Maximally informative dimensions: Analyzing neural responses to natural signals.

Size: px
Start display at page:

Download "Maximally informative dimensions: Analyzing neural responses to natural signals."

Transcription

1 Maximally infrmative dimensins: Analyzing neural respnses t natural signals. Tatyana Sharpee, Nicle C. Rust, and William Bialek Slan Swartz Center fr Theretical Neurbilgy, Department f Physilgy University f Califrnia at San Francisc, San Francisc, Califrnia Center fr Neural Science, New Yrk University, New Yrk, NY Department f Physics, Princetn University, Princetn, New Jersey sharpee@phy.ucsf.edu, rust@cns.nyu.edu, wbialek@princetn.edu We prpse a methd that allws fr a rigrus statistical analysis f neural respnses t natural stimuli, which are nn-gaussian and exhibit strng crrelatins. We have in mind a mdel in which neurns are selective fr a small number f stimulus dimensins ut f the high dimensinal stimulus space, but within this subspace the respnses can be arbitrarily nnlinear. Therefre we maximize the mutual infrmatin between the sequence f elicited neural respnses and an ensemble f stimuli that has been prjected n trial directins in the stimulus space. The prcedure can be dne iteratively by increasing the number f directins with respect t which infrmatin is maximized. Thse directins that allw the recvery f all f the infrmatin between spikes and the full unprjected stimuli describe the relevant subspace. If the dimensinality f the relevant subspace indeed is much smaller than that f the verall stimulus space, it may becme experimentally feasible t map ut the neurn s input-utput functin even under fully natural stimulus cnditins. This cntrasts with methds based n crrelatins functins (reverse crrelatin, spike-triggered cvariance,...) which all require simplified stimulus statistics if we are t use them rigrusly. 1 Intrductin Frm lfactin t visin and auditin, there is an increasing need, and a grwing number f experiments [1]-[8] that study respnses f sensry neurns t natural stimuli. Natural stimuli have specific statistical prperties [9, 10], and therefre sample nly a subspace f all pssible spatial and tempral frequencies explred during stimulatin with white nise. Observing the full dynamic range f neural respnses may require using stimulus ensembles which apprximate thse ccurring in nature, and it is an attractive hypthesis that the neural representatin f these natural signals may be ptimized in sme way. Finally, sme neurn respnses are strngly nnlinear and adaptive, and may nt be predicted frm a cmbinatin f respnses t simple stimuli. It has als been shwn that the variability in neural respnse decreases substantially when dynamical, rather than static, stimuli are used [11, 1]. Fr all these reasns, it wuld be attractive t have a rigrus methd f analyzing neural respnses t cmplex, naturalistic inputs. The stimuli analyzed by sensry neurns are intrinsically high-dimensinal, with dimen-

2 > sins. Fr example, in the case f visual neurns, input is specified as light intensity n a grid f at least pixels. The dimensinality increases further if the time dependence is t be explred as well. Full explratin f such a large parameter space is beynd the cnstraints f experimental data cllectin. Hwever, prgress can be made prvided we make certain assumptins abut hw the respnse has been generated. In the simplest mdel, the prbability f respnse can be described by ne receptive field (RF) [13]. The receptive field can be thught f as a special directin in the stimulus space such that the neurn s respnse depends nly n a prjectin f a given stimulus nt. This special directin is the ne fund by the reverse crrelatin methd [13, 14]. In a mre general case, the prbability f the respnse depends n prjectins "!#, $ &('' '*), f the stimulus n a set f vectrs +,, -'' '.0/ : :&; = ' ' &:&; where is the prbability f a spike given a stimulus : and is the average firing rate. In what fllws we will call the subspace spanned by the set f vectrs +, / the relevant subspace (RS). Even thugh the ideas develped belw can be used t analyze input-utput functins with respect t different neural respnses, we settle n a single spike as the respnse f interest. ) Eq. (1) in itself is nt yet a simplificatin if the dimensinality f the RS is equal t the dimensinality A f the stimulus space. )CB In this paper we will use the idea f dimensinality reductin [15, 16] and assume that A. The input-utput functin > in Eq. (1) can be strngly nnlinear, but it is presumed t depend nly n a small number f prjectins. This assumptin appears t be less stringent than that f apprximate linearity which ne makes when characterizing neurn s respnse in terms f Wiener kernels. )EDF The mst difficult part in recnstructing the input-utput functin is t find the RS. Fr, a descriptin in (1) / is just as valid, since we did nt make any terms f any linear cmbinatin f vectrs +, assumptins as t a particular frm f nnlinear functin >. We might hwever prefer ne crdinate system ver anther if it, fr example, leads t sparser prbability distributins r mre statistically independent variables. Once the relevant subspace is knwn, the prbability :G; # becmes a functin f nly few parameters, and it becmes feasible t map this functin experimentally, inverting the prbability distributins accrding t Bayes rule: +# /H 13 +# / ; &: 13 +# / If stimuli are crrelated Gaussian nise, then the neural respnse can be characterized by the spike-triggered cvariance methd [15, 16]. It can be shwn that the dimensinality f the RS is equal t the number f nn-zer eigenvalues f a matrix given by a difference between cvariance matrices f all presented stimuli and stimuli cnditinal n a spike. Mrever, the RS is spanned by the eigenvectrs assciated with the nn-zer eigenvalues multiplied by the inverse f the a priri cvariance matrix. Cmpared t the reverse crrelatin methd, we are n lnger limited t finding nly ne f the relevant directins. Hwever because f the necessity t prbe a tw-pint crrelatin functin, the spiketriggered cvariance methd requires better sampling f distributins f inputs cnditinal n a spike. In this paper we investigate whether it is pssible t lift the requirement fr stimuli t be Gaussian. When using natural stimuli, which are certainly nn-gaussian, the RS cannt be fund by the spike-triggered cvariance methd. Similarly, the reverse crrelatin methd des nt give the crrect RF, even in the simplest case where the input-utput functin (1) depends nly n ne prjectin. Hwever, vectrs that span the RS are clearly special directins in the stimulus space. This ntin can be quantified by Shannn infrmatin, and ()

3 e X e g ' an ptimizatin prblem can be frmulated t find the RS. Therefre the current implementatin f the dimensinality reductin idea is cmplimentary t the clustering f stimuli dne in the infrmatin bttleneck methd [17]; see als Ref. 13 [18]. Nn infrmatin based measures f similarity between prbability distributins # 13 and ; 4?6I79&: have als been prpsed [19]. We illustrate hw the ptimizatin scheme f maximizing infrmatin as functin f directin in the stimulus space wrks with natural stimuli fr mdel rientatin sensitive cells with ne and tw relevant directins, much like simple and cmplex cells fund in primary visual crtex. It is als pssible t estimate average errrs in the recnstructin. The advantage f this ptimizatin scheme is that it des nt rely n any specific statistical prperties f the stimulus ensemble, and can be used with natural stimuli. Infrmatin as an bjective functin When analyzing neural respnses, we cmpare the a priri prbability distributin f all presented stimuli with the prbability distributin f stimuli which lead t a spike. Fr Gaussian signals, the prbability distributin can be characterized by its secnd mment, the cvariance matrix. Hwever, an ensemble f natural stimuli is nt Gaussian, s that neither secnd nr any ther finite number f mments is sufficient t describe the prbability distributin. In this situatin, the Shannn infrmatin prvides a cnvenient way f cmparing tw prbability distributins. The average infrmatin carried by the arrival time f ne spike is given by [0] J K5L@M N0O QP&R 13 ; 4?687 9: GS TU WV 13 ; 4?6I79&: 0X 13 #ZY ; 4?687 9: The infrmatin per spike, as written in (3) is difficult t estimate experimentally, 13 since it requires either sampling f the high-dimensinal prbability distributin r a mdel f hw spikes were generated, J K5L@M N0O i.e. the knwledge f lw-dimensinal RS. Hwever it is pssible t calculate in a mdel-independent :G; way, if stimuli are presented multiple times t estimate the prbability distributin #. Then, J K[LM N*O ]\ 13^4?6I79&:_; 13^4?6I79&: &:_; S TU ` : badc where the average is taken ver all presented stimuli. Nte that fr a finite dataset f e repetitins, J K5L@M N0O the btained Kji5M kml@n M value K[L@M N*Op ef J K[L@M N*O hg will be Kji5M n kml@n Maverage larger than, with difference S q, where e is the number f different stimuli, and e K5L@M N0O J K[LM N0O is the number f elicited spikes [1] acrss all f the repetitins. The true value can als be fund by extraplating t esr []. The knwledge f the ttal infrmatin per spike will characterize the quality f the recnstructin f the neurn s input-utput relatin. Having in mind a mdel in which spikes are generated accrding t prjectin nt a lwdimensinal subspace, we start by prjecting all f the presented stimuli n a particular directin 1utv5wm; 4?6I79&: in the stimulus space, and frm prbability distributins xzy5{ [w! ; : R 1utv5w, Wzy5{ 5w! R. The infrmatin Jp x~} w,1mtv[wm; &: GS TU V 1utv5wm; 4?6I79&: 0X 1utG5w ZY prvides an invariant measure f hw much the ccurrence f a spike is determined by prjectin n the directin. It is a functin nly f directin in the stimulus space and des nt change when vectr is multiplied by a cnstant. This can be seen by nting that fr any prbability distributin and any cnstant 1m ZtG[w, W # 1utv5w X. When evaluated J, alng any vectr, ƒ J K[LM N0O J K[L@M N*O. The ttal infrmatin can be recvered alng ne particular directin nly if, and the RS is ne-dimensinal. R (3) (4) (5)

