An efficient dictionary learning algorithm for sparse representation
|
|
- Elisabeth Dennis
- 6 years ago
- Views:
Transcription
1 An efficient dictionary learning algorithm for pare repreentation Leyuan ang 1 and Shutao Li 1 1.College of Electrical and Information Engineering, Hunan Univerity, Changha, 41008, China fangleyuan@gmail.com, hutao_li@yahoo.com.cn Abtract Spare and redundant repreentation of data aume an ability to decribe ignal a linear combination of a few atom from a dictionary. If the model of the ignal i unknown, the dictionary can be learned from a et of training ignal. Like the K-SVD, many of the practical dictionary learning algorithm are compoed of two main part pare-coding and dictionary-update. hi paper firt propoe a Stagewie leat angle regreion (St-LARS) method for performing the pare-coding operation. he St-LARS applie a hard-threholding trategy into the original leat angle regreion (LARS) algorithm, which enable it to elect many atom at each iteration and thu reult in fat olution while till provide good reult. hen, a dictionary update method named approximated ingular value decompoition (ASVD) i ued on the dictionary update tage. It i a quick approximation of the exact SVD computation and can reduce the complexity of it. Experiment on both ynthetic data and 3-D image denoiing demontrate the advantage of the propoed algorithm over other dictionary learning method not only in term of better trained dictionary but alo in term of computation time. Key Word Dictionary learning, pare repreentation, leat angle regreion, hard threholding 1 INRODUCION Spare and redundant repreentation of a ignal y N K over a dictionary D (with K column denoted a atom) refer that one can find a linear combination of a few atom from D that i cloe to the ignal y. reviou work [1-4] have hown that modeling a ignal with uch a pare decompoition i very effective in many ignal and image proceing application. A fundamental conideration in employing the above pare model i the choice of the dictionary D. he majority of work on thi topic can be parted into two main categorie analytical-baed and learning-baed. he analytical approache contruct the dictionary, baing on variou type of wavelet [5], and it variant [6-7]. he learning approache recommend uing machine learning technique to infer the dictionary from a et of training example [8-10]. Advantage of thee approache are the finer dictionarie they produce compared to analytical approache, and the ignificantly better performance in application. However, the complexity contraint in thee algorithm often limit the ize of the dictionarie that can be trained, and the dimenion of ignal that can be proceed. So, reducing the complexity of thee learning algorithm i a tiff challenge for u. A we know, the main proce of many learning algorithm [8-10] can be divided into two tage pare coding and dictionary update. Spare coding i to find the paret olution of the training ignal, which dominate the complexity of the dictionary learning. Commonly ued trategie conducting the pare coding are typically baed on greedy puruit and convex relaxation. Greedy puruit employ ome greedy algorithm (e.g. the matching puruit (M) [11] and orthogonal matching puruit (OM) [1]) to get an approximate olution of the pare repreentation. Since the M and OM are that only one atom i elected at each N iteration, Donoho et al. have recently propoed the tagewie orthogonal matching puruit (StOM) [13] by applying a threhold trategy named fale dicover rate (DR) control into the atom election proce. So, the StOM can elect more than one atom for each iteration and thu ha low computational complexity than the M and OM, epecially for large-cale problem. However, the high non-convexity of thee greedy algorithm uually can not find the optimal olution and thi will enable the dictionary learning algorithm to get caught in local minimal or even addle point [, 14]. he convex relaxation approache (e.g. the bai puruit (B) [4] and the LASSO [15]) ue the 1 -norm a a parene meaure and have been proven to obtain the paret olution. But they run much lower than the above greedy algorithm, and thu will create heavy computational burden to the whole learning algorithm. A fat algorithm called leat angle regreion (LARS) introduced in [16] can make a mall modification to olve the LASSO problem and the computational complexity of it i very cloe to that of the greedy method. But, the LARS alo jut allow one atom to be choen in the atom election proce, and therefore there i a trong incentive to elect more atom for each iteration in order to accelerate the convergence, a in [13]. ollowing thi idea, we propoe a method named St-LARS by applying a hard-threholding trategy into the LARS. hi can enable it to elect more than one atom at each iteration while till keep good performance a the LARS. It i worthwhile to note that for the hard-threholding i le time conuming than the threholding controlled by the DR, the St-LARS generally can run much fater than the StOM, a will be demontrated in the experimental part. A a reult, the St-LARS greatly accelerate the tage of the pare coding and give a good pare olution for the dictionary update tage. Updating the dictionary, when the pare repreentation i found, i comparatively eaier. he SVD decompoition ued /10/$ IEEE
2 in the K-SVD ha been hown to be a wier method than many other approache [8-10, 14, 17]. In thi paper, we replace the exact SVD computation with a impler approximated one (ASVD) [18] and thu obtain further acceleration for the dictionary learning algorithm. he ret of thi paper i organized a follow. In Section, we introduce the propoed dictionary learning algorithm. Our experimental reult on both ynthetic data and 3-D image denoiing are preented in ection 3. Section 4 conclude thi paper and ugget future work. DICIONARY LEARNING Dictionary learning i the tak of learning or training a dictionary uch that it i well adapted to it training data. Uually the objective i to give pare repreentation of the training et, making the total error a mall a poible, i.e. minimizing the um of quared error. Let the training data contitute the column in a matrix Y and the pare coefficient vector are the column in matrix X. he objective function of the dictionary learning can be tated formally a a minimization problem min Y DX ubject to X, (1) D,X where the function denote the -norm. A practical optimization trategy can be found by plitting the problem into two part which are alternately olved within an iterative loop [, 8-10]. he two part are 1) Spare coding keeping D fixed, find X; ) Dictionary update keeping X fixed, find D. Since the propoed dictionary learning algorithm i alo baed on the two part, the following ubection will give the detailed decription of our improvement on thee part..1 Spare coding Conider olving the optimization problem (1) with -norm penalty over the pare matrix X while keeping the dictionary D fixed. hi problem can be olved by optimizing over each vector x of the pare matrix individually min y x ubjectto x t, x D () where the y repreent one ignal of the training matrix Y. Notice that the above optimization tak can be eaily changed to be 1 min y D x + λ x. (3) x or a proper choice of λ, the two problem are equivalent. If the p in the function i et to 0, the above problem i known to be N-hard in general [19] and ome greedy algorithm [11-13] can not get the optimal pare olution of it. he B [4] and LASSO [15] replace the combinatorial 0 -norm with the 1 -norm. A reported in [4], they achieve parer olution but are lower than the greedy approache for mot experiment. In [16], an algorithm called LARS i introduced to olve the LASSO problem with minor modification and it computational complexity i linear with the ize of input ignal a the greedy algorithm. However, the LARS compute the olution by only adding one atom to it active et at each iteration. herefore, we adapt the idea from [13], and propoe the St-LARS algorithm to olve the problem (3). y ˆx c I dx xi ig. 1. he cheme of the St-LARS algorithm. he cheme of St-LARS i preented in ig. 1, which conit of the following even tep. Step 1 Set initial olution x 0 = 0, initial reidual r=y, 0 initial etimate I 0 = φ, initial threhold λ 0 = D y and counter =1. Step Apply the D to the current reidual r -1, getting a vector of reidual correlation c c ( k) = dk r -1, k = 1,..., K, (4) for each correponding atom ( K i the number of atom in D, d k denote the k-th atom in D, and c ( k ) tand for the k-th element in c ). he reidual correlation c are uppoed to contain a mall fraction of ignificant non-zero. Step 3 erform hard threholding with threhold λ to find the ignificant non-zero in c I = { k c( k) > λ}. (5) he λ i calculated by λ = μλ 1, where the μ i a threhold decreae tep. r
