UVA$CS$6316$$ $Fall$2015$Graduate:$$ Machine$Learning$$ $ $Lecture$9:$Support$Vector$Machine$ (Cont.$Revised$Advanced$Version)$
|
|
- Alexis Walters
- 5 years ago
- Views:
Transcription
1 9/30/15 UVACS6316 Fall015Graduate: MachneLearnng Lecture9:SupportVectorMachne (Cont.RevsedAdvancedVerson) Dr.YanjunQ UnverstyofVrgna Departmentof ComputerScence ATT: there exst some nconsstency of math notatons across sldes. Notatons n each slde are mostly selfconsstent. 1 Whereweare?! FvemajorsecHonsofthscourse " Regresson(supervsed) " ClassfcaHon(supervsed) " Unsupervsedmodels " Learnngtheory " Graphcalmodels 9/30/15
2 Whereweare?! ThreemajorsecHonsforclassfcaHon We can dvde the large varety of classfcaton approaches nto roughly three major types 1. Dscrmnatve - drectly estmate a decson rule/boundary - e.g., support vector machne, decson tree. Generatve: - buld a generatve statstcal model - e.g., Bayesan networks 3. Instance based classfers - Use observaton drectly (no models) - e.g. K nearest neghbors 9/30/15 3 ADataset forbnary classfcahon Output as Bnary Class Label: 1 or -1 Data/ponts/nstances/examples/samples/records:[rows] Features/a0rbutes/dmensons/ndependent3varables/covarates/ predctors/regressors:[columns,exceptthelast] Target/outcome/response/label/dependent3varable:specal 9/30/15 columntobepredcted[lastcolumn] 4
3 Max margn classfers Instead of fttng all ponts, focus on boundary ponts Learn a boundary that leads to the largest margn from ponts on both sdes x Why? Intutve, makes sense Some theoretcal support Works well n practce 9/30/15 x 5 1 WhenlnearlySeparableCase Thedecsonboundaryshouldbeasfarawayfrom thedataofbothclassesaspossble Class 1 W s a p-dm vector; b s a scalar Class -1 9/30/15 6
4 Today " SupportVectorMachne(SVM) # HstoryofSVM # LargeMargnLnearClassfer # DefneMargn(M)ntermsofmodelparameter # OpHmzaHontolearnmodelparameters(w,b) # Nonlnearlyseparablecase # OpHmzaHonwthdualform # Nonlneardecsonboundary # PracHcalGude 9/30/15 7 Today " SupportVectorMachne(SVM) # HstoryofSVM # LargeMargnLnearClassfer # DefneMargn(M)ntermsofmodelparameter # OpHmzaHontolearnmodelparameters(w,b) # Nonlnearlyseparablecase # OpHmzaHonwthdualform # Nonlneardecsonboundary # PracHcalGude 9/30/15 8
5 Optmzaton Step.e. learnng optmal parameter for SVM Predct class +1 M M = w T w w T x+b=+1 w T x+b=0 w T x+b=-1 Predct class -1 Mn (w T w)/ subject to the followng constrants: For all x n class + 1 w T x+b >= 1 For all x n class - 1 w T x+b <= -1 } A total of n constrants f we have n nput samples 9/30/15 9 Optmzaton Step.e. learnng optmal parameter for SVM Predct class +1 M M = w T w w T x+b=+1 w T x+b=0 w T x+b=-1 Predct class -1 Mn (w T w)/ subject to the followng constrants: For all x n class + 1 w T x+b >= 1 For all x n class - 1 w T x+b <= -1 } A argmn w,b subject to x Dtran : y total of n constrants f we have n nput samples 9/30/15 10 p =1 w ( x w + b) 1