4 + ) / { > Y + ' { J, By analgy with (5), ne culd als calculate infrmatin '' ' alng a set f several directins + '' ' / based n the multi-pint prbability distributins: 1mt ˆ Š Š Š t0œ w,ž ; 4?6I79&: W y 5w,Ž Ž _! Ž ; &: R 1mt ˆ Š Š Š t*œ, wpž /x y [wpž Ž _! Ž R ' ) If we are successful in finding all f the directins in the input-utput relatin (1), then J K[LM N0O the infrmatin evaluated alng the fund set will be equal t the ttal infrmatin. When we calculate infrmatin alng J K[LM a N*O set f vectrs that are slightly ff frm the RS, the answer is, f curse, smaller than and is quadratic in deviatins { Ž. One can therefre find the RS by maximizing infrmatin with respect t vectrs simultaneusly. The infrmatin des nt increase if mre vectrs utside the RS are included int the calculatin. On the ther hand, the result f ptimizatin with respect t the number f vectrs may deviate frm the RS if stimuli are crrelated. The deviatin is als 13^4?687 '' prprtinal t a weighted average f ' '.. Fr uncrrelated stimuli, any vectr r Jp Jp a set f vectrs that maximizes belngs t the RS. T find the RS, we first maximize J K[L@M N*O, and cmpare this maximum with, which is estimated accrding t (4). If the difference exceeds that expected frm finite sampling crrectins, we increment the number f directins with respect t which infrmatin is simultaneusly maximized. J, The infrmatin as defined by (5) is a cntinuus functin, whse gradient can be cmputed t J w,1 } t [w V y5 ; w : y^ ; wp 1mtv[w"; : w 1utG[w (6) tj š Since infrmatin des nt change with the length f the vectr, (which can als be seen frm (6) directly), unnecessary evaluatins f infrmatin fr multiples f are avided by maximizing alng the gradient. As an ptimizatin algrithm, we have used a cmbinatin f gradient ascent and simulated annealing algrithms: successive line maximizatins were dne alng the directin f the gradient. During line maximizatins, a pint with a smaller value f infrmatin was accepted accrding t Bltzmann statistics, :@œv6 V ^J, Ž Jp Ž with prbability?0xžy. The effective temperature T is reduced upn cmpletin f each line maximizatin. 3 Discussin We tested the scheme f lking fr the mst infrmative directins n mdel neurns that respnd t stimuli derived frm natural scenes. As stimuli we used patches f digitized t 8-bit scale phts, in which n crrectins were made fr camera s light intensity transfrmatin functin. Our gal is t demnstrate that even thugh spatial crrelatins present in natural scenes are nn-gaussian, they can be successfully remved frm the estimate f vectrs defining the RS. 3.1 Simple Cell Our first example is taken t mimic prperties f simple cells fund in the primary visual crtex. A mdel phase and rientatin sensitive cell has a single relevant directin shwn in Fig. 1(a). A given frame leads t a spike if prjectin Ÿ!, reaches a threshld value in the presence f nise: &:&; &: F x y^ (7)

5 Figure 1: Analysis f a mdel simple cell with RF shwn in (a). The spike-triggered is shwn in (b). Panel (c) shws an attempt t remve crrelatins accrd- ; (d) vectr kh? fund by maximizing average Kji5 ing t reverse crrelatin methd, vª@«ž «Ž &:&; K[i5 infrmatin; (e) The prbability f a spike!h ± (crsses) is cmpared t :&; 8' used in generating spikes (slid line). Parameters ²~ kh 0 khm ³ I' and kh 0 khm ³ [ kh 0 and khm ³ are the maximum and minimum values f ver the ensemble Jp f presented stimuli.] (f) Cnvergence f the algrithm accrding t infrmatin and prjectin!_ as a functin f inverse effective temperature žµ. where 5w Gaussian randm variable f variance ² mdels additive nise, and functin 3 w D fr, and zer therwise. Tgether with the RF, the parameters fr threshld and the nise variance ² determine the input-utput functin. The spike-triggered average (STA), shwn in Fig. 1(b), is bradened because f spatial crrelatins present in natural stimuli. If stimuli were drawn frm a Gaussian prbability distributin, they culd be decrrelated by multiplying K[i5 by the inverse f the a priri cvariance matrix, accrding t the reverse crrelatin methd. The prcedure is nt valid fr nn-gaussian stimuli and nnlinear input-utput functins (1). The result f such a decrrelatin is shwn in Fig. 1(c). It is clearly missing the structure f the mdel filter. Hwever, it is pssible t btain a gd estimate f it by maximizing infrmatin directly, see panel (d). A typical prgress f the simulated annealing algrithm with decreasing temperature ž is shwn in panel (e). There we plt bth the infrmatin alng the vectr, and its prjectin n. The final value f prjectin depends 8* n the size f the data set, see belw. In the example 8' shwn in Fig. 1 there were ¹ spikes with average prbability f spike per frame. Having recnstructed the RF, ne can prceed t sample 13 the nnlinear input-utput functin. This is dne by cnstructing histgrams fr º!( kh 0 13 and º!p kh? ; 4?6I79&: f prjectins nt vectr 13^4?687 kh 0 9:G; fund by maximizing infrmatin, and taking their rati. In Fig. 1(e) we cmpare»!( kh? (crsses) with the prbability 13^4?687 9:_; used in the mdel (slid line). 3. Estimated deviatin frm the ptimal directin When infrmatin is calculated with respect t a finite data set, the vectr which maxi- J mizes will deviate frm the true RF. The deviatin { arises because the prbability distributins are estimated frm experimental histgrams and differ frm the