3 Step 4 Compute the update direction dx by projecting the reidual r onto ubpace panned by the column of D belonging to I dx (I ) = ( D D ) D r. (6) 1 I I I Step 5 Update the olution x by augmenting x 1 in the direction dx with a tep ize ε (typically choen equally to the threhold decreae tep) x = x 1 + εdx. (7) Step 6 Contruct a new D x uing the x. hen, the current reidual can be calculated by r = y D x. (8) Step 7 Check the topping condition. he procedure top with the output of final olution ˆx when the -norm of current reidual r reache an error goal, or when the threhold λ i le than a pre-choen value. If it i not atified, we et = + 1 and go to the tep. hi algorithm i a fat tagewie approximation to LARS and provide a good approximation to the olution of (3). It i till note that the computation in the tep 4 will uually not be carried out explicitly for it high computation cot. So, we employ a progreive Choleky update proce to reduce the work involved in the matrix inverion, and thu further accelerate the St-LARS. or more detail about the Choleky proce, the reader can be found in [18, 0].. Dictionary update Given fixed pare matrix X, the dictionary D can be updated by olving the following problem min Y DX. (9) D In general, thi problem can be olved uing gradient decent with iterative projection [17] or other leat quare method [8]. However, it can be much more efficiently olved uing the SVD decompoition. or a given atom k, the quadratic term in (9) i rewritten a Y d g d g = E d g, (10) j k j j k k k k k where the g j are the row of pare matrix X, the d k denote the atom of the dictionary D and the E k tand for the reidual matrix. After the SVD decompoition of the matrix E k, both the atom d k and g k can be updated. o avoid the introduction of new non-zero in X, the update ue only the ignal vector whoe current repreentation ue the atom d k. Since the exact SVD computation i time conuming, adopting a much quicker approach to get a olution of (10) i a more exciting option. herefore, our implementation here ue the ASVD, which employ a ingle iteration of alternate-optimization over the atom d k and the pare matrix row g k, which i given by d k= Ekg k/ Ekg k, (11) g k = ( Ek) d k. hi operation ha been proven to both reduce the complexity and give very cloe reult to the full decompoition of E k [18]. 3 EXERIMENS We evaluate the propoed method with ynthetic and real data. Uing ynthetic data with random dictionarie help u to examine the ability of the propoed method to recover dictionarie exactly (to within an acceptable quared error). hi ynthetic experiment i very imilar to that in [10], except for the dimenion of the data and the dictionary we generate are lightly higher. o tet the performance on real data, we chooe the 3-D C volume, which need a comparatively larger dictionary. So thee experiment are to demontrate that our method i more applicable to large-cale dictionary learning problem than other approache. 3.1 Synthetic experiment In thee imulation, a dictionary D i firt generated by normalizing a matrix with i.i.d. uniform random entrie. hen, we produce 1500 data ignal of dimenion 50, each created by a linear combination of ten different dictionary atom, with independent location and uniformly ditributed i.i.d. coefficient in random. After that, white Gauian noie with varying ignal-to-noie ratio (SNR) i added to the reulting data ignal. he training dictionary i initialized with data ignal and the number of the training iteration in thi etting i et to 100. If the quared error between the learnt and the true dictionary element i below 0., it i claified a correctly recovery. o allow a fair comparion, the imulation are repeated five time and the average value are calculated. In ig., the dictionary recovery ucce rate of our method are compared with that of the K-SVD [10], StOM-ASVD, and LARS-ASVD. he StOM-ASVD and LARS-ASVD are the method which firt apply the StOM and LARS to perform the pare coding repectively, and then employ the ASVD to conduct the dictionary update. he topping criteria for the LARS in LARS-ASVD and St-LARS in our method are λ = 0.3 in (3). or the OM in the K-SVD and StOM in the StOM-ASVD, they are topped when the fixed number of ten coefficient are found. he threhold decreae tep μ and the tep ize ε in (7) are both choen to A can be een from the ig., the LARS-ASVD and our method recover nearly the ame number of atom and perform better than the other two method for all the teted cae. In ig. 3, we alo compare the computation time of the four algorithm for the above imulation. he imulation are done in the environment of an AMD Athlon CU.81 GHz with a.00 GB RAM C, operating under Matlab A we can ee, the peed advantage of our method i obviou and it i about two time fater than the other approache. It i till notice that the StOM-ASVD can not run very fat in thee
4 imulation and even lower than the K-SVD, though the StOM can alo elect many atom at each iteration on the pare coding tage. he main reaon for thi i that the DR In thi ubection, our method i teted on two 3-D C volume Viible Male-Head and Viible emale-ankle. Like the K-SVD denoiing algorithm in [1], we firt train an over-complete dictionary uing block from the noiy image, and then denoie the image uing thi dictionary. he peak-ignal-to-noie ratio (NSR) i ued a objective denoiing meaure. he ize of the training block i and the ize of the dictionary i he number of the training block i choen to and the number of the training iteration i et to 5. We hould mention that ome newly denoiing technique, e.g. multi-cale framework [] and non-local imultaneou pare-coding [1], can be ued here to further improve our performance. However, our work here only focue on the original denoiing formulation for implicity, and thu do not conduct a thorough comparion with ome of the tate-of-the-art work. ig.. Comparion of the dictionary recovery ucce rate uing different dictionary learning method (a) (b) ig. 3. Comparion of the computation cot of the dictionary learning method control which calculate the threhold for each iteration of the StOM conume large computation, and the computational complexity of it i far larger than the hard-threholding ued in our method. We gue that if the StOM applie in larger dictionary or higher dimenional ignal, the StOM-ASVD will become fater than the K-SVD D image denoiing experiment able 1 C denoiing reult uing the K-SVD, and our method. Value repreent SNR (db), and are averaged over 5 execution. he bet reult in thi table are labeled in bold. Noie level Vi.. Ankle Vi. M. Head σ K-SVD Our method K-SVD Our method (c) (d) ig. 4. Denoiing reult for Viible emale-ankle. Image are mainly provided for qualitative evaluation. (a) Original image ( σ = 50); (b) Noiy image; (c) Denoied uing the K-SVD; (d) Denoied uing our method. he denoiing reult are ummarized in able 1. We can ee from the table that the SNR gain of our method over the K-SVD i conitent, though not very large. Some actual denoiing reult are hown in ig. 4. It can be een that the viual quality of the K-SVD i very cloe to that of our method. Although the difference in viual quality are typically mall, the main appeal of our method i it ubtantially better complexity, a depicted in able. A can be een, the complexity advantage of our method tranlate to a 4 reduction in denoiing time compared to the K-SVD. So, it can be eaily concluded that our method i more uitable for large-cale dictionary learning problem.