6 Optmzaton Revew: Ingredents ObjecHvefuncHon Varables Constrants Fndvaluesofthevarables thatmnmzeormaxmzetheobjechvefunchon whlesahsfyngtheconstrants 9/30/15 11 Optmzaton wth Quadratc programmng (QP) Quadratc programmng solves optmzaton problems of the followng form: mn U u T Ru + d T u + c subject to n nequalty constrants: a 11 u 1 + a 1 u +... b 1!!! a n1 u 1 + a n u +... b n and k equvalency constrants: a n +1,1 u 1 + a n +1, u +... = b n +1!!! Quadratc term When a problem can be specfed as a QP problem we can use solvers that are better than gradent descent or smulated annealng a n +k,1 u 1 + a n +k, u +... = b n +k 9/30/15 1
7 SVM as a QP problem w T x+b=+1 w T x+b=0 w T x+b=-1 Predct class +1 Predct class -1 M M = w T w R as I matrx, d as zero vector, c as 0 value mn U u T Ru + d T u + c subject to n nequalty constrants: a 11 u 1 + a 1 u +... b 1!!! Mn (w T w)/ subject to the followng nequalty constrants: For all x n class + 1 w T x+b >= 1 For all x n class - 1 w T x+b <= -1 } A total of n constrants f we have n nput samples a n1 u 1 + a n u +... b n and k equvalency constrants: a n +1,1 u 1 + a n +1, u +...= b n +1!!! a n +k,1 u 1 + a n +k, u +...= b n +k 9/30/15 13 Today " SupportVectorMachne(SVM) # HstoryofSVM # LargeMargnLnearClassfer # DefneMargn(M)ntermsofmodelparameter # OpHmzaHontolearnmodelparameters(w,b) # Nonlnearlyseparablecase # OpHmzaHonwthdualform # Nonlneardecsonboundary # PracHcalGude 9/30/15 14
8 Non lnearly separable case So far we assumed that a lnear plane can perfectly separate the ponts But ths s not usally the case - nose, outlers Hard to solve (two mnmzaton problems) How can we convert ths to a QP problem? - Mnmze tranng errors? mn w T w mn #errors - Penalze tranng errors: mn w T w+c*(#errors) Hard to encode n a QP problem 9/30/15 15 Non lnearly separable case Instead of mnmzng the number of msclassfed ponts we can mnmze the dstance between these ponts and ther correct plane The new optmzaton problem s: ε k +1 plane ε -1 plane j w T n w mn w! + Cε =1 subject to the followng nequalty constrants: For all x n class + 1 ε w T x +b >= 1- For all x n class - 1 ε w T x +b <= -1+ Wat. Are we mssng somethng? 9/30/15 16
9 Fnal optmzaton for non lnearly separable case The new optmzaton problem s: w T n w mn w + Cε =1 subject to the followng nequalty constrants: }A total of n constrants For all! ε 0 } Another n constrants 9/30/15 17 Where we are Two optmzaton problems: For the separable and non separable cases w T n w T w w mn mn w + Cε w =1 For all x n class + 1 w T x+b >= 1 For all x n class - 1 w T x+b <=-1 For all! ε 0 9/30/15 18
10 ModelSelecHon,fndrghtC Selectthe rght penalty parameter C 9/30/15 19 Today " SupportVectorMachne(SVM) # HstoryofSVM # LargeMargnLnearClassfer # DefneMargn(M)ntermsofmodelparameter # OpHmzaHontolearnmodelparameters(w,b) # Nonlnearlyseparablecase # OpHmzaHonwthdualform # Nonlneardecsonboundary # PracHcalGude 9/30/15 0
