6 Matrix Concentration Bounds
|
|
- Hope Wilkerson
- 6 years ago
- Views:
Transcription
1 6 Matix Concentation Bounds Concentation bounds ae inequalities that bound pobabilities of deviations by a andom vaiable fom some value, often its mean. Infomally, they show the pobability that a andom vaiable deviates fom its expectation is small. A basic example of andom vaiables being concentated aound the mean is stated by law of lage numbes that says unde mild conditions, sum of independent andom vaiables is close to thei expectation with a high pobability. Concentation bounds, which ae also known as tail bounds, ae a majo tool in analyzing aveage behaviou of algoithms, e.g. estimating the failue pobability o establishing high pobability bounds on unning time and space usage. In this lectue, we give an intoduction to some of concentation bounds on andom vaiables and andom matices. 6.1 Basic Tail Bounds If X is a andom vaiable and a 2 R is a eal value, then P[X a] and P[X apple a] ae called uppe tail and lowe tail of the distibution on X, espectively. Concentation bounds might bound one of the tails o both. We stat with the most basic yet fundamental tail bound, called as Makov s Inequality. Theoem (Makov s Inequality). Let X be a non-negative andom vaiable. Then fo all a>0 P(X Poof. Define an indicato andom vaiable I a = And that completes the poof. a) apple E[X] a ( 1 if X a 0 othewise. Note in both cases X ai a, theefoe E[X] a E[I a ] = a P(X a) Makov s inequality is the best possible tail bound one can get if all he knows is non-negativity of andom vaiable and its expectation, yet it is often too weak to yeild useful esults. Tail bounds can be impoved if moe infomation about X, e.g. its vaiance, is available. Below, we state a significantly stonge bound that employs exta statistics. Theoem (Chebyshev s Inequality). Fo any andom vaiable X, and any a>0 P ( X E[X] a) apple Va[X] a 2 Poof. Let s apply Makov s inequality on andom vaiable X =(Y E[Y ]) 2 and scala a = b 2, P (Y E[Y ]) 2 b 2 apple E (Y E[Y ]) 2 Note that P (Y E[Y ]) 2 b 2 = P ( Y E[Y ] b), and E (Y E[Y ]) 2 = Va(Y ) is the definition of vaiance. Theefoe P ( Y E[Y ] b) apple Va(Y ) b 2 b 2
2 The use of Chebyshev s inequality indicates that with pobability lage than 1 1/a 2, a andom vaiable X will fall within a times the standad deviation aound E[X]. Fo example, 75% of the times a andom value falls in the inteval [E[X] 2Va(X), E[X]+2Va(X)]. Both Makov s and Chebyshev s inequalities povide polynomially decaying bounds in amount of deviation (i.e. a in the fomula). Moe inteesting ae concentation bounds in which deviation pobabilities decay exponentially in the distance fom mean. Theoem (Chenoff-Hoeffding Inequality). Conside a set of independent andom vaiables {X 1,X 2,,X }. If we know each andom vaiable is bounded as a i apple X i apple b i with i = b i a i, then fo M = P X i and any paamete 2 (0, 1/2) 2 2 P ( M E[M] >) apple 2exp P Although Chenoff-Hoeffding inequality povides a stonge exponentially deceasing bound, it equies andom vaiables to be independent a condition that neithe the Makov s no the Chebyshev s inequalities equie. Befoe we move on to concentation bounds fo matices, we state a concete vesion of Chenoff bound that has an analogous in matices. This vesion bounds the tail distibution of a sum of independent 0, 1 andom vaiables that ae not necessaily distibuted identically. Theoem (Chenoff s Inequality). Let {X 1,X 2,,X } be a sequence of independent 0, 1 andom vaiables. Define M = P X i and µ = E[M]. Then fo any >0: e µ P (M (1 + )µ) apple (1 + ) Matix Concentation Bounds In applications, it is common that a andom matix can be expessed as a sum of independent andom matices[1]. Fo example, the covaiance of X 2 R n d can be witten as X T X = P n xt i x i whee x i denotes i-th ow of X. In this section, we state two common bounds on andom matices[1] Matix Chenoff Bound Chenoff s Inequality has an analogous in matix setting; the 0, 1 andom vaiables tanslate to positivesemidefinite andom matices which ae unifomly bounded on thei eigenvalues. Theoem (Matix Chenoff bound). Let {X 1,X 2,,X } be a finite sequence of independent d d dimensional andom matices such that each X i is positive semi-definite (X i 0) and is uppe bounded on eigenvalues, i.e. max(x i ) apple R fo a positive constant R. Define µ min = min ( P E[X i]) and µ max = max ( P E[X i]), then fo any 2 [0, 1] And fo any 0 P P min max E! X i min! X i! apple apple (1 )µ min apple d.! X i 2 i e (1 ) µ min R log d! apple apple (1 + )µ max apple d. e (1 + ) 1+ µ min /R µ max/r