6 e Ž X e J J e 1 v max Figure : Prjectin f vectr kh N 1 spike that maximizes infrmatin nk[lm RF N*O functin f the number f spikes t shw the linear scaling in (slid line is a fit). X e 10 5 is pltted as a distributins fund in the limit n infinite data size. Fr a simple cell, the quality f recnstructin can be characterized by the prjectin!_, where bth and are nrmalized, ¼ and { is by definitin rthgnal t. The deviatin { ¼, where is the Hessian f infrmatin. Its structure is similar t that f a cvariance matrix: ¼ Ž¾½ Sq } wp13[w"; : wàs q 13[wm; &: 13[w Á yh { ½ ; wp y^ Ž ; w y^ ½ ; wp (8) When averaged ver pssible utcmes f N trials, the gradient f infrmatin is zer fr the ptimal directin. Here in rder t evaluate y^{ Â-à V ¼ y J_Ä" ¼ Y, we need t knw the variance w J f the gradient f. By discretizing bth the space f stimuli and pssible prjectins, and assuming that the prbability f generating a spike is independent fr different bins, ne culd btain that y J Ž J ½ ¼ Ž¾½ K[LM N*O S q. Therefre an expected errr in the recnstructin f the ptimal filter is inversely prprtinal t the number f spikes and is given by:!8 y^{ ÂuÃ*Å V ¼ K[LM N0O Y S q where Âuà Šmeans that the trace is taken in the subspace rthgnal t the mdel filter, since by definitin {#Æ!-. In Fig. we plt the average prjectin f the nrmalized recnstructed vectr n the RF, and shw that it scales with the number f spikes. 3.3 Cmplex Cell A sequence f spikes frm a mdel cell with tw relevant directins was simulated by prjecting each f the stimuli n vectrs that differ by ÇuX in their spatial phase, taken t mimic prperties f cmplex cells, see Fig. 3. A particular frame leads t a spike accrding t a lgical OR, that is if either zÿ!p,, zÿ!p, r exceeds a threshld in the presence f nise. Similarly t (7), :_; : # É> xzy5 *; ; ËÊF *; ; where and are independent Gaussian variables. The sampling f this input-utput functin by ur particular set f natural stimuli is shwn in Fig. 3(c). Sme, especially large, cmbinatins f values f and are nt present in the ensemble. (9) (10)

7 Ì Ð t Ì Ì Ð t Ì We start by maximizing infrmatin with respect t ne directin. Cntrary t analysis fr a simple cell, ne ptimal directin recvers nly abut J K[L@M N*O 60 f the ttal infrmatin per spike. This is significantly different frm the ttal fr stimuli drawn frm natural scenes, where due t crrelatins even a randm vectr has a high prbability f explaining 60 f ttal infrmatin per spike. We therefre g n t maximize infrmatin with respect t tw directins. An example f the recnstructin f input-utput functin f a cmplex cell is given in Fig. 3. Vectrs and Jp that maximize are nt rthgnal, and are als rtated with respect t and. Hwever, the quality f recnstructin is independent f a particular chice f basis with the RS. The apprpriate measure f similarity ˆ5Ð between the tw planes is the dt prduct f their nrmals. In the example f Fig. 3, Î?Ï Î?Ï¾Ñ t ˆ I' Ó. ÍÌ!I ÒÌ Ñ Maximizing infrmatin with respect t tw directins requires a significantly slwer cling rate, and cnsequently lnger cmputatinal times. Hwever, an expected errr in the recnstructin, Ô ˆ^Ð ÍÌ Î?Ï Î0Ï Ñ!, t ˆ ÍÌ Ñ, fllws a e K[LM N0O behavir, similarly t (9), and is rughly twice that fr a simple cell given the same number f spikes. In this calculatin there were spikes. (a) e (1) (b) e () (c) P(spike s (1),s () ) mdel (d) v 1 30 (e) v (f) P(spike s v 1,s v ) recnstructin and Figure 3: Analysis f a mdel cmplex cell with relevant directins shwn in (a) and (b). Spikes are generated I' accrding t an OR input-utput functin > with the threshld kh? khm ³ 8' and nise variance ²¹ kh? khm ³. Panel (c) shws hw the input-utput functin is sampled by ur ensemble f stimuli. Dark. Belw, pixels fr large values f and 13 crrespnd t cases where x we shw vectrs and J, fund by maximizing infrmatin tgether with the crrespnding input-utput functin with respect t prjectins! and!. In cnclusin, features f the stimulus that are mst relevant fr generating the respnse f a neurn can be fund by maximizing infrmatin between the sequence f respnses and the prjectin f stimuli n trial vectrs within the stimulus space. Calculated in this manner, infrmatin becmes a functin f directin in a stimulus space. Thse directins that maximize the infrmatin and accunt fr the ttal infrmatin per respnse f interest span the relevant subspace. This analysis allws the recnstructin f the relevant subspace withut assuming a particular frm f the input-utput functin. It can be strngly nnlinear within the relevant subspace, and is t be estimated frm experimental histgrams. Mst imprtantly, this methd can be used with any stimulus ensemble, even thse that are strngly nn-gaussian as in the case f natural images. Acknwledgments We thank K. D. Miller fr many helpful discussins. Wrk at UCSF was supprted in part by the Slan and Swartz Fundatins and by a training grant frm the NIH. Our cllab-

8 ratin began at the Marine Bilgical Labratry in a curse supprted by grants frm NIMH and the Hward Hughes Medical Institute. References [1] F. Rieke, D. A. Bdnar, and W. Bialek. Naturalistic stimuli increase the rate and efficiency f infrmatin transmissin by primary auditry afferents. Prc. R. Sc. Lnd. B, 6:59 65, (1995). [] W. E. Vinje and J. L. Gallant. Sparse cding and decrrelatin in primary visual crtex during natural visin. Science, 87: , 000. [3] F. E. Theunissen, K. Sen, and A. J. Dupe. Spectral-tempral receptive fields f nnlinear auditry neurns btained using natural sunds. J. Neursci., 0: , 000. [4] G. D. Lewen, W. Bialek, and R. R. de Ruyter van Steveninck. Neural cding f naturalistic mtin stimuli. Netwrk: Cmput. Neural Syst., 1:317 39, 001. [5] N. J. Vickers, T. A. Christensen, T. Baker, and J. G. Hildebrand. Odur-plume dynamics influence the brain s lfactry cde. Nature, 410: , 001. [6] K. Sen, F. E. Theunissen, and A. J. Dupe. Feature analysis f natural sunds in the sngbird auditry frebrain. J. Neurphysil., 86: , 001. [7] D. L. Ringach, M. J. Hawken, and R. Shapley. Receptive field structure f neurns in mnkey visual crtex revealed by stimulatin with natural image sequences. Jurnal f Visin, :1 4, 00. [8] W. E. Vinje and J. L. Gallant. Natural stimulatin f the nnclassical receptive field increases infrmatin transmissin efficiency in V1. J. Neursci., : , 00. [9] D. L. Ruderman and W. Bialek. Statistics f natural images: scaling in the wds. Phys. Rev. Lett., 73: , [10] D. J. Field. Relatins between the statistics f natural images and the respnse prperties f crtical cells. J. Opt. Sc. Am. A, 4: , [11] P. Kara, P. Reinagel, and R. C. Reid. Lw respnse variability in simultaneusly recrded retinal, thalamic, and crtical neurns. Neurn, 7: , 000. [1] R. R. de Ruyter van Steveninck, G. D. Lewen, S. P. Strng, R. Kberle, and W. Bialek. Reprducibility and variability in neural spike trains. Science, 75: , [13] F. Rieke, D. Warland, R. R. de Ruyter van Steveninck, and W. Bialek. Spikes: Explring the neural cde. MIT Press, Cambridge, [14] E. de Ber and P. Kuyper. Triggered crrelatin. IEEE Trans. Bimed. Eng., 15: , [15] N. Brenner, W. Bialek, and R. R. de Ruyter van Steveninck. Adaptive rescaling maximizes infrmatin transmissin. Neurn, 6:695 70, 000. [16] R. R. de Ruyter van Steveninck and W. Bialek. Real-time perfrmance f a mvement-sensitive neurn in the blwfly visual system: cding and infrmatin transfer in shrt spike sequences. Prc. R. Sc. Lnd. B, 34: , [17] N. Tishby, F. C. Pereira, and W. Bialek. The infrmatin bttleneck methd. In Prceedings f the 37th Allertn Cnference n Cmmunicatin, Cntrl and Cmputing, edited by B. Hajek & R. S. Sreenivas. University f Illinis, , [18] A. G. Dimitrv and J. P. Miller. Neural cding and decding: cmmunicatin channels and quantizatin. Netwrk: Cmput. Neural Syst., 1:441 47, 001. [19] L. Paninski. Cnvergence prperties f sme spike-triggered analysis techniques. In Advances in Neural Infrmatin Prcessing 15, edited by 003. [0] N. Brenner, S. P. Strng, R. Kberle, W Bialek, and R. R. de Ruyter van Steveninck. Synergy in a neural cde. Neural Cmp., 1: , 000. [1] A. Treves and S. Panzeri. The upward bias in measures f infrmatin derived frm limited data samples. Neural Cmp., 7:399, [] S. P. Strng, R. Kberle, R. R. de Ruyter van Steveninck, and W. Bialek. Entrpy and infrmatin in neural spike trains. Phys. Rev. Lett., 80:197 00, 1998.