5 able Running time (in econd) of the K-SVD, and our method for the reult in able 1. Method/σ Vi.. Ankle Vi. M. Head K-SVD Our method CONCLUSION In thi paper, we combine the St-LARS and ASVD into an efficient dictionary learning algorithm. he St-LARS give a better olution for the pare coding and greatly reduce the complexity of it by imply exploiting a hard-threholding trategy. he ASVD i a quick approximation way for updating dictionary, and thu further accelerate the whole dictionary learning algorithm. he experimental reult demontrate the uperior performance of the propoed method in term of training better dictionary and reducing the computational complexity. So thi learning algorithm i very uitable for ome large-cale ignal proceing application (like the 3-D C denoiing above). However, the full potential of thi algorithm i needed to be further explored, and other ignal proceing tak (involving color image denoiing, deblurring, and inpainting) are expected to benefit from it. ACKNOWLEDGEMENS hi paper i upported by the National Natural Science oundation of China (No ), the h.d. rogram oundation of Minitry of Education of China (No ), the Key roject of Chinee Minitry of Education (009-10), and the Open roject rogram of National Laboratory of attern Recognition, China. REERENCES [1] M. Elad and M. Aharon. Image denoiing via pare and redundant repreentation over learned dictionarie. IEEE ran. Image proce., 15(1) , 006. [] J. Mairal, G. Sapiro, and M. Elad. Learning multicale pare repreentation for image and video retoration. SIAM J. Multicale Model. Simul., 7(1) 14-41, 008. [3] J. Mairal,. Bach, J. once, G. Sapiro, and A. Zierman. Non-local pare model for image retoration. roc. ICCV 009, okyo, Japan, 009, pp [4] S. S. Chen, D. L. Donoho, and M. A. Saunder. Atomic decompoition by bai puruit. SIAM Rev., 43(1) , 001. [5] S. Mallat. A wavelet tour of ignal proceing, third edition. Academic re, 009. [6] M. N. Do and M. Vetterli. he contourlet tranform an efficient directional multireolution image repreentation. IEEE ran. Image roce., 14(1) , 005. [7] D. L. Donoho. Wedgelet nearly minimax etimation of edge. he Annal of Statitic, 7(3) , [8] K. Engan, S. O. Aae, and J. Hakon Huoy. Method of optimal direction for frame deign. roc. ICASS 1999, Wahington, USA, 1999, Vol.5, pp [9] S. Leage, R. Gribonval,. Bimbot, and L. Benaroya. Learning union of orthonormal bae with threhold threholded ingular value decompoition. roc. ICASS 005, hiladelphia, USA, 005, Vol.5, pp [10] M. Aharon, M. Elad, and A. M. Brucktein. he K-SVD An algorithm for deigning of overcomplete dictionarie for pare repreentation. IEEE ran. Signal roce., 54(11) , 006. [11] S. Mallat and Z. Zhang. Matching puruit with time-frequency dictionarie. IEEE ran. Signal roce., 41(1) , [1] Y. C. ati, R. Rezaiifar, and. S. Krihnapraad. Orthogonal matching puruit Recurive function approximation with application to wavelet decompoition. roc. AilomarSS, California, USA, 1993, Vol.1, pp [13] D. L. Donoho, Y. aig, I. Drori, and J.-L. Starck. Spare olution of underdetermined linear equation by tagewie orthogonal matching puruit. 006 [Online]. Available http//tat.tanford.edu/~idrori/stom.pdf, reprint. [14] R. Rubintein, A. M. Brucktein, and M. Elad. Dictionarie for pare repreentation modeling. roceeding of the IEEE, 98(6) , 010. [15] R. ibhirani. Regreion hrinkage and election via the lao. J. Royal. Statit. Soc B., 58(1) 67-88, [16] B. Efron,. Hatie, I. Johnton, and R. ibhirani. Leat angle regreion. Ann. Statit., 3() , 004. [17] Y. Cenor and S. A. Zenio. arallel optimization theory, Algorithm and Application, Oxford Univerity re, [18] R. Rubintein, M. Zibulevky, and M. Elad. Efficient implementation of the K-SVD algorithm uing batch orthogonal matching puruit. echnical Report-CS echnion, 008. [19] G. Davi, S. Mallat, and M. Avellaneda. Adaptive greedy approximation. Contructive approximation, 13(1) 57-58, [0]. Blumenath and M. Davie. Gradient puruit. IEEE ran Signal roce., 56(6) , 008. [1] K. Dabov, A. oi, V. Katkovnik, and K. Egiazarian. Image denoiing by pare 3-D tranform-domain collaborative filtering. IEEE ran. Image roce., 16(8) , 007.