3 E max! X i apple 1.8 µ max + R log d Hee, we state an example to demonstate the application of Matix Chenoff bound. Example [Random Submatix]. [1] Conside a fix matix C =[c 1,c 2,,c n ] 2 R d n. Let s constuct a submatix Z 2 R d n C by andomly picking columns of C with pobability s/n. If column c i is selected, we set i-th column of Z to c i (i.e. z i = c i ), othewise z i =0. We ae inteested in finding the expected lagest singula value of Z and expected smallest singula values of Z. Solution: Let s define a andom vaiable i fo each column c i. If c i is selected, we set i =1and othewise we set it to zeo. Theefoe i is a 0, 1 andom vaiable as following ( 1 with pobability s/n i = 0 othewise Putting them in a diagonal matix =diag( 1, 2,, n) 2 R n n allows us to wite Z as Z = C. In ode to be able to use the Matix Chenoff bound, we need to convet singula values to eigenvalues. Note that singula values of Z ae elated to eigenvalues of ZZ T as i(z) 2 = i (ZZ T ); theefoe if we conside singula values and eigen values indexed in descending ode, we can bound lagest singula value of Z as and smallest singula value as 1(Z) 2 = 1 (ZZ T ) d(z) 2 = d (ZZ T ) Next, we wite ZZ T in tems of sum of andom matices that satisfy conditions of theoem 6.2.1: If we denote each summand as X i, ZZ T = ZZ T =(C )(C ) T = C 2 C T = C C T = ic i c T i = X i E[ZZ T ]= E[X i ]= s n ic i c T i c i c T i = s n CCT Note that X i s ae independent, positive semi-definite, and bounded as max(x i ) applekc i k 2. So the maximum eigenvalue of all summands is bounded as R = max kc ik 2. The mean paametes satisfy 1appleiapplen and µ max = n (E[ZZ T ]) = s n 1 (C) 2 µ min = d (E[ZZ T ]) = s n d (C) 2 Applying Matix Chenoff inequality we obtain E 1(Z) 2 = E d(zz T ) apple 1.8 ( s n ) 1(C) 2 + max 1appleiapplen kc ik 2 log d E d(z) 2 = E d(zz T ) 0.6 ( s n ) d(c) 2 max 1appleiapplen kc ik 2 log d As this bound shows andom matix Z gets a shae of the spectum of C in popotion to the numbe of columns it picks. In othe wods, matix C has n columns which andom matix Z includes about s of them in expectation, and the bound is showing each singula value i (Z) 2 of Z inheits an s/n shae of i(c) 2 fo all 1 apple i apple d.
4 6.2.2 Matix Benstein Inequality Benstein Inequality dops the condition of positive semi-definitness and concens a sum of zeo-mean bounded andom matices. Theoem (Matix Benstein Inequality). Let {X 1,X 2,,X } be a set of independent d 1 d 2 dimensional andom matices with E[X i ]=0and P kx i k 2 apple R fo a positive constant R and fo all X i s. Define vaiance paamete as = max E(X ixi T ), P E(XT i X i), then fo any t 0 t 2 /2 P X i t apple (d 1 + d 2 )exp + Rt/3 E X i apple p 2 log(d 1 + d 2 )+ R 3 log(d 1 + d 2 ) Example [Randomize Matix Multiplication]. [1] Conside matix B 2 R d1 n with unit nom columns (i.e. kb i k =1, 81 apple i apple n) and matix C 2 R n d 2 with unit nom ows (i.e. kc i k =1, 81 apple i apple n). We ae inteested in computing an appoximation to the poduct BC 2 R d 1 d 2 using a andom sample of columns of B and ows of C. solution: Fist note that if we constuct a andom vaiable S =(nb j c j ) 2 R d 1 d 2 whee j 2 [1,n] is sampled unifomly at andom, then S will be an unbiased estimato fo the poduct BC. This is tue because: E[S] =E nb j c j = n (b k c k P(k = j)) = n (b k c k ( 1 n )) = (b k c k )=BC k=1 Although S is an unbiased estimato fo BC, it has a lage vaiance. In ode to educe the vaiance we compute m independent copy of S and aveage them as Z = 1 P m m S i.now Z is a bette appoximation fo BC. The eo we ae inteested to compute is m X e = E [kz BCk 2 ]=E ( 1 m S 1 i m BC) 2 Theefoe each summand is X i = 1 m (S i BC). Note that summands ae independent and E[X i ]=0. Befoe we bound spectal nom of summands, we use symmetization technique to say E [kz BCk] apple 2 m m E X i S i Whee i ae independenet Rademache andom vaiables, independent of S i s. Theefoe k=1 kx i k 2 applek i S i kappleks i k = nkb j kkc i k = n Theefoe the uppe bound R on spectal nom of any summand is n. Now we compute the vaiance in two steps: E[S i Si T ]= n 2 (b j c j )(b j c j ) T P(j = i) = n 2 kc j k 2 b j b T j = n 2 BB T And E[S T i S i ]= n 2 (b j c j ) T (b j c j )P(j = i) = k=1 n 2 kb j k 2 c jt c j = n 2 C T C
5 Theefoe we obtain = max{kmn BB T k, kmn C T Ck} = mn max{kbk 2, kck 2 }. Applying the Benstein Bound E [kz BCk] apple 2 m m E X i S i apple 2 p 2mn log(d1 + d 2 ) max{kbk 2, kck 2 } + n m 3 log(d 1 + d 2 ) = 8n m log(d 1 + d 2 ) max{kbk 2, kck 2 } + 2n 3m log(d 1 + d 2 ) In ode to make sense out of this inequality, let s take a look at numeic ank of B and C. As we know, numeic ank of a matix is defined as nank(a) =kak 2 F /kak2 2. Since B and C has unit columns and unit ows, espectively then kbk 2 F = kck2 F = n. Then the geometic mean of numeic anks will be µ = p s n nank(b) nank(c) = 2 n kbk 2 = 2 kck2 2 kbk 2 kck 2 Conveting the eo bound to a elative scale we obtain E [kz BCk] kbk 2 kck 2 8n apple m log(d 1 + d 2 ) max{kbk 2, kck 2 } + 2 kbk 2 kck 2 3 8n apple m log(d d 2 ) min{kbk 2, kck 2 } s n 8 = kbk 2 kck 2 m log(d 1 + d 2 ) 8µn apple m log(d 1 + d 2 )+ 2 µ log(d 1 + d 2 ) 3 m n log(d 1 + d 2 ) mkbk 2 kck 2 µ log(d 1 + d 2 ) m kbk 2 kck 2 min 2 {kbk 2, kck 2 } µ log(d 1 + d 2 ) m Note last inequality follows since min 2 {kbk 2, kck 2 } =min{kbk 2 2, kck2 2 } and 1 applekbk 2, kck 2 apple n. Setting the elative eo to be less than, the numbe of samples (m) will be m = ( 2 µn log(d 1 +d 2 )). This shows the numbe of samples we need to appoximate poduct of two matices is popotional to the geometic mean of thei numeic ank. And that makes sense because if two matices of high ank, it means thee ae many distinct diections in thei spectum, theefoe we need to pick many samples to get a good appoximation to thei poduct.