Pattern Recognition 2014 Support Vector Machines

Pattern Recognition 2014 Support Vector Machines Pattern Recgnitin 2014 Supprt Vectr Machines Ad Feelders Universiteit Utrecht Ad Feelders ( Universiteit Utrecht ) Pattern Recgnitin 1 / 55 Overview 1 Separable Case 2 Kernel Functins 3 Allwing Errrs (Sft

More information

, which yields. where z1. and z2

, which yields. where z1. and z2 The Gaussian r Nrmal PDF, Page 1 The Gaussian r Nrmal Prbability Density Functin Authr: Jhn M Cimbala, Penn State University Latest revisin: 11 September 13 The Gaussian r Nrmal Prbability Density Functin

More information

ENSC Discrete Time Systems. Project Outline. Semester

ENSC Discrete Time Systems. Project Outline. Semester ENSC 49 - iscrete Time Systems Prject Outline Semester 006-1. Objectives The gal f the prject is t design a channel fading simulatr. Upn successful cmpletin f the prject, yu will reinfrce yur understanding

More information

Distributions, spatial statistics and a Bayesian perspective

Distributions, spatial statistics and a Bayesian perspective Distributins, spatial statistics and a Bayesian perspective Dug Nychka Natinal Center fr Atmspheric Research Distributins and densities Cnditinal distributins and Bayes Thm Bivariate nrmal Spatial statistics

More information

the results to larger systems due to prop'erties of the projection algorithm. First, the number of hidden nodes must

the results to larger systems due to prop'erties of the projection algorithm. First, the number of hidden nodes must M.E. Aggune, M.J. Dambrg, M.A. El-Sharkawi, R.J. Marks II and L.E. Atlas, "Dynamic and static security assessment f pwer systems using artificial neural netwrks", Prceedings f the NSF Wrkshp n Applicatins

More information

Checking the resolved resonance region in EXFOR database

Checking the resolved resonance region in EXFOR database Checking the reslved resnance regin in EXFOR database Gttfried Bertn Sciété de Calcul Mathématique (SCM) Oscar Cabells OECD/NEA Data Bank JEFF Meetings - Sessin JEFF Experiments Nvember 0-4, 017 Bulgne-Billancurt,

More information

T Algorithmic methods for data mining. Slide set 6: dimensionality reduction

T Algorithmic methods for data mining. Slide set 6: dimensionality reduction T-61.5060 Algrithmic methds fr data mining Slide set 6: dimensinality reductin reading assignment LRU bk: 11.1 11.3 PCA tutrial in mycurses (ptinal) ptinal: An Elementary Prf f a Therem f Jhnsn and Lindenstrauss,

More information

A Matrix Representation of Panel Data

A Matrix Representation of Panel Data web Extensin 6 Appendix 6.A A Matrix Representatin f Panel Data Panel data mdels cme in tw brad varieties, distinct intercept DGPs and errr cmpnent DGPs. his appendix presents matrix algebra representatins

More information

David HORN and Irit OPHER. School of Physics and Astronomy. Raymond and Beverly Sackler Faculty of Exact Sciences

David HORN and Irit OPHER. School of Physics and Astronomy. Raymond and Beverly Sackler Faculty of Exact Sciences Cmplex Dynamics f Neurnal Threshlds David HORN and Irit OPHER Schl f Physics and Astrnmy Raymnd and Beverly Sackler Faculty f Exact Sciences Tel Aviv University, Tel Aviv 69978, Israel hrn@neurn.tau.ac.il

More information

Lecture 10, Principal Component Analysis

Lecture 10, Principal Component Analysis Principal Cmpnent Analysis Lecture 10, Principal Cmpnent Analysis Ha Helen Zhang Fall 2017 Ha Helen Zhang Lecture 10, Principal Cmpnent Analysis 1 / 16 Principal Cmpnent Analysis Lecture 10, Principal

More information

The blessing of dimensionality for kernel methods

The blessing of dimensionality for kernel methods fr kernel methds Building classifiers in high dimensinal space Pierre Dupnt Pierre.Dupnt@ucluvain.be Classifiers define decisin surfaces in sme feature space where the data is either initially represented

More information

Part 3 Introduction to statistical classification techniques

Part 3 Introduction to statistical classification techniques Part 3 Intrductin t statistical classificatin techniques Machine Learning, Part 3, March 07 Fabi Rli Preamble ØIn Part we have seen that if we knw: Psterir prbabilities P(ω i / ) Or the equivalent terms

More information

CHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS

CHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS CHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS 1 Influential bservatins are bservatins whse presence in the data can have a distrting effect n the parameter estimates and pssibly the entire analysis,

More information

STIMULUS ENCODING BY MUL TIDIMENSIONAL RECEPTIVE FIELDS IN SINGLE CELLS AND CELL POPULATIONS IN VI OF A WAKE MONKEY

STIMULUS ENCODING BY MUL TIDIMENSIONAL RECEPTIVE FIELDS IN SINGLE CELLS AND CELL POPULATIONS IN VI OF A WAKE MONKEY STIMULUS ENCODING BY MUL TIDIMENSIONAL RECEPTIVE FIELDS IN SINGLE CELLS AND CELL POPULATIONS IN VI OF A WAKE MONKEY Edward Stern Center fr Neural Cmputatin and Department f Neurbilgy Life Sciences Institute

More information

k-nearest Neighbor How to choose k Average of k points more reliable when: Large k: noise in attributes +o o noise in class labels

k-nearest Neighbor How to choose k Average of k points more reliable when: Large k: noise in attributes +o o noise in class labels Mtivating Example Memry-Based Learning Instance-Based Learning K-earest eighbr Inductive Assumptin Similar inputs map t similar utputs If nt true => learning is impssible If true => learning reduces t

More information

Modelling of Clock Behaviour. Don Percival. Applied Physics Laboratory University of Washington Seattle, Washington, USA

Modelling of Clock Behaviour. Don Percival. Applied Physics Laboratory University of Washington Seattle, Washington, USA Mdelling f Clck Behaviur Dn Percival Applied Physics Labratry University f Washingtn Seattle, Washingtn, USA verheads and paper fr talk available at http://faculty.washingtn.edu/dbp/talks.html 1 Overview

More information

Least Squares Optimal Filtering with Multirate Observations

Least Squares Optimal Filtering with Multirate Observations Prc. 36th Asilmar Cnf. n Signals, Systems, and Cmputers, Pacific Grve, CA, Nvember 2002 Least Squares Optimal Filtering with Multirate Observatins Charles W. herrien and Anthny H. Hawes Department f Electrical

More information

Determining the Accuracy of Modal Parameter Estimation Methods

Determining the Accuracy of Modal Parameter Estimation Methods Determining the Accuracy f Mdal Parameter Estimatin Methds by Michael Lee Ph.D., P.E. & Mar Richardsn Ph.D. Structural Measurement Systems Milpitas, CA Abstract The mst cmmn type f mdal testing system