11 Where we are Two optmzaton problems: For the separable and non separable cases w T n w Mn (w T mn w)/ w + Cε =1 For all! ε 0 Instead of solvng these QPs drectly we wll solve a dual formulaton of the SVM optmzaton problem The man reason for swtchng to ths type of representaton s that t would allow us to use a neat trck that wll make our lves easer (and the run tme faster) 9/30/15 1 Optmzaton Revew: ConstranedOpHmzaHon mn u u s.t.u>=b Allowed mn Case1: b Global mn Allowed mn Case: b Global mn 9/30/15
12 Optmzaton Revew: ConstranedOpHmzaHonwthLagrange Whenequalconstrants!opHmzef(x),3subjecttog (x)=0 MethodofLagrangemulHplers:converttoa hgherfdmensonalproblem Mnmze f ( x) + g ( x) λ w.r.t. ( x1 x n 1 k ; λ λ ) 9/30/15 IntroducngaLagrangemulHplerforeachconstrant ConstructtheLagranganfortheorgnalopHmzaHonproblem 3 OpHmzaHonRevew:Lagrangan (moregeneralstandardform) 9/30/15 FromStanford ConvexOpHmzaHon Boyd&Vandenberghe 4
13 mn u u s.t.u>=b 9/30/15 5 mn u u s.t.u>=b Dual Prmal 9/30/15 6
14 Optmzaton Revew: LagranganDualty ThePrmalProblem mn w Prmal: s.t. ThegeneralzedLagrangan:! 9/30/15! =1 =1 the 's( 0)and 'sarecalledthelagaranganmulhplers Lemma:! AreNwrPenPrmal: k f 0 (w) f (w) 0,!!! = 1,,k h (w)= 0,!!! = 1,,l L(w,α,β)= f 0 (w)+ α f (w) + β h (w) max!l(w,α,β)= f 0 (w) f!w!satsfes!prmal!constrants α,β,α 0 o/w mn w maxα, β, α 0 L ( w, α, β ) l ErcXng@CMU,006f008 7 Optmzaton Revew: Dual Func Recall that Lagrange multplers can be appled to turn the followng problem: L3 9/30/15 8
15 Optmzaton Revew: LagranganDualty,cont. RecallthePrmalProblem: mn w maxα, β, α 0 L ( w, α, β ) TheDualProblem: max α β, α 0 Theorem(weakdualty): d, mn w L ( w, α, β ) * = maxα, β, α 0 mnw L ( w, α, β) mnw maxα, β, α 0 L ( w, α, β) = p * Theorem(strongdualty): Iffthereexstasaddlepontof,wehave 9/30/15 L ( w, α, β ) * d = p * 9 OpHmzaHonRevew:LagrangedualfuncHon Inf(.):greatest lowerbound 9/30/15 30
16 An alternatve representaton of the SVM QP We wll start wth the lnearly separable case Instead of encodng the correct classfcaton rule and constrant we wll use Lagrange multples to encode t as part of the our mnmzaton problem Mn (w T w)/ s.t. (w T x +b)y >= 1 Recall that Lagrange multplers can be appled to turn the followng problem: N ( ) L prmal = 1! w α y (w x + b) 1 =1 9/30/15 31 w T w mn w,b max α α [(wt x + b)y 1] α! 0 9/30/15 3
17 TheDualProblem WemnmzeLwthrespecttowandbfrst:! Notethat(*)mples: max mn, L ( w, b, α) α 0 Plus(***)backtoL,andusng(**),wehave: w b tran w L(w,b,α )! = w α y x = 0,! =1 tran b L(w,b,α )! = α y = 0, w =! tran =1 =1 α y x () * () ** () *** 9/30/15 L(w,b,α )= α 1 =1!,j=1 α α j y y j (x T x j ) 33 Summary:DualforSVM Solvngforwthatgvesmaxmummargn: 1. CombneobjecHvefuncHonandconstrantsntonew objechvefunchon,usnglagrangemulhplers3\alpha 3 N ( ) L prmal = 1! w α y (w x + b) 1 =1. TomnmzethsLagrangan,wetakedervaHvesofwandb andsetthemto0: 9/30/15 34
18 Summary:DualforSVM 3. SubsHtuHngandrearrangnggvesthedualoftheLagrangan: N L dual = α 1 α =1 α j y y j x x j,j! whchwetrytomaxmze(notmnmze). 4. Oncewehavethe\alpha,wecansubsHtutentoprevous equahonstogetwandb. 5. ThsdefneswandbaslnearcombnaHonsofthetranng data. 9/30/15 35 Optmzaton Revew: DualProblem Solvngdualproblemfthedual formseaserthanprmalform Needtochangeprmal mnmzahontodual maxmzahon(or!needto changeprmalmaxmzahonto dualmnmzahon) Onlyvaldwhentheorgnal ophmzahonproblemsconvex/ concave(strongdualty) PrmalProblem,e.g., x * = argmn f (x) x subject to h(x) = c DualProblem, e.g., λ * = argmax λ g(λ) Strong dualty g(λ) = nf( f (x) + λ(h(x) c)) x 9/30/15 36
