radom Variables Estimatig the bias of a coi Questio: We wat to estimate the proportio p of Democrats i the US populatio, by takig

More information

Resolution Proofs of Generalized Pigeonhole Principles

Resolution Proofs of Generalized Pigeonhole Principles Resolutio Proofs of Geeralized Pigeohole Priciples Samuel R. Buss Departmet of Mathematics Uiversity of Califoria, Berkeley Győrgy Turá Departmet of Mathematics, Statistics, ad Computer Sciece Uiversity

More information

5 Birkhoff s Ergodic Theorem

5 Birkhoff s Ergodic Theorem 5 Birkhoff s Ergodic Theorem Amog the most useful of the various geeralizatios of KolmogorovâĂŹs strog law of large umbers are the ergodic theorems of Birkhoff ad Kigma, which exted the validity of the

More information

Journal of Multivariate Analysis. Superefficient estimation of the marginals by exploiting knowledge on the copula

Journal of Multivariate Analysis. Superefficient estimation of the marginals by exploiting knowledge on the copula Joural of Multivariate Aalysis 102 (2011) 1315 1319 Cotets lists available at ScieceDirect Joural of Multivariate Aalysis joural homepage: www.elsevier.com/locate/jmva Superefficiet estimatio of the margials

More information

Lecture 4: April 10, 2013

Lecture 4: April 10, 2013 TTIC/CMSC 1150 Mathematical Toolkit Sprig 01 Madhur Tulsiai Lecture 4: April 10, 01 Scribe: Haris Agelidakis 1 Chebyshev s Iequality recap I the previous lecture, we used Chebyshev s iequality to get a

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

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

Lecture 4: Unique-SAT, Parity-SAT, and Approximate Counting

Lecture 4: Unique-SAT, Parity-SAT, and Approximate Counting Advaced Complexity Theory Sprig 206 Lecture 4: Uique-SAT, Parity-SAT, ad Approximate Coutig Prof. Daa Moshkovitz Scribe: Aoymous Studet Scribe Date: Fall 202 Overview I this lecture we begi talkig about

More information

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

Endcoding Complexity Versus Minimum Distance

Endcoding Complexity Versus Minimum Distance Edcodig Complexity Versus Miimum Distace Louay M.J. Bazzi, Sajoy K. Mitter Abstract We establish a boud o the miimum distace of a biary error correctig code give costraits o the computatioal time-space

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

Expectation and Variance of a random variable

Expectation and Variance of a random variable Chapter 11 Expectatio ad Variace of a radom variable The aim of this lecture is to defie ad itroduce mathematical Expectatio ad variace of a fuctio of discrete & cotiuous radom variables ad the distributio

More information

2000 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes

2000 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes 2000 IEEE. Persoal use of this material is permitted. However, permissio to reprit/republish this material for advertisig or promotioal purposes or for creatig ew collective works for resale or redistributio

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

Polynomial identity testing and global minimum cut

Polynomial identity testing and global minimum cut CHAPTER 6 Polyomial idetity testig ad global miimum cut I this lecture we will cosider two further problems that ca be solved usig probabilistic algorithms. I the first half, we will cosider the problem

More information

Message Authentication Codes. Reading: Chapter 4 of Katz & Lindell

Message Authentication Codes. Reading: Chapter 4 of Katz & Lindell Message Autheticatio Codes Readig: Chapter 4 of Katz & Lidell 1 Message autheticatio Bob receives a message m from Alice, he wats to ow (Data origi autheticatio) whether the message was really set by Alice.

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

Chapter IV Integration Theory

Chapter IV Integration Theory Chapter IV Itegratio Theory Lectures 32-33 1. Costructio of the itegral I this sectio we costruct the abstract itegral. As a matter of termiology, we defie a measure space as beig a triple (, A, µ), where

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

Vector Permutation Code Design Algorithm. Danilo SILVA and Weiler A. FINAMORE

Vector Permutation Code Design Algorithm. Danilo SILVA and Weiler A. FINAMORE Iteratioal Symposium o Iformatio Theory ad its Applicatios, ISITA2004 Parma, Italy, October 10 13, 2004 Vector Permutatio Code Desig Algorithm Dailo SILVA ad Weiler A. FINAMORE Cetro de Estudos em Telecomuicações

More information

Multiterminal source coding with complementary delivery

Multiterminal source coding with complementary delivery Iteratioal Symposium o Iformatio Theory ad its Applicatios, ISITA2006 Seoul, Korea, October 29 November 1, 2006 Multitermial source codig with complemetary delivery Akisato Kimura ad Tomohiko Uyematsu