More information

Math Foundations 20 Work Plan

Math Foundations 20 Work Plan Math Fundatins 20 Wrk Plan Units / Tpics 20.8 Demnstrate understanding f systems f linear inequalities in tw variables. Time Frame December 1-3 weeks 6-10 Majr Learning Indicatrs Identify situatins relevant

More information

3.4 Shrinkage Methods Prostate Cancer Data Example (Continued) Ridge Regression

3.4 Shrinkage Methods Prostate Cancer Data Example (Continued) Ridge Regression 3.3.4 Prstate Cancer Data Example (Cntinued) 3.4 Shrinkage Methds 61 Table 3.3 shws the cefficients frm a number f different selectin and shrinkage methds. They are best-subset selectin using an all-subsets

More information

Sequential Allocation with Minimal Switching

Sequential Allocation with Minimal Switching In Cmputing Science and Statistics 28 (1996), pp. 567 572 Sequential Allcatin with Minimal Switching Quentin F. Stut 1 Janis Hardwick 1 EECS Dept., University f Michigan Statistics Dept., Purdue University

More information

COMP 551 Applied Machine Learning Lecture 11: Support Vector Machines

COMP 551 Applied Machine Learning Lecture 11: Support Vector Machines COMP 551 Applied Machine Learning Lecture 11: Supprt Vectr Machines Instructr: (jpineau@cs.mcgill.ca) Class web page: www.cs.mcgill.ca/~jpineau/cmp551 Unless therwise nted, all material psted fr this curse

More information

Internal vs. external validity. External validity. This section is based on Stock and Watson s Chapter 9.

Internal vs. external validity. External validity. This section is based on Stock and Watson s Chapter 9. Sectin 7 Mdel Assessment This sectin is based n Stck and Watsn s Chapter 9. Internal vs. external validity Internal validity refers t whether the analysis is valid fr the ppulatin and sample being studied.

More information

CHAPTER 3 INEQUALITIES. Copyright -The Institute of Chartered Accountants of India

CHAPTER 3 INEQUALITIES. Copyright -The Institute of Chartered Accountants of India CHAPTER 3 INEQUALITIES Cpyright -The Institute f Chartered Accuntants f India INEQUALITIES LEARNING OBJECTIVES One f the widely used decisin making prblems, nwadays, is t decide n the ptimal mix f scarce

More information

SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis

SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical mdel fr micrarray data analysis David Rssell Department f Bistatistics M.D. Andersn Cancer Center, Hustn, TX 77030, USA rsselldavid@gmail.cm

More information

CS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007

CS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007 CS 477/677 Analysis f Algrithms Fall 2007 Dr. Gerge Bebis Curse Prject Due Date: 11/29/2007 Part1: Cmparisn f Srting Algrithms (70% f the prject grade) The bjective f the first part f the assignment is

More information

Chapter 3: Cluster Analysis

Chapter 3: Cluster Analysis Chapter 3: Cluster Analysis } 3.1 Basic Cncepts f Clustering 3.1.1 Cluster Analysis 3.1. Clustering Categries } 3. Partitining Methds 3..1 The principle 3.. K-Means Methd 3..3 K-Medids Methd 3..4 CLARA

More information

NUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION

NUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION NUROP Chinese Pinyin T Chinese Character Cnversin NUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION CHIA LI SHI 1 AND LUA KIM TENG 2 Schl f Cmputing, Natinal University f Singapre 3 Science

More information

IAML: Support Vector Machines

IAML: Support Vector Machines 1 / 22 IAML: Supprt Vectr Machines Charles Suttn and Victr Lavrenk Schl f Infrmatics Semester 1 2 / 22 Outline Separating hyperplane with maimum margin Nn-separable training data Epanding the input int

More information

COMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification

COMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification COMP 551 Applied Machine Learning Lecture 5: Generative mdels fr linear classificatin Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Jelle Pineau Class web page: www.cs.mcgill.ca/~hvanh2/cmp551

More information

February 28, 2013 COMMENTS ON DIFFUSION, DIFFUSIVITY AND DERIVATION OF HYPERBOLIC EQUATIONS DESCRIBING THE DIFFUSION PHENOMENA

February 28, 2013 COMMENTS ON DIFFUSION, DIFFUSIVITY AND DERIVATION OF HYPERBOLIC EQUATIONS DESCRIBING THE DIFFUSION PHENOMENA February 28, 2013 COMMENTS ON DIFFUSION, DIFFUSIVITY AND DERIVATION OF HYPERBOLIC EQUATIONS DESCRIBING THE DIFFUSION PHENOMENA Mental Experiment regarding 1D randm walk Cnsider a cntainer f gas in thermal

More information

Performance Bounds for Detect and Avoid Signal Sensing

Performance Bounds for Detect and Avoid Signal Sensing Perfrmance unds fr Detect and Avid Signal Sensing Sam Reisenfeld Real-ime Infrmatin etwrks, University f echnlgy, Sydney, radway, SW 007, Australia samr@uts.edu.au Abstract Detect and Avid (DAA) is a Cgnitive

More information

MATHEMATICS SYLLABUS SECONDARY 5th YEAR

MATHEMATICS SYLLABUS SECONDARY 5th YEAR Eurpean Schls Office f the Secretary-General Pedaggical Develpment Unit Ref. : 011-01-D-8-en- Orig. : EN MATHEMATICS SYLLABUS SECONDARY 5th YEAR 6 perid/week curse APPROVED BY THE JOINT TEACHING COMMITTEE

More information

INTRODUCTION TO THE PHYSIOLOGY OF PERCEPTION Chapter 2

INTRODUCTION TO THE PHYSIOLOGY OF PERCEPTION Chapter 2 Intr t the Physilgy f Perceptin 1 Perceptin (PSY 4204) Christine L. Ruva, Ph.D. INTRODUCTION TO THE PHYSIOLOGY OF PERCEPTION Chapter 2 RECEPTORS & NEURAL PROCESSING Theme f Chapter: We d nt just perceive

More information

COMP 551 Applied Machine Learning Lecture 9: Support Vector Machines (cont d)

COMP 551 Applied Machine Learning Lecture 9: Support Vector Machines (cont d) COMP 551 Applied Machine Learning Lecture 9: Supprt Vectr Machines (cnt d) Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Class web page: www.cs.mcgill.ca/~hvanh2/cmp551 Unless therwise

More information

Department of Electrical Engineering, University of Waterloo. Introduction

Department of Electrical Engineering, University of Waterloo. Introduction Sectin 4: Sequential Circuits Majr Tpics Types f sequential circuits Flip-flps Analysis f clcked sequential circuits Mre and Mealy machines Design f clcked sequential circuits State transitin design methd

More information

1 The limitations of Hartree Fock approximation

1 The limitations of Hartree Fock approximation Chapter: Pst-Hartree Fck Methds - I The limitatins f Hartree Fck apprximatin The n electrn single determinant Hartree Fck wave functin is the variatinal best amng all pssible n electrn single determinants

More information

initially lcated away frm the data set never win the cmpetitin, resulting in a nnptimal nal cdebk, [2] [3] [4] and [5]. Khnen's Self Organizing Featur

initially lcated away frm the data set never win the cmpetitin, resulting in a nnptimal nal cdebk, [2] [3] [4] and [5]. Khnen's Self Organizing Featur Cdewrd Distributin fr Frequency Sensitive Cmpetitive Learning with One Dimensinal Input Data Aristides S. Galanpuls and Stanley C. Ahalt Department f Electrical Engineering The Ohi State University Abstract

More information

Perfrmance f Sensitizing Rules n Shewhart Cntrl Charts with Autcrrelated Data Key Wrds: Autregressive, Mving Average, Runs Tests, Shewhart Cntrl Chart