6
7 Bibliogaphy [1] Joel A Topp. Use-fiendly tail bounds fo sums of andom matices. Foundations of Computational Mathematics, 12(4): ,
Surveillance Points in High Dimensional Spaces
Société de Calcul Mathématique SA Tools fo decision help since 995 Suveillance Points in High Dimensional Spaces by Benad Beauzamy Januay 06 Abstact Let us conside any compute softwae, elying upon a lage
More informationGoodness-of-fit for composite hypotheses.
Section 11 Goodness-of-fit fo composite hypotheses. Example. Let us conside a Matlab example. Let us geneate 50 obsevations fom N(1, 2): X=nomnd(1,2,50,1); Then, unning a chi-squaed goodness-of-fit test
More informationStanford University CS259Q: Quantum Computing Handout 8 Luca Trevisan October 18, 2012
Stanfod Univesity CS59Q: Quantum Computing Handout 8 Luca Tevisan Octobe 8, 0 Lectue 8 In which we use the quantum Fouie tansfom to solve the peiod-finding poblem. The Peiod Finding Poblem Let f : {0,...,
More information10/04/18. P [P(x)] 1 negl(n).
Mastemath, Sping 208 Into to Lattice lgs & Cypto Lectue 0 0/04/8 Lectues: D. Dadush, L. Ducas Scibe: K. de Boe Intoduction In this lectue, we will teat two main pats. Duing the fist pat we continue the
More information3.1 Random variables
3 Chapte III Random Vaiables 3 Random vaiables A sample space S may be difficult to descibe if the elements of S ae not numbes discuss how we can use a ule by which an element s of S may be associated
More informationMultiple Criteria Secretary Problem: A New Approach
J. Stat. Appl. Po. 3, o., 9-38 (04 9 Jounal of Statistics Applications & Pobability An Intenational Jounal http://dx.doi.og/0.785/jsap/0303 Multiple Citeia Secetay Poblem: A ew Appoach Alaka Padhye, and
More informationHypothesis Test and Confidence Interval for the Negative Binomial Distribution via Coincidence: A Case for Rare Events
Intenational Jounal of Contempoay Mathematical Sciences Vol. 12, 2017, no. 5, 243-253 HIKARI Ltd, www.m-hikai.com https://doi.og/10.12988/ijcms.2017.7728 Hypothesis Test and Confidence Inteval fo the Negative
More information6 PROBABILITY GENERATING FUNCTIONS
6 PROBABILITY GENERATING FUNCTIONS Cetain deivations pesented in this couse have been somewhat heavy on algeba. Fo example, detemining the expectation of the Binomial distibution (page 5.1 tuned out to
More informationON INDEPENDENT SETS IN PURELY ATOMIC PROBABILITY SPACES WITH GEOMETRIC DISTRIBUTION. 1. Introduction. 1 r r. r k for every set E A, E \ {0},
ON INDEPENDENT SETS IN PURELY ATOMIC PROBABILITY SPACES WITH GEOMETRIC DISTRIBUTION E. J. IONASCU and A. A. STANCU Abstact. We ae inteested in constucting concete independent events in puely atomic pobability
More informationA generalization of the Bernstein polynomials
A genealization of the Benstein polynomials Halil Ouç and Geoge M Phillips Mathematical Institute, Univesity of St Andews, Noth Haugh, St Andews, Fife KY16 9SS, Scotland Dedicated to Philip J Davis This
More informationThe Substring Search Problem
The Substing Seach Poblem One algoithm which is used in a vaiety of applications is the family of substing seach algoithms. These algoithms allow a use to detemine if, given two chaacte stings, one is
More informationIntroduction to Mathematical Statistics Robert V. Hogg Joeseph McKean Allen T. Craig Seventh Edition
Intoduction to Mathematical Statistics Robet V. Hogg Joeseph McKean Allen T. Caig Seventh Edition Peason Education Limited Edinbugh Gate Halow Essex CM2 2JE England and Associated Companies thoughout the
More informationNumerical Integration
MCEN 473/573 Chapte 0 Numeical Integation Fall, 2006 Textbook, 0.4 and 0.5 Isopaametic Fomula Numeical Integation [] e [ ] T k = h B [ D][ B] e B Jdsdt In pactice, the element stiffness is calculated numeically.