Efficient Methods of Doppler Processing for Coexisting Land and Weather Clutter
Efficient Method of Doppler Proceing for Coexiting Land and Weather Clutter Ça gatay Candan and A Özgür Yılmaz Middle Eat Technical Univerity METU) Ankara, Turkey ccandan@metuedutr, aoyilmaz@metuedutr
More informationWhite Rose Research Online URL for this paper: Version: Accepted Version
Thi i a repoitory copy of Identification of nonlinear ytem with non-peritent excitation uing an iterative forward orthogonal leat quare regreion algorithm. White Roe Reearch Online URL for thi paper: http://eprint.whiteroe.ac.uk/107314/
More informationON THE APPROXIMATION ERROR IN HIGH DIMENSIONAL MODEL REPRESENTATION. Xiaoqun Wang
Proceeding of the 2008 Winter Simulation Conference S. J. Maon, R. R. Hill, L. Mönch, O. Roe, T. Jefferon, J. W. Fowler ed. ON THE APPROXIMATION ERROR IN HIGH DIMENSIONAL MODEL REPRESENTATION Xiaoqun Wang
More informationGroup Sparse Coding with a Laplacian Scale Mixture Prior
Group Spare Coding with a Laplacian Scale Mixture Prior Pierre J. Garrigue IQ Engine, Inc. Berkeley, CA 94704 pierre.garrigue@gmail.com Bruno A. Olhauen Helen Will Neurocience Intitute School of Optometry
More informationSocial Studies 201 Notes for March 18, 2005
1 Social Studie 201 Note for March 18, 2005 Etimation of a mean, mall ample ize Section 8.4, p. 501. When a reearcher ha only a mall ample ize available, the central limit theorem doe not apply to the
More informationSuggested Answers To Exercises. estimates variability in a sampling distribution of random means. About 68% of means fall
Beyond Significance Teting ( nd Edition), Rex B. Kline Suggeted Anwer To Exercie Chapter. The tatitic meaure variability among core at the cae level. In a normal ditribution, about 68% of the core fall
More informationReal Sources (Secondary Sources) Phantom Source (Primary source) LS P. h rl. h rr. h ll. h lr. h pl. h pr
Ecient frequency domain ltered-x realization of phantom ource iet C.W. ommen, Ronald M. Aart, Alexander W.M. Mathijen, John Gara, Haiyan He Abtract A phantom ound ource i a virtual ound image which can
More informationSocial Studies 201 Notes for November 14, 2003
1 Social Studie 201 Note for November 14, 2003 Etimation of a mean, mall ample ize Section 8.4, p. 501. When a reearcher ha only a mall ample ize available, the central limit theorem doe not apply to the
More informationμ + = σ = D 4 σ = D 3 σ = σ = All units in parts (a) and (b) are in V. (1) x chart: Center = μ = 0.75 UCL =
Our online Tutor are available 4*7 to provide Help with Proce control ytem Homework/Aignment or a long term Graduate/Undergraduate Proce control ytem Project. Our Tutor being experienced and proficient
More informationMolecular Dynamics Simulations of Nonequilibrium Effects Associated with Thermally Activated Exothermic Reactions
Original Paper orma, 5, 9 7, Molecular Dynamic Simulation of Nonequilibrium Effect ociated with Thermally ctivated Exothermic Reaction Jerzy GORECKI and Joanna Natalia GORECK Intitute of Phyical Chemitry,
More informationDetermination of the local contrast of interference fringe patterns using continuous wavelet transform
Determination of the local contrat of interference fringe pattern uing continuou wavelet tranform Jong Kwang Hyok, Kim Chol Su Intitute of Optic, Department of Phyic, Kim Il Sung Univerity, Pyongyang,
More informationStochastic Optimization with Inequality Constraints Using Simultaneous Perturbations and Penalty Functions
Stochatic Optimization with Inequality Contraint Uing Simultaneou Perturbation and Penalty Function I-Jeng Wang* and Jame C. Spall** The John Hopkin Univerity Applied Phyic Laboratory 11100 John Hopkin
More informationProblem Set 8 Solutions
Deign and Analyi of Algorithm April 29, 2015 Maachuett Intitute of Technology 6.046J/18.410J Prof. Erik Demaine, Srini Devada, and Nancy Lynch Problem Set 8 Solution Problem Set 8 Solution Thi problem
More informationImage Denoising Based on Non-Local Low-Rank Dictionary Learning
Advanced cience and Technology Letter Vol.11 (AT 16) pp.85-89 http://dx.doi.org/1.1457/atl.16. Iage Denoiing Baed on Non-Local Low-Rank Dictionary Learning Zhang Bo 1 1 Electronic and Inforation Engineering
More informationFast Convolutional Sparse Coding (FCSC)
Fat Convolutional Spare Coding (FCSC) Bailey ong Department of Computer Science Univerity of California, Irvine bhkong@ic.uci.edu Charle C. Fowlke Department of Computer Science Univerity of California,
More information[Saxena, 2(9): September, 2013] ISSN: Impact Factor: INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH TECHNOLOGY
[Saena, (9): September, 0] ISSN: 77-9655 Impact Factor:.85 IJESRT INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH TECHNOLOGY Contant Stre Accelerated Life Teting Uing Rayleigh Geometric Proce
More informationPreemptive scheduling on a small number of hierarchical machines
Available online at www.ciencedirect.com Information and Computation 06 (008) 60 619 www.elevier.com/locate/ic Preemptive cheduling on a mall number of hierarchical machine György Dóa a, Leah Eptein b,
More informationAlternate Dispersion Measures in Replicated Factorial Experiments
Alternate Diperion Meaure in Replicated Factorial Experiment Neal A. Mackertich The Raytheon Company, Sudbury MA 02421 Jame C. Benneyan Northeatern Univerity, Boton MA 02115 Peter D. Krau The Raytheon
More informationIN this paper, we focus on the following composite optimization
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 1 VR-: A Simple Stochatic Variance Reduction Method for Machine Learning Fanhua Shang, Member, IEEE, Kaiwen Zhou, Hongying Liu, Jame Cheng, Ivor W. Tang,
More informationSignal Modeling: From Convolutional Sparse Coding to Convolutional Neural Networks
Signal Modeling: From Convolutional Spare Coding to Convolutional Neural Network Vardan Papyan The Computer Science Department Technion Irael Intitute of technology Haifa 32000, Irael Joint work with Jeremia
More informationA Constraint Propagation Algorithm for Determining the Stability Margin. The paper addresses the stability margin assessment for linear systems
A Contraint Propagation Algorithm for Determining the Stability Margin of Linear Parameter Circuit and Sytem Lubomir Kolev and Simona Filipova-Petrakieva Abtract The paper addree the tability margin aement
More informationProximal Iteratively Reweighted Algorithm with Multiple Splitting for Nonconvex Sparsity Optimization