Perfrmance f Sensitizing Rules n Shewhart Cntrl Charts with Autcrrelated Data Key Wrds: Autregressive, Mving Average, Runs Tests, Shewhart Cntrl Chart Perfrmance f Sensitizing Rules n Shewhart Cntrl Charts with Autcrrelated Data Sandy D. Balkin Dennis K. J. Lin y Pennsylvania State University, University Park, PA 16802 Sandy Balkin is a graduate student

More information

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeff Reading: Chapter 2 STATS 202: Data mining and analysis September 27, 2017 1 / 20 Supervised vs. unsupervised learning In unsupervised

More information

THERMAL-VACUUM VERSUS THERMAL- ATMOSPHERIC TESTS OF ELECTRONIC ASSEMBLIES

THERMAL-VACUUM VERSUS THERMAL- ATMOSPHERIC TESTS OF ELECTRONIC ASSEMBLIES PREFERRED RELIABILITY PAGE 1 OF 5 PRACTICES PRACTICE NO. PT-TE-1409 THERMAL-VACUUM VERSUS THERMAL- ATMOSPHERIC Practice: Perfrm all thermal envirnmental tests n electrnic spaceflight hardware in a flight-like

More information

Introduction: A Generalized approach for computing the trajectories associated with the Newtonian N Body Problem

Introduction: A Generalized approach for computing the trajectories associated with the Newtonian N Body Problem A Generalized apprach fr cmputing the trajectries assciated with the Newtnian N Bdy Prblem AbuBar Mehmd, Syed Umer Abbas Shah and Ghulam Shabbir Faculty f Engineering Sciences, GIK Institute f Engineering

More information

ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS

ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS Particle Acceleratrs, 1986, Vl. 19, pp. 99-105 0031-2460/86/1904-0099/$15.00/0 1986 Grdn and Breach, Science Publishers, S.A. Printed in the United States f America ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS

More information

Cells though to send feedback signals from the medulla back to the lamina o L: Lamina Monopolar cells

Cells though to send feedback signals from the medulla back to the lamina o L: Lamina Monopolar cells Classificatin Rules (and Exceptins) Name: Cell type fllwed by either a clumn ID (determined by the visual lcatin f the cell) r a numeric identifier t separate ut different examples f a given cell type

More information

CAUSAL INFERENCE. Technical Track Session I. Phillippe Leite. The World Bank

CAUSAL INFERENCE. Technical Track Session I. Phillippe Leite. The World Bank CAUSAL INFERENCE Technical Track Sessin I Phillippe Leite The Wrld Bank These slides were develped by Christel Vermeersch and mdified by Phillippe Leite fr the purpse f this wrkshp Plicy questins are causal

More information

7 TH GRADE MATH STANDARDS

7 TH GRADE MATH STANDARDS ALGEBRA STANDARDS Gal 1: Students will use the language f algebra t explre, describe, represent, and analyze number expressins and relatins 7 TH GRADE MATH STANDARDS 7.M.1.1: (Cmprehensin) Select, use,

More information

Tree Structured Classifier

Tree Structured Classifier Tree Structured Classifier Reference: Classificatin and Regressin Trees by L. Breiman, J. H. Friedman, R. A. Olshen, and C. J. Stne, Chapman & Hall, 98. A Medical Eample (CART): Predict high risk patients

More information

Support-Vector Machines

Support-Vector Machines Supprt-Vectr Machines Intrductin Supprt vectr machine is a linear machine with sme very nice prperties. Haykin chapter 6. See Alpaydin chapter 13 fr similar cntent. Nte: Part f this lecture drew material

More information

Dead-beat controller design

Dead-beat controller design J. Hetthéssy, A. Barta, R. Bars: Dead beat cntrller design Nvember, 4 Dead-beat cntrller design In sampled data cntrl systems the cntrller is realised by an intelligent device, typically by a PLC (Prgrammable

More information

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff

Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeoff Lecture 2: Supervised vs. unsupervised learning, bias-variance tradeff Reading: Chapter 2 STATS 202: Data mining and analysis September 27, 2017 1 / 20 Supervised vs. unsupervised learning In unsupervised

More information

Bootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) >

Bootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) > Btstrap Methd > # Purpse: understand hw btstrap methd wrks > bs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(bs) > mean(bs) [1] 21.64625 > # estimate f lambda > lambda = 1/mean(bs);

More information

Source Coding and Compression

Source Coding and Compression Surce Cding and Cmpressin Heik Schwarz Cntact: Dr.-Ing. Heik Schwarz heik.schwarz@hhi.fraunhfer.de Heik Schwarz Surce Cding and Cmpressin September 22, 2013 1 / 60 PartI: Surce Cding Fundamentals Heik

More information

making triangle (ie same reference angle) ). This is a standard form that will allow us all to have the X= y=

making triangle (ie same reference angle) ). This is a standard form that will allow us all to have the X= y= Intrductin t Vectrs I 21 Intrductin t Vectrs I 22 I. Determine the hrizntal and vertical cmpnents f the resultant vectr by cunting n the grid. X= y= J. Draw a mangle with hrizntal and vertical cmpnents

More information

A New Evaluation Measure. J. Joiner and L. Werner. The problems of evaluation and the needed criteria of evaluation

A New Evaluation Measure. J. Joiner and L. Werner. The problems of evaluation and the needed criteria of evaluation III-l III. A New Evaluatin Measure J. Jiner and L. Werner Abstract The prblems f evaluatin and the needed criteria f evaluatin measures in the SMART system f infrmatin retrieval are reviewed and discussed.

More information

COMP 551 Applied Machine Learning Lecture 4: Linear classification

COMP 551 Applied Machine Learning Lecture 4: Linear classification COMP 551 Applied Machine Learning Lecture 4: Linear classificatin Instructr: Jelle Pineau (jpineau@cs.mcgill.ca) Class web page: www.cs.mcgill.ca/~jpineau/cmp551 Unless therwise nted, all material psted

More information

Collocation Map for Overcoming Data Sparseness

Collocation Map for Overcoming Data Sparseness Cllcatin Map fr Overcming Data Sparseness Mnj Kim, Yung S. Han, and Key-Sun Chi Department f Cmputer Science Krea Advanced Institute f Science and Technlgy Taejn, 305-701, Krea mj0712~eve.kaist.ac.kr,

More information

ENGI 4430 Parametric Vector Functions Page 2-01

ENGI 4430 Parametric Vector Functions Page 2-01 ENGI 4430 Parametric Vectr Functins Page -01. Parametric Vectr Functins (cntinued) Any nn-zer vectr r can be decmpsed int its magnitude r and its directin: r rrˆ, where r r 0 Tangent Vectr: dx dy dz dr

More information

FIELD QUALITY IN ACCELERATOR MAGNETS

FIELD QUALITY IN ACCELERATOR MAGNETS FIELD QUALITY IN ACCELERATOR MAGNETS S. Russenschuck CERN, 1211 Geneva 23, Switzerland Abstract The field quality in the supercnducting magnets is expressed in terms f the cefficients f the Furier series

More information

Pressure And Entropy Variations Across The Weak Shock Wave Due To Viscosity Effects

Pressure And Entropy Variations Across The Weak Shock Wave Due To Viscosity Effects Pressure And Entrpy Variatins Acrss The Weak Shck Wave Due T Viscsity Effects OSTAFA A. A. AHOUD Department f athematics Faculty f Science Benha University 13518 Benha EGYPT Abstract:-The nnlinear differential

More information

Analysis on the Stability of Reservoir Soil Slope Based on Fuzzy Artificial Neural Network

Analysis on the Stability of Reservoir Soil Slope Based on Fuzzy Artificial Neural Network Research Jurnal f Applied Sciences, Engineering and Technlgy 5(2): 465-469, 2013 ISSN: 2040-7459; E-ISSN: 2040-7467 Maxwell Scientific Organizatin, 2013 Submitted: May 08, 2012 Accepted: May 29, 2012 Published:

More information

The general linear model and Statistical Parametric Mapping I: Introduction to the GLM

The general linear model and Statistical Parametric Mapping I: Introduction to the GLM The general linear mdel and Statistical Parametric Mapping I: Intrductin t the GLM Alexa Mrcm and Stefan Kiebel, Rik Hensn, Andrew Hlmes & J-B J Pline Overview Intrductin Essential cncepts Mdelling Design

More information

Physics 2010 Motion with Constant Acceleration Experiment 1

Physics 2010 Motion with Constant Acceleration Experiment 1 . Physics 00 Mtin with Cnstant Acceleratin Experiment In this lab, we will study the mtin f a glider as it accelerates dwnhill n a tilted air track. The glider is supprted ver the air track by a cushin

More information

Inference in the Multiple-Regression

Inference in the Multiple-Regression Sectin 5 Mdel Inference in the Multiple-Regressin Kinds f hypthesis tests in a multiple regressin There are several distinct kinds f hypthesis tests we can run in a multiple regressin. Suppse that amng

More information

Particle Size Distributions from SANS Data Using the Maximum Entropy Method. By J. A. POTTON, G. J. DANIELL AND B. D. RAINFORD

Particle Size Distributions from SANS Data Using the Maximum Entropy Method. By J. A. POTTON, G. J. DANIELL AND B. D. RAINFORD 3 J. Appl. Cryst. (1988). 21,3-8 Particle Size Distributins frm SANS Data Using the Maximum Entrpy Methd By J. A. PTTN, G. J. DANIELL AND B. D. RAINFRD Physics Department, The University, Suthamptn S9

More information

A Few Basic Facts About Isothermal Mass Transfer in a Binary Mixture

A Few Basic Facts About Isothermal Mass Transfer in a Binary Mixture Few asic Facts but Isthermal Mass Transfer in a inary Miture David Keffer Department f Chemical Engineering University f Tennessee first begun: pril 22, 2004 last updated: January 13, 2006 dkeffer@utk.edu

More information

Enhancing Performance of MLP/RBF Neural Classifiers via an Multivariate Data Distribution Scheme

Enhancing Performance of MLP/RBF Neural Classifiers via an Multivariate Data Distribution Scheme Enhancing Perfrmance f / Neural Classifiers via an Multivariate Data Distributin Scheme Halis Altun, Gökhan Gelen Nigde University, Electrical and Electrnics Engineering Department Nigde, Turkey haltun@nigde.edu.tr

More information

What is Statistical Learning?

What is Statistical Learning? What is Statistical Learning? Sales 5 10 15 20 25 Sales 5 10 15 20 25 Sales 5 10 15 20 25 0 50 100 200 300 TV 0 10 20 30 40 50 Radi 0 20 40 60 80 100 Newspaper Shwn are Sales vs TV, Radi and Newspaper,

More information

Synchronous Motor V-Curves

Synchronous Motor V-Curves Synchrnus Mtr V-Curves 1 Synchrnus Mtr V-Curves Intrductin Synchrnus mtrs are used in applicatins such as textile mills where cnstant speed peratin is critical. Mst small synchrnus mtrs cntain squirrel

More information

Multiple Source Multiple. using Network Coding

Multiple Source Multiple. using Network Coding Multiple Surce Multiple Destinatin Tplgy Inference using Netwrk Cding Pegah Sattari EECS, UC Irvine Jint wrk with Athina Markpulu, at UCI, Christina Fraguli, at EPFL, Lausanne Outline Netwrk Tmgraphy Gal,

More information

Building to Transformations on Coordinate Axis Grade 5: Geometry Graph points on the coordinate plane to solve real-world and mathematical problems.

Building to Transformations on Coordinate Axis Grade 5: Geometry Graph points on the coordinate plane to solve real-world and mathematical problems. Building t Transfrmatins n Crdinate Axis Grade 5: Gemetry Graph pints n the crdinate plane t slve real-wrld and mathematical prblems. 5.G.1. Use a pair f perpendicular number lines, called axes, t define

More information

OF SIMPLY SUPPORTED PLYWOOD PLATES UNDER COMBINED EDGEWISE BENDING AND COMPRESSION

OF SIMPLY SUPPORTED PLYWOOD PLATES UNDER COMBINED EDGEWISE BENDING AND COMPRESSION U. S. FOREST SERVICE RESEARCH PAPER FPL 50 DECEMBER U. S. DEPARTMENT OF AGRICULTURE FOREST SERVICE FOREST PRODUCTS LABORATORY OF SIMPLY SUPPORTED PLYWOOD PLATES UNDER COMBINED EDGEWISE BENDING AND COMPRESSION

More information

Application of ILIUM to the estimation of the T eff [Fe/H] pair from BP/RP

Application of ILIUM to the estimation of the T eff [Fe/H] pair from BP/RP Applicatin f ILIUM t the estimatin f the T eff [Fe/H] pair frm BP/RP prepared by: apprved by: reference: issue: 1 revisin: 1 date: 2009-02-10 status: Issued Cryn A.L. Bailer-Jnes Max Planck Institute fr

More information

Smoothing, penalized least squares and splines

Smoothing, penalized least squares and splines Smthing, penalized least squares and splines Duglas Nychka, www.image.ucar.edu/~nychka Lcally weighted averages Penalized least squares smthers Prperties f smthers Splines and Reprducing Kernels The interplatin

More information

MODULE FOUR. This module addresses functions. SC Academic Elementary Algebra Standards:

MODULE FOUR. This module addresses functions. SC Academic Elementary Algebra Standards: MODULE FOUR This mdule addresses functins SC Academic Standards: EA-3.1 Classify a relatinship as being either a functin r nt a functin when given data as a table, set f rdered pairs, r graph. EA-3.2 Use

More information

Analysis of Curved Bridges Crossing Fault Rupture Zones

Analysis of Curved Bridges Crossing Fault Rupture Zones Analysis f Curved Bridges Crssing Fault Rupture Znes R.K.Gel, B.Qu & O.Rdriguez Dept. f Civil and Envirnmental Engineering, Califrnia Plytechnic State University, San Luis Obisp, CA 93407, USA SUMMARY:

More information

ELT COMMUNICATION THEORY

ELT COMMUNICATION THEORY ELT 41307 COMMUNICATION THEORY Matlab Exercise #2 Randm variables and randm prcesses 1 RANDOM VARIABLES 1.1 ROLLING A FAIR 6 FACED DICE (DISCRETE VALIABLE) Generate randm samples fr rlling a fair 6 faced

More information

Revision: August 19, E Main Suite D Pullman, WA (509) Voice and Fax

Revision: August 19, E Main Suite D Pullman, WA (509) Voice and Fax .7.4: Direct frequency dmain circuit analysis Revisin: August 9, 00 5 E Main Suite D Pullman, WA 9963 (509) 334 6306 ice and Fax Overview n chapter.7., we determined the steadystate respnse f electrical

More information

Admin. MDP Search Trees. Optimal Quantities. Reinforcement Learning

Admin. MDP Search Trees. Optimal Quantities. Reinforcement Learning Admin Reinfrcement Learning Cntent adapted frm Berkeley CS188 MDP Search Trees Each MDP state prjects an expectimax-like search tree Optimal Quantities The value (utility) f a state s: V*(s) = expected

More information

BASD HIGH SCHOOL FORMAL LAB REPORT

BASD HIGH SCHOOL FORMAL LAB REPORT BASD HIGH SCHOOL FORMAL LAB REPORT *WARNING: After an explanatin f what t include in each sectin, there is an example f hw the sectin might lk using a sample experiment Keep in mind, the sample lab used