More informationOn the Poisson Approximation to the Negative Hypergeometric Distribution
BULLETIN of the Malaysian Mathematical Sciences Society http://mathusmmy/bulletin Bull Malays Math Sci Soc (2) 34(2) (2011), 331 336 On the Poisson Appoximation to the Negative Hypegeometic Distibution
More information556: MATHEMATICAL STATISTICS I
556: MATHEMATICAL STATISTICS I CHAPTER 5: STOCHASTIC CONVERGENCE The following efinitions ae state in tems of scala anom vaiables, but exten natually to vecto anom vaiables efine on the same obability
More information1 Explicit Explore or Exploit (E 3 ) Algorithm
2.997 Decision-Making in Lage-Scale Systems Mach 3 MIT, Sping 2004 Handout #2 Lectue Note 9 Explicit Exploe o Exploit (E 3 ) Algoithm Last lectue, we studied the Q-leaning algoithm: [ ] Q t+ (x t, a t
More informationLecture 28: Convergence of Random Variables and Related Theorems
EE50: Pobability Foundations fo Electical Enginees July-Novembe 205 Lectue 28: Convegence of Random Vaiables and Related Theoems Lectue:. Kishna Jagannathan Scibe: Gopal, Sudhasan, Ajay, Swamy, Kolla An
More informationNew problems in universal algebraic geometry illustrated by boolean equations
New poblems in univesal algebaic geomety illustated by boolean equations axiv:1611.00152v2 [math.ra] 25 Nov 2016 Atem N. Shevlyakov Novembe 28, 2016 Abstact We discuss new poblems in univesal algebaic
More informationCSCE 478/878 Lecture 4: Experimental Design and Analysis. Stephen Scott. 3 Building a tree on the training set Introduction. Outline.
In Homewok, you ae (supposedly) Choosing a data set 2 Extacting a test set of size > 3 3 Building a tee on the taining set 4 Testing on the test set 5 Repoting the accuacy (Adapted fom Ethem Alpaydin and
More informationRandom Variables and Probability Distribution Random Variable
Random Vaiables and Pobability Distibution Random Vaiable Random vaiable: If S is the sample space P(S) is the powe set of the sample space, P is the pobability of the function then (S, P(S), P) is called
More information16 Modeling a Language by a Markov Process
K. Pommeening, Language Statistics 80 16 Modeling a Language by a Makov Pocess Fo deiving theoetical esults a common model of language is the intepetation of texts as esults of Makov pocesses. This model
More information4/18/2005. Statistical Learning Theory
Statistical Leaning Theoy Statistical Leaning Theoy A model of supevised leaning consists of: a Envionment - Supplying a vecto x with a fixed but unknown pdf F x (x b Teache. It povides a desied esponse
More informationUnobserved Correlation in Ascending Auctions: Example And Extensions
Unobseved Coelation in Ascending Auctions: Example And Extensions Daniel Quint Univesity of Wisconsin Novembe 2009 Intoduction In pivate-value ascending auctions, the winning bidde s willingness to pay
More information763620SS STATISTICAL PHYSICS Solutions 2 Autumn 2012
763620SS STATISTICAL PHYSICS Solutions 2 Autumn 2012 1. Continuous Random Walk Conside a continuous one-dimensional andom walk. Let w(s i ds i be the pobability that the length of the i th displacement
More informationLET a random variable x follows the two - parameter
INTERNATIONAL JOURNAL OF MATHEMATICS AND SCIENTIFIC COMPUTING ISSN: 2231-5330, VOL. 5, NO. 1, 2015 19 Shinkage Bayesian Appoach in Item - Failue Gamma Data In Pesence of Pio Point Guess Value Gyan Pakash
More informationPROBLEM SET #1 SOLUTIONS by Robert A. DiStasio Jr.
POBLM S # SOLUIONS by obet A. DiStasio J. Q. he Bon-Oppenheime appoximation is the standad way of appoximating the gound state of a molecula system. Wite down the conditions that detemine the tonic and
More informationQuasi-Randomness and the Distribution of Copies of a Fixed Graph
Quasi-Randomness and the Distibution of Copies of a Fixed Gaph Asaf Shapia Abstact We show that if a gaph G has the popety that all subsets of vetices of size n/4 contain the coect numbe of tiangles one
More informationA Crash Course in (2 2) Matrices
A Cash Couse in ( ) Matices Seveal weeks woth of matix algeba in an hou (Relax, we will only stuy the simplest case, that of matices) Review topics: What is a matix (pl matices)? A matix is a ectangula
More informationFall 2014 Randomized Algorithms Oct 8, Lecture 3
Fall 204 Randomized Algoithms Oct 8, 204 Lectue 3 Pof. Fiedich Eisenband Scibes: Floian Tamè In this lectue we will be concened with linea pogamming, in paticula Clakson s Las Vegas algoithm []. The main
More informationCentral Coverage Bayes Prediction Intervals for the Generalized Pareto Distribution
Statistics Reseach Lettes Vol. Iss., Novembe Cental Coveage Bayes Pediction Intevals fo the Genealized Paeto Distibution Gyan Pakash Depatment of Community Medicine S. N. Medical College, Aga, U. P., India
More informationA Multivariate Normal Law for Turing s Formulae
A Multivaiate Nomal Law fo Tuing s Fomulae Zhiyi Zhang Depatment of Mathematics and Statistics Univesity of Noth Caolina at Chalotte Chalotte, NC 28223 Abstact This pape establishes a sufficient condition
More informationWeighted least-squares estimators of parametric functions of the regression coefficients under a general linear model
Ann Inst Stat Math (2010) 62:929 941 DOI 10.1007/s10463-008-0199-8 Weighted least-squaes estimatos of paametic functions of the egession coefficients unde a geneal linea model Yongge Tian Received: 9 Januay
More informationInformation Retrieval Advanced IR models. Luca Bondi