Proceeding of the Twenty-Eighth AAAI Conference on Artificial Intelligence Proimal Iteratively Reweighted Algorithm with Multiple Splitting for Nonconve Sparity Optimization Canyi Lu 1, Yunchao Wei, Zhouchen
More informationEvolutionary Algorithms Based Fixed Order Robust Controller Design and Robustness Performance Analysis
Proceeding of 01 4th International Conference on Machine Learning and Computing IPCSIT vol. 5 (01) (01) IACSIT Pre, Singapore Evolutionary Algorithm Baed Fixed Order Robut Controller Deign and Robutne
More informationLecture 4 Topic 3: General linear models (GLMs), the fundamentals of the analysis of variance (ANOVA), and completely randomized designs (CRDs)
Lecture 4 Topic 3: General linear model (GLM), the fundamental of the analyi of variance (ANOVA), and completely randomized deign (CRD) The general linear model One population: An obervation i explained
More informationZ a>2 s 1n = X L - m. X L = m + Z a>2 s 1n X L = The decision rule for this one-tail test is
M09_BERE8380_12_OM_C09.QD 2/21/11 3:44 PM Page 1 9.6 The Power of a Tet 9.6 The Power of a Tet 1 Section 9.1 defined Type I and Type II error and their aociated rik. Recall that a repreent the probability
More informationCHAPTER 4 DESIGN OF STATE FEEDBACK CONTROLLERS AND STATE OBSERVERS USING REDUCED ORDER MODEL
98 CHAPTER DESIGN OF STATE FEEDBACK CONTROLLERS AND STATE OBSERVERS USING REDUCED ORDER MODEL INTRODUCTION The deign of ytem uing tate pace model for the deign i called a modern control deign and it i
More informationFinding the location of switched capacitor banks in distribution systems based on wavelet transform
UPEC00 3t Aug - 3rd Sept 00 Finding the location of witched capacitor bank in ditribution ytem baed on wavelet tranform Bahram nohad Shahid Chamran Univerity in Ahvaz bahramnohad@yahoo.com Mehrdad keramatzadeh
More informationAsymptotics of ABC. Paul Fearnhead 1, Correspondence: Abstract
Aymptotic of ABC Paul Fearnhead 1, 1 Department of Mathematic and Statitic, Lancater Univerity Correpondence: p.fearnhead@lancater.ac.uk arxiv:1706.07712v1 [tat.me] 23 Jun 2017 Abtract Thi document i due
More informationClustering Methods without Given Number of Clusters
Clutering Method without Given Number of Cluter Peng Xu, Fei Liu Introduction A we now, mean method i a very effective algorithm of clutering. It mot powerful feature i the calability and implicity. However,
More informationImproving the Efficiency of a Digital Filtering Scheme for Diabatic Initialization
1976 MONTHLY WEATHER REVIEW VOLUME 15 Improving the Efficiency of a Digital Filtering Scheme for Diabatic Initialization PETER LYNCH Met Éireann, Dublin, Ireland DOMINIQUE GIARD CNRM/GMAP, Météo-France,
More informationYoram Gat. Technical report No. 548, March Abstract. A classier is said to have good generalization ability if it performs on
A bound concerning the generalization ability of a certain cla of learning algorithm Yoram Gat Univerity of California, Berkeley Technical report No. 548, March 999 Abtract A claier i aid to have good
More informationNew Construction of Deterministic Compressed Sensing Matrices via Singular Linear Spaces over Finite Fields
New Contruction of Determinitic Compreed Sening Matrice via Singular Linear Space over Finite Field XUEMEI LIU College of Science Civil Aviation Univerity of China Tianjin300300 CHINA m-liu776@63com YINGMO
More informationAn estimation approach for autotuning of event-based PI control systems
Acta de la XXXIX Jornada de Automática, Badajoz, 5-7 de Septiembre de 08 An etimation approach for autotuning of event-baed PI control ytem Joé Sánchez Moreno, María Guinaldo Loada, Sebatián Dormido Departamento
More informationCHAPTER 8 OBSERVER BASED REDUCED ORDER CONTROLLER DESIGN FOR LARGE SCALE LINEAR DISCRETE-TIME CONTROL SYSTEMS
CHAPTER 8 OBSERVER BASED REDUCED ORDER CONTROLLER DESIGN FOR LARGE SCALE LINEAR DISCRETE-TIME CONTROL SYSTEMS 8.1 INTRODUCTION 8.2 REDUCED ORDER MODEL DESIGN FOR LINEAR DISCRETE-TIME CONTROL SYSTEMS 8.3
More informationLOW ORDER MIMO CONTROLLER DESIGN FOR AN ENGINE DISTURBANCE REJECTION PROBLEM. P.Dickinson, A.T.Shenton
LOW ORDER MIMO CONTROLLER DESIGN FOR AN ENGINE DISTURBANCE REJECTION PROBLEM P.Dickinon, A.T.Shenton Department of Engineering, The Univerity of Liverpool, Liverpool L69 3GH, UK Abtract: Thi paper compare
More informationApplication of Gradient Projection for Sparse Reconstruction to Compressed Sensing for Image Reconstruction of Electrical Capacitance Tomography
Journal of Electrical and Electronic Engineering 08; 6(): 46-5 http://www.ciencepublihinggroup.com/j/jeee doi: 0.648/j.jeee.08060. ISS: 39-63 (Print); ISS: 39-605 (Online) Application of Gradient Projection
More informationOptimal Coordination of Samples in Business Surveys
Paper preented at the ICES-III, June 8-, 007, Montreal, Quebec, Canada Optimal Coordination of Sample in Buine Survey enka Mach, Ioana Şchiopu-Kratina, Philip T Rei, Jean-Marc Fillion Statitic Canada New
More informationCodes Correcting Two Deletions
1 Code Correcting Two Deletion Ryan Gabry and Frederic Sala Spawar Sytem Center Univerity of California, Lo Angele ryan.gabry@navy.mil fredala@ucla.edu Abtract In thi work, we invetigate the problem of
More informationarxiv: v3 [eess.sp] 18 Dec 2017
Compreed Sening by Shortet-Solution Guided Decimation Mutian Shen 3, Pan Zhang, and Hai-Jun Zhou,,4 Key Laboratory for Theoretical Phyic, Intitute of Theoretical Phyic, Chinee Academy of Science, Beijing
More informationDistributed dynamic modeling and monitoring for. large-scale industrial processes under closed-loop. control
Ditributed dynamic modeling and monitoring for large-cale indutrial procee under cloed-loop control Wenqing Li 1, Chunhui Zhao 1,2*, Biao Huang 3 1. State Key Laboratory of Indutrial Control echnology,
More informationLecture 21. The Lovasz splitting-off lemma Topics in Combinatorial Optimization April 29th, 2004
18.997 Topic in Combinatorial Optimization April 29th, 2004 Lecture 21 Lecturer: Michel X. Goeman Scribe: Mohammad Mahdian 1 The Lovaz plitting-off lemma Lovaz plitting-off lemma tate the following. Theorem
More informationNeural Network Linearization of Pressure Force Sensor Transfer Characteristic
Acta Polytechnica Hungarica Vol., No., 006 Neural Network Linearization of Preure Force Senor Tranfer Characteritic Jozef Vojtko, Irena Kováčová, Ladilav Madaráz, Dobrolav Kováč Faculty of Electrical Engineering
More informationA Study on Simulating Convolutional Codes and Turbo Codes