More information

1996 Engineering Systems Design and Analysis Conference, Montpellier, France, July 1-4, 1996, Vol. 7, pp

1996 Engineering Systems Design and Analysis Conference, Montpellier, France, July 1-4, 1996, Vol. 7, pp THE POWER AND LIMIT OF NEURAL NETWORKS T. Y. Lin Department f Mathematics and Cmputer Science San Jse State University San Jse, Califrnia 959-003 tylin@cs.ssu.edu and Bereley Initiative in Sft Cmputing*

More information

Verification of Quality Parameters of a Solar Panel and Modification in Formulae of its Series Resistance

Verification of Quality Parameters of a Solar Panel and Modification in Formulae of its Series Resistance Verificatin f Quality Parameters f a Slar Panel and Mdificatin in Frmulae f its Series Resistance Sanika Gawhane Pune-411037-India Onkar Hule Pune-411037- India Chinmy Kulkarni Pune-411037-India Ojas Pandav

More information

arxiv:hep-ph/ v1 2 Jun 1995

arxiv:hep-ph/ v1 2 Jun 1995 WIS-95//May-PH The rati F n /F p frm the analysis f data using a new scaling variable S. A. Gurvitz arxiv:hep-ph/95063v1 Jun 1995 Department f Particle Physics, Weizmann Institute f Science, Rehvt 76100,

More information

SAMPLING DYNAMICAL SYSTEMS

SAMPLING DYNAMICAL SYSTEMS SAMPLING DYNAMICAL SYSTEMS Melvin J. Hinich Applied Research Labratries The University f Texas at Austin Austin, TX 78713-8029, USA (512) 835-3278 (Vice) 835-3259 (Fax) hinich@mail.la.utexas.edu ABSTRACT

More information

Computational modeling techniques

Computational modeling techniques Cmputatinal mdeling techniques Lecture 11: Mdeling with systems f ODEs In Petre Department f IT, Ab Akademi http://www.users.ab.fi/ipetre/cmpmd/ Mdeling with differential equatins Mdeling strategy Fcus

More information

Figure 1a. A planar mechanism.

Figure 1a. A planar mechanism. ME 5 - Machine Design I Fall Semester 0 Name f Student Lab Sectin Number EXAM. OPEN BOOK AND CLOSED NOTES. Mnday, September rd, 0 Write n ne side nly f the paper prvided fr yur slutins. Where necessary,

More information

Lecture 13: Markov Chain Monte Carlo. Gibbs sampling

Lecture 13: Markov Chain Monte Carlo. Gibbs sampling Lecture 13: Markv hain Mnte arl Gibbs sampling Gibbs sampling Markv chains 1 Recall: Apprximate inference using samples Main idea: we generate samples frm ur Bayes net, then cmpute prbabilities using (weighted)

More information

General Chemistry II, Unit I: Study Guide (part I)

General Chemistry II, Unit I: Study Guide (part I) 1 General Chemistry II, Unit I: Study Guide (part I) CDS Chapter 14: Physical Prperties f Gases Observatin 1: Pressure- Vlume Measurements n Gases The spring f air is measured as pressure, defined as the

More information

4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression

4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression 4th Indian Institute f Astrphysics - PennState Astrstatistics Schl July, 2013 Vainu Bappu Observatry, Kavalur Crrelatin and Regressin Rahul Ry Indian Statistical Institute, Delhi. Crrelatin Cnsider a tw

More information

Resampling Methods. Chapter 5. Chapter 5 1 / 52

Resampling Methods. Chapter 5. Chapter 5 1 / 52 Resampling Methds Chapter 5 Chapter 5 1 / 52 1 51 Validatin set apprach 2 52 Crss validatin 3 53 Btstrap Chapter 5 2 / 52 Abut Resampling An imprtant statistical tl Pretending the data as ppulatin and

More information

Admissibility Conditions and Asymptotic Behavior of Strongly Regular Graphs

Admissibility Conditions and Asymptotic Behavior of Strongly Regular Graphs Admissibility Cnditins and Asympttic Behavir f Strngly Regular Graphs VASCO MOÇO MANO Department f Mathematics University f Prt Oprt PORTUGAL vascmcman@gmailcm LUÍS ANTÓNIO DE ALMEIDA VIEIRA Department

More information

Methods for Determination of Mean Speckle Size in Simulated Speckle Pattern

Methods for Determination of Mean Speckle Size in Simulated Speckle Pattern 0.478/msr-04-004 MEASUREMENT SCENCE REVEW, Vlume 4, N. 3, 04 Methds fr Determinatin f Mean Speckle Size in Simulated Speckle Pattern. Hamarvá, P. Šmíd, P. Hrváth, M. Hrabvský nstitute f Physics f the Academy

More information

(Communicated at the meeting of January )

(Communicated at the meeting of January ) Physics. - Establishment f an Abslute Scale fr the herm-electric Frce. By G. BOR ELlUS. W. H. KEESOM. C. H. JOHANSSON and J. O. LND E. Supplement N0. 69b t the Cmmunicatins frm the Physical Labratry at

More information

Lecture 6: Phase Space and Damped Oscillations

Lecture 6: Phase Space and Damped Oscillations Lecture 6: Phase Space and Damped Oscillatins Oscillatins in Multiple Dimensins The preius discussin was fine fr scillatin in a single dimensin In general, thugh, we want t deal with the situatin where:

More information

2.161 Signal Processing: Continuous and Discrete Fall 2008

2.161 Signal Processing: Continuous and Discrete Fall 2008 MIT OpenCurseWare http://cw.mit.edu 2.161 Signal Prcessing: Cntinuus and Discrete Fall 2008 Fr infrmatin abut citing these materials r ur Terms f Use, visit: http://cw.mit.edu/terms. Massachusetts Institute

More information

MATCHING TECHNIQUES. Technical Track Session VI. Emanuela Galasso. The World Bank

MATCHING TECHNIQUES. Technical Track Session VI. Emanuela Galasso. The World Bank MATCHING TECHNIQUES Technical Track Sessin VI Emanuela Galass The Wrld Bank These slides were develped by Christel Vermeersch and mdified by Emanuela Galass fr the purpse f this wrkshp When can we use

More information

Equilibrium of Stress

Equilibrium of Stress Equilibrium f Stress Cnsider tw perpendicular planes passing thrugh a pint p. The stress cmpnents acting n these planes are as shwn in ig. 3.4.1a. These stresses are usuall shwn tgether acting n a small

More information

Emphases in Common Core Standards for Mathematical Content Kindergarten High School

Emphases in Common Core Standards for Mathematical Content Kindergarten High School Emphases in Cmmn Cre Standards fr Mathematical Cntent Kindergarten High Schl Cntent Emphases by Cluster March 12, 2012 Describes cntent emphases in the standards at the cluster level fr each grade. These

More information

CESAR Science Case The differential rotation of the Sun and its Chromosphere. Introduction. Material that is necessary during the laboratory

CESAR Science Case The differential rotation of the Sun and its Chromosphere. Introduction. Material that is necessary during the laboratory Teacher s guide CESAR Science Case The differential rtatin f the Sun and its Chrmsphere Material that is necessary during the labratry CESAR Astrnmical wrd list CESAR Bklet CESAR Frmula sheet CESAR Student

More information

Chapter 9 Vector Differential Calculus, Grad, Div, Curl

Chapter 9 Vector Differential Calculus, Grad, Div, Curl Chapter 9 Vectr Differential Calculus, Grad, Div, Curl 9.1 Vectrs in 2-Space and 3-Space 9.2 Inner Prduct (Dt Prduct) 9.3 Vectr Prduct (Crss Prduct, Outer Prduct) 9.4 Vectr and Scalar Functins and Fields

More information

ChE 471: LECTURE 4 Fall 2003

ChE 471: LECTURE 4 Fall 2003 ChE 47: LECTURE 4 Fall 003 IDEL RECTORS One f the key gals f chemical reactin engineering is t quantify the relatinship between prductin rate, reactr size, reactin kinetics and selected perating cnditins.

More information