Advanced IR models Luca Bondi Advanced IR models 2 (LSI) Pobabilistic Latent Semantic Analysis (plsa) Vecto Space Model 3 Stating point: Vecto Space Model Documents and queies epesented as vectos in the
More informationTemporal-Difference Learning
.997 Decision-Making in Lage-Scale Systems Mach 17 MIT, Sping 004 Handout #17 Lectue Note 13 1 Tempoal-Diffeence Leaning We now conside the poblem of computing an appopiate paamete, so that, given an appoximation
More informationFailure Probability of 2-within-Consecutive-(2, 2)-out-of-(n, m): F System for Special Values of m
Jounal of Mathematics and Statistics 5 (): 0-4, 009 ISSN 549-3644 009 Science Publications Failue Pobability of -within-consecutive-(, )-out-of-(n, m): F System fo Special Values of m E.M.E.. Sayed Depatment
More informationChapter 3: Theory of Modular Arithmetic 38
Chapte 3: Theoy of Modula Aithmetic 38 Section D Chinese Remainde Theoem By the end of this section you will be able to pove the Chinese Remainde Theoem apply this theoem to solve simultaneous linea conguences
More informationRigid Body Dynamics 2. CSE169: Computer Animation Instructor: Steve Rotenberg UCSD, Winter 2018
Rigid Body Dynamics 2 CSE169: Compute Animation nstucto: Steve Rotenbeg UCSD, Winte 2018 Coss Poduct & Hat Opeato Deivative of a Rotating Vecto Let s say that vecto is otating aound the oigin, maintaining
More informationPhysics 505 Homework No. 9 Solutions S9-1
Physics 505 Homewok No 9 s S9-1 1 As pomised, hee is the tick fo summing the matix elements fo the Stak effect fo the gound state of the hydogen atom Recall, we need to calculate the coection to the gound
More informationAuchmuty High School Mathematics Department Advanced Higher Notes Teacher Version
The Binomial Theoem Factoials Auchmuty High School Mathematics Depatment The calculations,, 6 etc. often appea in mathematics. They ae called factoials and have been given the notation n!. e.g. 6! 6!!!!!
More informationSOME GENERAL NUMERICAL RADIUS INEQUALITIES FOR THE OFF-DIAGONAL PARTS OF 2 2 OPERATOR MATRICES
italian jounal of pue and applied mathematics n. 35 015 (433 44) 433 SOME GENERAL NUMERICAL RADIUS INEQUALITIES FOR THE OFF-DIAGONAL PARTS OF OPERATOR MATRICES Watheq Bani-Domi Depatment of Mathematics
More information(a) Unde zeo-bias conditions, thee ae no lled states on one side of the junction which ae at the same enegy as the empty allowed states on the othe si
1 Esaki Diode hen the concentation of impuity atoms in a pn-diode is vey high, the depletion laye width is educed to about 1 nm. Classically, a caie must have an enegy at least equal to the potential-baie
More informationChem 453/544 Fall /08/03. Exam #1 Solutions
Chem 453/544 Fall 3 /8/3 Exam # Solutions. ( points) Use the genealized compessibility diagam povided on the last page to estimate ove what ange of pessues A at oom tempeatue confoms to the ideal gas law
More informationPhysics Tutorial V1 2D Vectors
Physics Tutoial V1 2D Vectos 1 Resolving Vectos & Addition of Vectos A vecto quantity has both magnitude and diection. Thee ae two ways commonly used to mathematically descibe a vecto. y (a) The pola fom:,
More information15.081J/6.251J Introduction to Mathematical Programming. Lecture 6: The Simplex Method II
15081J/6251J Intoduction to Mathematical Pogamming ectue 6: The Simplex Method II 1 Outline Revised Simplex method Slide 1 The full tableau implementation Anticycling 2 Revised Simplex Initial data: A,
More informationIntroduction Common Divisors. Discrete Mathematics Andrei Bulatov
Intoduction Common Divisos Discete Mathematics Andei Bulatov Discete Mathematics Common Divisos 3- Pevious Lectue Integes Division, popeties of divisibility The division algoithm Repesentation of numbes
More information2. Electrostatics. Dr. Rakhesh Singh Kshetrimayum 8/11/ Electromagnetic Field Theory by R. S. Kshetrimayum
2. Electostatics D. Rakhesh Singh Kshetimayum 1 2.1 Intoduction In this chapte, we will study how to find the electostatic fields fo vaious cases? fo symmetic known chage distibution fo un-symmetic known
More informationEM Boundary Value Problems
EM Bounday Value Poblems 10/ 9 11/ By Ilekta chistidi & Lee, Seung-Hyun A. Geneal Desciption : Maxwell Equations & Loentz Foce We want to find the equations of motion of chaged paticles. The way to do
More informationTHE NUMBER OF TWO CONSECUTIVE SUCCESSES IN A HOPPE-PÓLYA URN
TH NUMBR OF TWO CONSCUTIV SUCCSSS IN A HOPP-PÓLYA URN LARS HOLST Depatment of Mathematics, Royal Institute of Technology S 100 44 Stocholm, Sweden -mail: lholst@math.th.se Novembe 27, 2007 Abstact In a
More informationWeb-based Supplementary Materials for. Controlling False Discoveries in Multidimensional Directional Decisions, with
Web-based Supplementay Mateials fo Contolling False Discoveies in Multidimensional Diectional Decisions, with Applications to Gene Expession Data on Odeed Categoies Wenge Guo Biostatistics Banch, National
More informationPearson s Chi-Square Test Modifications for Comparison of Unweighted and Weighted Histograms and Two Weighted Histograms
Peason s Chi-Squae Test Modifications fo Compaison of Unweighted and Weighted Histogams and Two Weighted Histogams Univesity of Akueyi, Bogi, v/noduslód, IS-6 Akueyi, Iceland E-mail: nikolai@unak.is Two
More informationTHE CONE THEOREM JOEL A. TROPP. Abstract. We prove a fixed point theorem for functions which are positive with respect to a cone in a Banach space.