A Study on Simulating Convolutional Code and Turbo Code Final Report By Daniel Chang July 27, 2001 Advior: Dr. P. Kinman Executive Summary Thi project include the deign of imulation of everal convolutional
More informationChapter 2 Sampling and Quantization. In order to investigate sampling and quantization, the difference between analog
Chapter Sampling and Quantization.1 Analog and Digital Signal In order to invetigate ampling and quantization, the difference between analog and digital ignal mut be undertood. Analog ignal conit of continuou
More informationUSING NONLINEAR CONTROL ALGORITHMS TO IMPROVE THE QUALITY OF SHAKING TABLE TESTS
October 12-17, 28, Beijing, China USING NONLINEAR CONTR ALGORITHMS TO IMPROVE THE QUALITY OF SHAKING TABLE TESTS T.Y. Yang 1 and A. Schellenberg 2 1 Pot Doctoral Scholar, Dept. of Civil and Env. Eng.,
More informationSingular perturbation theory
Singular perturbation theory Marc R. Rouel June 21, 2004 1 Introduction When we apply the teady-tate approximation (SSA) in chemical kinetic, we typically argue that ome of the intermediate are highly
More informationJan Purczyński, Kamila Bednarz-Okrzyńska Estimation of the shape parameter of GED distribution for a small sample size
Jan Purczyńki, Kamila Bednarz-Okrzyńka Etimation of the hape parameter of GED ditribution for a mall ample ize Folia Oeconomica Stetinenia 4()/, 35-46 04 Folia Oeconomica Stetinenia DOI: 0.478/foli-04-003
More informationComplex CORDIC-like Algorithms for Linearly Constrained MVDR Beamforming
Complex CORDIC-like Algorithm for Linearly Contrained MVDR Beamforming Mariu Otte (otte@dt.e-technik.uni-dortmund.de) Information Proceing Lab Univerity of Dortmund Otto Hahn Str. 4 44221 Dortmund, Germany
More informationControl of Delayed Integrating Processes Using Two Feedback Controllers R MS Approach
Proceeding of the 7th WSEAS International Conference on SYSTEM SCIENCE and SIMULATION in ENGINEERING (ICOSSSE '8) Control of Delayed Integrating Procee Uing Two Feedback Controller R MS Approach LIBOR
More informationOne Class of Splitting Iterative Schemes
One Cla of Splitting Iterative Scheme v Ciegi and V. Pakalnytė Vilniu Gedimina Technical Univerity Saulėtekio al. 11, 2054, Vilniu, Lithuania rc@fm.vtu.lt Abtract. Thi paper deal with the tability analyi
More informationCDMA Signature Sequences with Low Peak-to-Average-Power Ratio via Alternating Projection
CDMA Signature Sequence with Low Peak-to-Average-Power Ratio via Alternating Projection Joel A Tropp Int for Comp Engr and Sci (ICES) The Univerity of Texa at Autin 1 Univerity Station C0200 Autin, TX
More informationJul 4, 2005 turbo_code_primer Revision 0.0. Turbo Code Primer
Jul 4, 5 turbo_code_primer Reviion. Turbo Code Primer. Introduction Thi document give a quick tutorial on MAP baed turbo coder. Section develop the background theory. Section work through a imple numerical
More informationThe Use of MDL to Select among Computational Models of Cognition
The Ue of DL to Select among Computational odel of Cognition In J. yung, ark A. Pitt & Shaobo Zhang Vijay Balaubramanian Department of Pychology David Rittenhoue Laboratorie Ohio State Univerity Univerity
More informationMATEMATIK Datum: Tid: eftermiddag. A.Heintz Telefonvakt: Anders Martinsson Tel.:
MATEMATIK Datum: 20-08-25 Tid: eftermiddag GU, Chalmer Hjälpmedel: inga A.Heintz Telefonvakt: Ander Martinon Tel.: 073-07926. Löningar till tenta i ODE och matematik modellering, MMG5, MVE6. Define what
More informationarxiv: v1 [math.oc] 16 Jan 2012
A Reduced Bai Method for the Simulation of American Option Bernard Haadonk, Julien Salomon and Barbara Wohlmuth arxiv:121.3289v1 [math.oc] 16 Jan 212 Abtract We preent a reduced bai method for the imulation
More informationFactor Analysis with Poisson Output
Factor Analyi with Poion Output Gopal Santhanam Byron Yu Krihna V. Shenoy, Department of Electrical Engineering, Neurocience Program Stanford Univerity Stanford, CA 94305, USA {gopal,byronyu,henoy}@tanford.edu
More informationWavelet Analysis in EPR Spectroscopy
Vol. 108 (2005) ACTA PHYSICA POLONICA A No. 1 Proceeding of the XXI International Meeting on Radio and Microwave Spectrocopy RAMIS 2005, Poznań-Bȩdlewo, Poland, April 24 28, 2005 Wavelet Analyi in EPR
More informationResearch Article Reliability of Foundation Pile Based on Settlement and a Parameter Sensitivity Analysis
Mathematical Problem in Engineering Volume 2016, Article ID 1659549, 7 page http://dxdoiorg/101155/2016/1659549 Reearch Article Reliability of Foundation Pile Baed on Settlement and a Parameter Senitivity
More informationAPPLICATION OF THE SINGLE IMPACT MICROINDENTATION FOR NON- DESTRUCTIVE TESTING OF THE FRACTURE TOUGHNESS OF NONMETALLIC AND POLYMERIC MATERIALS
APPLICATION OF THE SINGLE IMPACT MICROINDENTATION FOR NON- DESTRUCTIVE TESTING OF THE FRACTURE TOUGHNESS OF NONMETALLIC AND POLYMERIC MATERIALS REN A. P. INSTITUTE OF APPLIED PHYSICS OF THE NATIONAL ACADEMY
More informationCHEAP CONTROL PERFORMANCE LIMITATIONS OF INPUT CONSTRAINED LINEAR SYSTEMS
Copyright 22 IFAC 5th Triennial World Congre, Barcelona, Spain CHEAP CONTROL PERFORMANCE LIMITATIONS OF INPUT CONSTRAINED LINEAR SYSTEMS Tritan Pérez Graham C. Goodwin Maria M. Serón Department of Electrical
More informationCompact finite-difference approximations for anisotropic image smoothing and painting
CWP-593 Compact finite-difference approximation for aniotropic image moothing and painting Dave Hale Center for Wave Phenomena, Colorado School of Mine, Golden CO 80401, USA ABSTRACT Finite-difference
More informationConvex Optimization-Based Rotation Parameter Estimation Using Micro-Doppler
Journal of Electrical Engineering 4 (6) 57-64 doi:.765/8-/6.4. D DAVID PUBLISHING Convex Optimization-Baed Rotation Parameter Etimation Uing Micro-Doppler Kyungwoo Yoo, Joohwan Chun, Seungoh Yoo and Chungho
More informationLINEAR system identification has been the subject of
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL., NO., 1 Set-Memberhip error-in-variable identification through convex relaxation technique Vito Cerone, Member, IEEE, Dario Piga, Member, IEEE, and Diego Regruto,
More informationSource slideplayer.com/fundamentals of Analytical Chemistry, F.J. Holler, S.R.Crouch. Chapter 6: Random Errors in Chemical Analysis
Source lideplayer.com/fundamental of Analytical Chemitry, F.J. Holler, S.R.Crouch Chapter 6: Random Error in Chemical Analyi Random error are preent in every meaurement no matter how careful the experimenter.