More information

Lecture 2: April 3, 2013

Lecture 2: April 3, 2013 TTIC/CMSC 350 Mathematical Toolkit Sprig 203 Madhur Tulsiai Lecture 2: April 3, 203 Scribe: Shubhedu Trivedi Coi tosses cotiued We retur to the coi tossig example from the last lecture agai: Example. Give,

More information

Lower Bounds on Black-Box Reductions of Hitting to Density Estimation

Lower Bounds on Black-Box Reductions of Hitting to Density Estimation Lower Bouds o Black-Box Reductios of Hittig to Desity Estimatio Roei Tell Departmet of Computer Sciece ad Applied Mathematics, Weizma Istitute of Sciece, Rehovot, Israel roei.tell@weizma.ac.il Abstract

More information

2 High-level Complexity vs. Concrete Complexity

2 High-level Complexity vs. Concrete Complexity COMS 6998: Advaced Complexity Sprig 2017 Lecture 1: Course Itroductio ad Boolea Formulas Lecturer: Rocco Servedio Scribes: Jiahui Liu, Kailash Karthik Meiyappa 1 Overview of Topics 1. Boolea formulas (examples,

More information

Advanced Stochastic Processes.

Advanced Stochastic Processes. Advaced Stochastic Processes. David Gamarik LECTURE 2 Radom variables ad measurable fuctios. Strog Law of Large Numbers (SLLN). Scary stuff cotiued... Outlie of Lecture Radom variables ad measurable fuctios.

More information

Binary classification, Part 1

Binary classification, Part 1 Biary classificatio, Part 1 Maxim Ragisky September 25, 2014 The problem of biary classificatio ca be stated as follows. We have a radom couple Z = (X,Y ), where X R d is called the feature vector ad Y

More information

Lecture 01: the Central Limit Theorem. 1 Central Limit Theorem for i.i.d. random variables

Lecture 01: the Central Limit Theorem. 1 Central Limit Theorem for i.i.d. random variables CSCI-B609: A Theorist s Toolkit, Fall 06 Aug 3 Lecture 0: the Cetral Limit Theorem Lecturer: Yua Zhou Scribe: Yua Xie & Yua Zhou Cetral Limit Theorem for iid radom variables Let us say that we wat to aalyze

More information

Information Theory and Coding

Information Theory and Coding Sol. Iformatio Theory ad Codig. The capacity of a bad-limited additive white Gaussia (AWGN) chael is give by C = Wlog 2 ( + σ 2 W ) bits per secod(bps), where W is the chael badwidth, is the average power

More information

Arithmetic Distribution Matching

Arithmetic Distribution Matching Arithmetic Distributio Matchig Sebastia Baur ad Georg Böcherer Istitute for Commuicatios Egieerig Techische Uiversität Müche, Germay Email: baursebastia@mytum.de,georg.boecherer@tum.de arxiv:48.393v [cs.it]

More information

Lecture Notes for Analysis Class

Lecture Notes for Analysis Class Lecture Notes for Aalysis Class Topological Spaces A topology for a set X is a collectio T of subsets of X such that: (a) X ad the empty set are i T (b) Uios of elemets of T are i T (c) Fiite itersectios

More information

2 Markov Chain Monte Carlo Sampling

2 Markov Chain Monte Carlo Sampling 22 Part I. Markov Chais ad Stochastic Samplig Figure 10: Hard-core colourig of a lattice. 2 Markov Chai Mote Carlo Samplig We ow itroduce Markov chai Mote Carlo (MCMC) samplig, which is a extremely importat

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

Quantum Lower Bounds for the Goldreich-Levin Problem

Quantum Lower Bounds for the Goldreich-Levin Problem Quatum Lower Bouds for the Goldreich-Levi Problem Mark Adcock 1, Richard Cleve 1, Kazuo Iwama, Raymod Putra,4, Shigeru Yamashita 3, 1 Departmet of Computer Sciece, Uiversity of Calgary, Caada Graduate

More information

Quantum Computing Lecture 7. Quantum Factoring

Quantum Computing Lecture 7. Quantum Factoring Quatum Computig Lecture 7 Quatum Factorig Maris Ozols Quatum factorig A polyomial time quatum algorithm for factorig umbers was published by Peter Shor i 1994. Polyomial time meas that the umber of gates

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