THE ONE THEOEM JOEL A. TOPP Abstact. We pove a fixed point theoem fo functions which ae positive with espect to a cone in a Banach space. 1. Definitions Definition 1. Let X be a eal Banach space. A subset
More informationWhen two numbers are written as the product of their prime factors, they are in factored form.
10 1 Study Guide Pages 420 425 Factos Because 3 4 12, we say that 3 and 4 ae factos of 12. In othe wods, factos ae the numbes you multiply to get a poduct. Since 2 6 12, 2 and 6 ae also factos of 12. The
More informationAbsorption Rate into a Small Sphere for a Diffusing Particle Confined in a Large Sphere
Applied Mathematics, 06, 7, 709-70 Published Online Apil 06 in SciRes. http://www.scip.og/jounal/am http://dx.doi.og/0.46/am.06.77065 Absoption Rate into a Small Sphee fo a Diffusing Paticle Confined in
More information8 Separation of Variables in Other Coordinate Systems
8 Sepaation of Vaiables in Othe Coodinate Systems Fo the method of sepaation of vaiables to succeed you need to be able to expess the poblem at hand in a coodinate system in which the physical boundaies
More informationInternet Appendix for A Bayesian Approach to Real Options: The Case of Distinguishing Between Temporary and Permanent Shocks
Intenet Appendix fo A Bayesian Appoach to Real Options: The Case of Distinguishing Between Tempoay and Pemanent Shocks Steven R. Genadie Gaduate School of Business, Stanfod Univesity Andey Malenko Gaduate
More informationBASIC ALGEBRA OF VECTORS
Fomulae Fo u Vecto Algeba By Mi Mohammed Abbas II PCMB 'A' Impotant Tems, Definitions & Fomulae 01 Vecto - Basic Intoduction: A quantity having magnitude as well as the diection is called vecto It is denoted
More informationRelated Rates - the Basics
Related Rates - the Basics In this section we exploe the way we can use deivatives to find the velocity at which things ae changing ove time. Up to now we have been finding the deivative to compae the
More information1. Review of Probability.
1. Review of Pobability. What is pobability? Pefom an expeiment. The esult is not pedictable. One of finitely many possibilities R 1, R 2,, R k can occu. Some ae pehaps moe likely than othes. We assign
More informationDivisibility. c = bf = (ae)f = a(ef) EXAMPLE: Since 7 56 and , the Theorem above tells us that
Divisibility DEFINITION: If a and b ae integes with a 0, we say that a divides b if thee is an intege c such that b = ac. If a divides b, we also say that a is a diviso o facto of b. NOTATION: d n means
More informationMAC Module 12 Eigenvalues and Eigenvectors
MAC 23 Module 2 Eigenvalues and Eigenvectos Leaning Objectives Upon completing this module, you should be able to:. Solve the eigenvalue poblem by finding the eigenvalues and the coesponding eigenvectos
More informationQIP Course 10: Quantum Factorization Algorithm (Part 3)
QIP Couse 10: Quantum Factoization Algoithm (Pat 3 Ryutaoh Matsumoto Nagoya Univesity, Japan Send you comments to yutaoh.matsumoto@nagoya-u.jp Septembe 2018 @ Tokyo Tech. Matsumoto (Nagoya U. QIP Couse
More informationConvergence Dynamics of Resource-Homogeneous Congestion Games: Technical Report
1 Convegence Dynamics of Resouce-Homogeneous Congestion Games: Technical Repot Richad Southwell and Jianwei Huang Abstact Many esouce shaing scenaios can be modeled using congestion games A nice popety
More informationOn the ratio of maximum and minimum degree in maximal intersecting families
On the atio of maximum and minimum degee in maximal intesecting families Zoltán Lóánt Nagy Lale Özkahya Balázs Patkós Máté Vize Mach 6, 013 Abstact To study how balanced o unbalanced a maximal intesecting
More informationEFFECTS OF FRINGING FIELDS ON SINGLE PARTICLE DYNAMICS. M. Bassetti and C. Biscari INFN-LNF, CP 13, Frascati (RM), Italy
Fascati Physics Seies Vol. X (998), pp. 47-54 4 th Advanced ICFA Beam Dynamics Wokshop, Fascati, Oct. -5, 997 EFFECTS OF FRININ FIELDS ON SINLE PARTICLE DYNAMICS M. Bassetti and C. Biscai INFN-LNF, CP
More informationn 1 Cov(X,Y)= ( X i- X )( Y i-y ). N-1 i=1 * If variable X and variable Y tend to increase together, then c(x,y) > 0
Covaiance and Peason Coelation Vatanian, SW 540 Both covaiance and coelation indicate the elationship between two (o moe) vaiables. Neithe the covaiance o coelation give the slope between the X and Y vaiable,
More informationMATH 220: SECOND ORDER CONSTANT COEFFICIENT PDE. We consider second order constant coefficient scalar linear PDEs on R n. These have the form
MATH 220: SECOND ORDER CONSTANT COEFFICIENT PDE ANDRAS VASY We conside second ode constant coefficient scala linea PDEs on R n. These have the fom Lu = f L = a ij xi xj + b i xi + c i whee a ij b i and
More informationSolution to HW 3, Ma 1a Fall 2016
Solution to HW 3, Ma a Fall 206 Section 2. Execise 2: Let C be a subset of the eal numbes consisting of those eal numbes x having the popety that evey digit in the decimal expansion of x is, 3, 5, o 7.