More informationAnnex-A: RTTOV9 Cloud validation
RTTOV-91 Science and Validation Plan Annex-A: RTTOV9 Cloud validation Author O Embury C J Merchant The Univerity of Edinburgh Intitute for Atmo. & Environ. Science Crew Building King Building Edinburgh
More information7.2 INVERSE TRANSFORMS AND TRANSFORMS OF DERIVATIVES 281
72 INVERSE TRANSFORMS AND TRANSFORMS OF DERIVATIVES 28 and i 2 Show how Euler formula (page 33) can then be ued to deduce the reult a ( a) 2 b 2 {e at co bt} {e at in bt} b ( a) 2 b 2 5 Under what condition
More informationDesign By Emulation (Indirect Method)
Deign By Emulation (Indirect Method he baic trategy here i, that Given a continuou tranfer function, it i required to find the bet dicrete equivalent uch that the ignal produced by paing an input ignal
More informationA Bluffer s Guide to... Sphericity
A Bluffer Guide to Sphericity Andy Field Univerity of Suex The ue of repeated meaure, where the ame ubject are teted under a number of condition, ha numerou practical and tatitical benefit. For one thing
More informationEFFICIENTLY implemented active set methods have
An Efficient Active Set Method for SVM raining without Singular Inner Problem Chritopher Sentelle, Georgio C. Anagnotopoulo and Michael Georgiopoulo Abtract Efficiently implemented active et method have
More informationinto a discrete time function. Recall that the table of Laplace/z-transforms is constructed by (i) selecting to get
Lecture 25 Introduction to Some Matlab c2d Code in Relation to Sampled Sytem here are many way to convert a continuou time function, { h( t) ; t [0, )} into a dicrete time function { h ( k) ; k {0,,, }}
More informationThe Hassenpflug Matrix Tensor Notation
The Haenpflug Matrix Tenor Notation D.N.J. El Dept of Mech Mechatron Eng Univ of Stellenboch, South Africa e-mail: dnjel@un.ac.za 2009/09/01 Abtract Thi i a ample document to illutrate the typeetting of
More informationAdvanced Digital Signal Processing. Stationary/nonstationary signals. Time-Frequency Analysis... Some nonstationary signals. Time-Frequency Analysis
Advanced Digital ignal Proceing Prof. Nizamettin AYDIN naydin@yildiz.edu.tr Time-Frequency Analyi http://www.yildiz.edu.tr/~naydin 2 tationary/nontationary ignal Time-Frequency Analyi Fourier Tranform
More informationTHE SPLITTING SUBSPACE CONJECTURE
THE SPLITTING SUBSPAE ONJETURE ERI HEN AND DENNIS TSENG Abtract We anwer a uetion by Niederreiter concerning the enumeration of a cla of ubpace of finite dimenional vector pace over finite field by proving
More informationSTOCHASTIC GENERALIZED TRANSPORTATION PROBLEM WITH DISCRETE DISTRIBUTION OF DEMAND
OPERATIONS RESEARCH AND DECISIONS No. 4 203 DOI: 0.5277/ord30402 Marcin ANHOLCER STOCHASTIC GENERALIZED TRANSPORTATION PROBLEM WITH DISCRETE DISTRIBUTION OF DEMAND The generalized tranportation problem
More informationGain and Phase Margins Based Delay Dependent Stability Analysis of Two- Area LFC System with Communication Delays
Gain and Phae Margin Baed Delay Dependent Stability Analyi of Two- Area LFC Sytem with Communication Delay Şahin Sönmez and Saffet Ayaun Department of Electrical Engineering, Niğde Ömer Halidemir Univerity,
More informationLearning Multiplicative Interactions
CSC2535 2011 Lecture 6a Learning Multiplicative Interaction Geoffrey Hinton Two different meaning of multiplicative If we take two denity model and multiply together their probability ditribution at each
More informationEstimation of Peaked Densities Over the Interval [0,1] Using Two-Sided Power Distribution: Application to Lottery Experiments
MPRA Munich Peronal RePEc Archive Etimation of Peaed Denitie Over the Interval [0] Uing Two-Sided Power Ditribution: Application to Lottery Experiment Krzyztof Konte Artal Invetment 8. April 00 Online
More informationCHAPTER 6. Estimation
CHAPTER 6 Etimation Definition. Statitical inference i the procedure by which we reach a concluion about a population on the bai of information contained in a ample drawn from that population. Definition.
More informationTRIPLE SOLUTIONS FOR THE ONE-DIMENSIONAL
GLASNIK MATEMATIČKI Vol. 38583, 73 84 TRIPLE SOLUTIONS FOR THE ONE-DIMENSIONAL p-laplacian Haihen Lü, Donal O Regan and Ravi P. Agarwal Academy of Mathematic and Sytem Science, Beijing, China, National
More informationConfusion matrices. True / False positives / negatives. INF 4300 Classification III Anne Solberg The agenda today: E.g., testing for cancer
INF 4300 Claification III Anne Solberg 29.10.14 The agenda today: More on etimating claifier accuracy Cure of dimenionality knn-claification K-mean clutering x i feature vector for pixel i i- The cla label
More informationDIFFERENTIAL EQUATIONS
DIFFERENTIAL EQUATIONS Laplace Tranform Paul Dawkin Table of Content Preface... Laplace Tranform... Introduction... The Definition... 5 Laplace Tranform... 9 Invere Laplace Tranform... Step Function...4
More informationSolutions. Digital Control Systems ( ) 120 minutes examination time + 15 minutes reading time at the beginning of the exam
BSc - Sample Examination Digital Control Sytem (5-588-) Prof. L. Guzzella Solution Exam Duration: Number of Quetion: Rating: Permitted aid: minute examination time + 5 minute reading time at the beginning
More informationOptimization model in Input output analysis and computable general. equilibrium by using multiple criteria non-linear programming.