More informationq i i=1 p i ln p i Another measure, which proves a useful benchmark in our analysis, is the chi squared divergence of p, q, which is defined by
CSISZÁR f DIVERGENCE, OSTROWSKI S INEQUALITY AND MUTUAL INFORMATION S. S. DRAGOMIR, V. GLUŠČEVIĆ, AND C. E. M. PEARCE Abstact. The Ostowski integal inequality fo an absolutely continuous function is used
More information1.2 Differential cross section
.2. DIFFERENTIAL CROSS SECTION Febuay 9, 205 Lectue VIII.2 Diffeential coss section We found that the solution to the Schodinge equation has the fom e ik x ψ 2π 3/2 fk, k + e ik x and that fk, k = 2 m
More informationChapter 5 Linear Equations: Basic Theory and Practice
Chapte 5 inea Equations: Basic Theoy and actice In this chapte and the next, we ae inteested in the linea algebaic equation AX = b, (5-1) whee A is an m n matix, X is an n 1 vecto to be solved fo, and
More informationOn a quantity that is analogous to potential and a theorem that relates to it
Su une quantité analogue au potential et su un théoème y elatif C R Acad Sci 7 (87) 34-39 On a quantity that is analogous to potential and a theoem that elates to it By R CLAUSIUS Tanslated by D H Delphenich
More informationAST 121S: The origin and evolution of the Universe. Introduction to Mathematical Handout 1
Please ead this fist... AST S: The oigin and evolution of the Univese Intoduction to Mathematical Handout This is an unusually long hand-out and one which uses in places mathematics that you may not be
More informationMultiple Experts with Binary Features
Multiple Expets with Binay Featues Ye Jin & Lingen Zhang Decembe 9, 2010 1 Intoduction Ou intuition fo the poect comes fom the pape Supevised Leaning fom Multiple Expets: Whom to tust when eveyone lies
More informationBoundary Layers and Singular Perturbation Lectures 16 and 17 Boundary Layers and Singular Perturbation. x% 0 Ω0æ + Kx% 1 Ω0æ + ` : 0. (9.
Lectues 16 and 17 Bounday Layes and Singula Petubation A Regula Petubation In some physical poblems, the solution is dependent on a paamete K. When the paamete K is vey small, it is natual to expect that
More information18.06 Problem Set 4 Solution
8.6 Poblem Set 4 Solution Total: points Section 3.5. Poblem 2: (Recommended) Find the lagest possible numbe of independent vectos among ) ) ) v = v 4 = v 5 = v 6 = v 2 = v 3 =. Solution (4 points): Since
More informationControl Chart Analysis of E k /M/1 Queueing Model
Intenational OPEN ACCESS Jounal Of Moden Engineeing Reseach (IJMER Contol Chat Analysis of E /M/1 Queueing Model T.Poongodi 1, D. (Ms. S. Muthulashmi 1, (Assistant Pofesso, Faculty of Engineeing, Pofesso,
More informationAPPENDIX. For the 2 lectures of Claude Cohen-Tannoudji on Atom-Atom Interactions in Ultracold Quantum Gases
APPENDIX Fo the lectues of Claude Cohen-Tannoudji on Atom-Atom Inteactions in Ultacold Quantum Gases Pupose of this Appendix Demonstate the othonomalization elation(ϕ ϕ = δ k k δ δ )k - The wave function
More informationProbablistically Checkable Proofs
Lectue 12 Pobablistically Checkable Poofs May 13, 2004 Lectue: Paul Beame Notes: Chis Re 12.1 Pobablisitically Checkable Poofs Oveview We know that IP = PSPACE. This means thee is an inteactive potocol
More informationGENLOG Multinomial Loglinear and Logit Models
GENLOG Multinomial Loglinea and Logit Models Notation This chapte descibes the algoithms used to calculate maximum-likelihood estimates fo the multinomial loglinea model and the multinomial logit model.
More informationworking pages for Paul Richards class notes; do not copy or circulate without permission from PGR 2004/11/3 10:50
woking pages fo Paul Richads class notes; do not copy o ciculate without pemission fom PGR 2004/11/3 10:50 CHAPTER7 Solid angle, 3D integals, Gauss s Theoem, and a Delta Function We define the solid angle,
More informationarxiv: v1 [math.co] 4 May 2017
On The Numbe Of Unlabeled Bipatite Gaphs Abdullah Atmaca and A Yavuz Ouç axiv:7050800v [mathco] 4 May 207 Abstact This pape solves a poblem that was stated by M A Haison in 973 [] This poblem, that has
More informationOn the Quasi-inverse of a Non-square Matrix: An Infinite Solution
Applied Mathematical Sciences, Vol 11, 2017, no 27, 1337-1351 HIKARI Ltd, wwwm-hikaicom https://doiog/1012988/ams20177273 On the Quasi-invese of a Non-squae Matix: An Infinite Solution Ruben D Codeo J
More informationMethod for Approximating Irrational Numbers
Method fo Appoximating Iational Numbes Eic Reichwein Depatment of Physics Univesity of Califonia, Santa Cuz June 6, 0 Abstact I will put foth an algoithm fo poducing inceasingly accuate ational appoximations
More informationAsymptotically Lacunary Statistical Equivalent Sequence Spaces Defined by Ideal Convergence and an Orlicz Function
"Science Stays Tue Hee" Jounal of Mathematics and Statistical Science, 335-35 Science Signpost Publishing Asymptotically Lacunay Statistical Equivalent Sequence Spaces Defined by Ideal Convegence and an
More informationNuclear Medicine Physics 02 Oct. 2007
Nuclea Medicine Physics Oct. 7 Counting Statistics and Eo Popagation Nuclea Medicine Physics Lectues Imaging Reseach Laboatoy, Radiology Dept. Lay MacDonald 1//7 Statistics (Summaized in One Slide) Type
More informationJournal of Inequalities in Pure and Applied Mathematics
Jounal of Inequalities in Pue and Applied Mathematics COEFFICIENT INEQUALITY FOR A FUNCTION WHOSE DERIVATIVE HAS A POSITIVE REAL PART S. ABRAMOVICH, M. KLARIČIĆ BAKULA AND S. BANIĆ Depatment of Mathematics
More informationMODULE 5a and 5b (Stewart, Sections 12.2, 12.3) INTRO: In MATH 1114 vectors were written either as rows (a1, a2,..., an) or as columns a 1 a. ...
MODULE 5a and 5b (Stewat, Sections 2.2, 2.3) INTRO: In MATH 4 vectos wee witten eithe as ows (a, a2,..., an) o as columns a a 2... a n and the set of all such vectos of fixed length n was called the vecto
More informationSUPPLEMENTARY MATERIAL CHAPTER 7 A (2 ) B. a x + bx + c dx
SUPPLEMENTARY MATERIAL 613 7.6.3 CHAPTER 7 ( px + q) a x + bx + c dx. We choose constants A and B such that d px + q A ( ax + bx + c) + B dx A(ax + b) + B Compaing the coefficients of x and the constant
More informationPhysics 2A Chapter 10 - Moment of Inertia Fall 2018
Physics Chapte 0 - oment of netia Fall 08 The moment of inetia of a otating object is a measue of its otational inetia in the same way that the mass of an object is a measue of its inetia fo linea motion.
More information1D2G - Numerical solution of the neutron diffusion equation
DG - Numeical solution of the neuton diffusion equation Y. Danon Daft: /6/09 Oveview A simple numeical solution of the neuton diffusion equation in one dimension and two enegy goups was implemented. Both
More informationEstimation of entropy rate and Rényi entropy rate for Markov chains
06 IEEE Intenational Symposium on Infomation Theoy Estimation of entopy ate and Rényi entopy ate fo Makov chains Sudeep Kamath, Segio Vedú Depatment of Electical Engineeing Pinceton Univesity e-mail: sukamath,
More informationBerkeley Math Circle AIME Preparation March 5, 2013
Algeba Toolkit Rules of Thumb. Make sue that you can pove all fomulas you use. This is even bette than memoizing the fomulas. Although it is best to memoize, as well. Stive fo elegant, economical methods.
More informationVector d is a linear vector function of vector d when the following relationships hold:
Appendix 4 Dyadic Analysis DEFINITION ecto d is a linea vecto function of vecto d when the following elationships hold: d x = a xxd x + a xy d y + a xz d z d y = a yxd x + a yy d y + a yz d z d z = a zxd
More informationHua Xu 3 and Hiroaki Mukaidani 33. The University of Tsukuba, Otsuka. Hiroshima City University, 3-4-1, Ozuka-Higashi
he inea Quadatic Dynamic Game fo Discete-ime Descipto Systems Hua Xu 3 and Hioai Muaidani 33 3 Gaduate School of Systems Management he Univesity of suuba, 3-9- Otsua Bunyo-u, oyo -0, Japan xuhua@gssm.otsua.tsuuba.ac.jp
More informationEncapsulation theory: the transformation equations of absolute information hiding.
1 Encapsulation theoy: the tansfomation equations of absolute infomation hiding. Edmund Kiwan * www.edmundkiwan.com Abstact This pape descibes how the potential coupling of a set vaies as the set is tansfomed,
More information1 Notes on Order Statistics
1 Notes on Ode Statistics Fo X a andom vecto in R n with distibution F, and π S n, define X π by and F π by X π (X π(1),..., X π(n) ) F π (x 1,..., x n ) F (x π 1 (1),..., x π 1 (n)); then the distibution
More information221B Lecture Notes Scattering Theory I
Why Scatteing? B Lectue Notes Scatteing Theoy I Scatteing of paticles off taget has been one of the most impotant applications of quantum mechanics. It is pobably the most effective way to study the stuctue
More informationGradient-based Neural Network for Online Solution of Lyapunov Matrix Equation with Li Activation Function
Intenational Confeence on Infomation echnology and Management Innovation (ICIMI 05) Gadient-based Neual Netwok fo Online Solution of Lyapunov Matix Equation with Li Activation unction Shiheng Wang, Shidong
More informationTHE JEU DE TAQUIN ON THE SHIFTED RIM HOOK TABLEAUX. Jaejin Lee
Koean J. Math. 23 (2015), No. 3, pp. 427 438 http://dx.doi.og/10.11568/kjm.2015.23.3.427 THE JEU DE TAQUIN ON THE SHIFTED RIM HOOK TABLEAUX Jaejin Lee Abstact. The Schensted algoithm fist descibed by Robinson
More informationHOW TO TEACH THE FUNDAMENTALS OF INFORMATION SCIENCE, CODING, DECODING AND NUMBER SYSTEMS?
6th INTERNATIONAL MULTIDISCIPLINARY CONFERENCE HOW TO TEACH THE FUNDAMENTALS OF INFORMATION SCIENCE, CODING, DECODING AND NUMBER SYSTEMS? Cecília Sitkuné Göömbei College of Nyíegyháza Hungay Abstact: The
More information