Optimization model in Input output analyi and computable general equilibrium by uing multiple criteria non-linear programming Jing He * Intitute of ytem cience, cademy of Mathematic and ytem cience Chinee
More informationA Simple Approach to Synthesizing Naïve Quantized Control for Reference Tracking
A Simple Approach to Syntheizing Naïve Quantized Control for Reference Tracking SHIANG-HUA YU Department of Electrical Engineering National Sun Yat-Sen Univerity 70 Lien-Hai Road, Kaohiung 804 TAIAN Abtract:
More informationEstimating floor acceleration in nonlinear multi-story moment-resisting frames
Etimating floor acceleration in nonlinear multi-tory moment-reiting frame R. Karami Mohammadi Aitant Profeor, Civil Engineering Department, K.N.Tooi Univerity M. Mohammadi M.Sc. Student, Civil Engineering
More informationOn the Observability of a Linear System with a Sparse Initial State
1 On the Obervability of a Linear Sytem with a Spare Initial State Geethu Joeph and Chandra R Murthy Senior Member IEEE Abtract In thi paper we addre the problem of obervability of a linear dynamic ytem
More informationSimple Observer Based Synchronization of Lorenz System with Parametric Uncertainty
IOSR Journal of Electrical and Electronic Engineering (IOSR-JEEE) ISSN: 78-676Volume, Iue 6 (Nov. - Dec. 0), PP 4-0 Simple Oberver Baed Synchronization of Lorenz Sytem with Parametric Uncertainty Manih
More informationAn Efficient Statistical Method for Image Noise Level Estimation
An Efficient Statitical Method for Image Noie Level Etimation Guangyong Chen 1, Fengyuan Zhu 1, and Pheng Ann Heng 1,2 1 Department of Computer Science and Engineering, The Chinee Univerity of Hong Kong
More informationGiven the following circuit with unknown initial capacitor voltage v(0): X(s) Immediately, we know that the transfer function H(s) is
EE 4G Note: Chapter 6 Intructor: Cheung More about ZSR and ZIR. Finding unknown initial condition: Given the following circuit with unknown initial capacitor voltage v0: F v0/ / Input xt 0Ω Output yt -
More informationSMALL-SIGNAL STABILITY ASSESSMENT OF THE EUROPEAN POWER SYSTEM BASED ON ADVANCED NEURAL NETWORK METHOD
SMALL-SIGNAL STABILITY ASSESSMENT OF THE EUROPEAN POWER SYSTEM BASED ON ADVANCED NEURAL NETWORK METHOD S.P. Teeuwen, I. Erlich U. Bachmann Univerity of Duiburg, Germany Department of Electrical Power Sytem
More informationA RECONFIGURABLE MARS CONSTELLATION FOR RADIO OCCULTATION MEASUREMENTS AND NAVIGATION
A RECONFIGURABLE MARS CONSTELLATION FOR RADIO OCCULTATION MEASUREMENTS AND NAVIGATION Iabelle Nann, Dario Izzo, Roger Walker ESA Advanced Concept Team, ESTEC (DG-X) Keplerlaan 1, 2201 AZ Noordwijk ZH,
More informationPOWER SYSTEM SMALL SIGNAL STABILITY ANALYSIS BASED ON TEST SIGNAL
POWE YEM MALL INAL ABILIY ANALYI BAE ON E INAL Zheng Xu, Wei hao, Changchun Zhou Zheang Univerity, Hangzhou, 37 PChina Email: hvdc@ceezueducn Abtract - In thi paper, a method baed on ome tet ignal (et
More informationComparing Means: t-tests for Two Independent Samples
Comparing ean: t-tet for Two Independent Sample Independent-eaure Deign t-tet for Two Independent Sample Allow reearcher to evaluate the mean difference between two population uing data from two eparate
More informationLecture 9: Shor s Algorithm
Quantum Computation (CMU 8-859BB, Fall 05) Lecture 9: Shor Algorithm October 7, 05 Lecturer: Ryan O Donnell Scribe: Sidhanth Mohanty Overview Let u recall the period finding problem that wa et up a a function
More informationColorado School of Mines. Computer Vision. Professor William Hoff Dept of Electrical Engineering &Computer Science.
Proeor William Ho Dept o Electrical Engineering &Computer Science http://inide.mine.edu/~who/ Uncertaint Uncertaint Let a that we have computed a reult (uch a poe o an object), rom image data How do we
More informationEstimation of Current Population Variance in Two Successive Occasions
ISSN 684-8403 Journal of Statitic Volume 7, 00, pp. 54-65 Etimation of Current Population Variance in Two Succeive Occaion Abtract Muhammad Azam, Qamruz Zaman, Salahuddin 3 and Javed Shabbir 4 The problem
More informationLateral vibration of footbridges under crowd-loading: Continuous crowd modeling approach
ateral vibration of footbridge under crowd-loading: Continuou crowd modeling approach Joanna Bodgi, a, Silvano Erlicher,b and Pierre Argoul,c Intitut NAVIER, ENPC, 6 et 8 av. B. Pacal, Cité Decarte, Champ
More informationPIPELINED DIVISION OF SIGNED NUMBERS WITH THE USE OF RESIDUE ARITHMETIC FOR SMALL NUMBER RANGE WITH THE PROGRAMMABLE GATE ARRAY
POZNAN UNIVE RSITY OF TE CHNOLOGY ACADE MIC JOURNALS No 76 Electrical Engineering 03 Robert SMYK* Zenon ULMAN* Maciej CZYŻAK* PIPELINED DIVISION OF SIGNED NUMBERS WITH THE USE OF RESIDUE ARITHMETIC FOR
More informationAN ADAPTIVE SIGNAL SEARCH ALGORITHM IN GPS RECEIVER
N PTIVE SIGNL SERH LGORITHM IN GPS REEIVER Item Type text; Proceeding uthor Li, Sun; Yinfeng, Wang; Qihan, Zhang Publiher International Foundation for Telemetering Journal International Telemetering onference
More informationPhysicsAndMathsTutor.com
1. A teacher wihe to tet whether playing background muic enable tudent to complete a tak more quickly. The ame tak wa completed by 15 tudent, divided at random into two group. The firt group had background
More information