19 Summary: Dual SVM for lnearly separable case Substtutng w nto our target functon and usng the addtonal constrant we get: Dual formulaton max α α 1 α y = 0 α 0,j α α j y y j x T x j EaserthanorgnalQP,moreeffcentalgorthmsexsttofnd 9/30/15 37 OpHmzaHonRevew: nf(.):greatestlowerbound 9/30/15 38
20 OpHmzaHonRevew: KeyforSVMDual 9/30/15 39 KKT=>Supportvectors NotetheKKTcondHonfffonlyafew 'scan benonzero!! ( ) = 0,!!!! = 1,,n! α y (w x + b) 1 Class 8 = =0 Callthetranngdataponts whose 'sarenonzerothe supportvectors(sv) 5 =0 7 =0 =0 4 =0 6 =1.4 1 =0.8 9/30/15 9 =0 Class 1 3 =0 40
21 Dual SVM - nterpretaton w = α x y α For that are 0, no nfluence 9/30/15 41 Dual SVM for lnearly separable case Our dual target functon: max α α 1 α y = 0 α 0 To evaluate a new sample x ts we need to compute: w T x ts + b =!,j α α j y y j x T x j α y x T x ts + b! = sgn α y x T x ts Dot product for all tranng samples Dot product wth tranng samples y ts + b 9/30/15 SV 4! ( )
22 Dual formulaton for lnearly non separable case Dual target functon: max α α 1 α y = 0 C >α! 0,,j α α j y y j x T x j HyperparameterC shouldbetuned throughkffoldscv The only dfference s that the \alpha are now bounded To evaluate a new sample x j we need to compute: w T x j + b = α y x T x j + b Thssverysmlartothe ophmzahonproblemnthelnear separablecase,exceptthattheres anupperboundcon now Onceagan,effcentalgorthmexst tofnd 9/30/15 43 FastSVMImplementaHons SMO:SequenHalMnmalOpHmzaHon SVMfLght LbSVM BSVM 9/30/15 44
23 SMO:SequenHalMnmalOpHmzaHon Keydea DvdethelargeQPproblemofSVMntoaseresof smallestpossbleqpproblems,whchcanbesolved analyhcallyandthusavodsusngahmefconsumng numercalqpntheloop(akndofsqpmethod). Spacecomplexty:O(n). SnceQPsgreatlysmplfed,mostHmefconsumngpartof SMOstheevaluaHonofdecsonfuncHon,thereforets veryfastforlnearsvmandsparsedata. 9/30/15 45 SMO Ateachstep,SMOchoosesLagrangemulHplersto jontlyophmze,fndtheophmalvaluesforthese mulhplersandupdatesthesvmtoreflectthenew ophmalvalues. Threecomponents AnanalyHcmethodtosolveforthetwoLagrange mulhplers AheursHcforchoosngwhchmulHplerstoopHmze AmethodforcompuHngbateachstep,sothattheKTT condhonsarefulflledforboththetwoexamples (correspondngtothetwomulhplers) 9/30/15 46
24 ChoosngWhchMulHplerstoOpHmze FrstmulHpler IterateovertheenHretranngset,andfndanexample thatvolatesthekttcondhon. SecondmulHpler MaxmzetheszeofsteptakendurngjontopHmzaHon. E 1 fe,wheree stheerroronthefthexample. 9/30/15 47 Today " SupportVectorMachne(SVM) # HstoryofSVM # LargeMargnLnearClassfer # DefneMargn(M)ntermsofmodelparameter # OpHmzaHontolearnmodelparameters(w,b) # Nonlnearlyseparablecase # OpHmzaHonwthdualform # Nonlneardecsonboundary # PracHcalGude 9/30/15 48
25 Classfyng n 1-d Can an SVM correctly classfy ths data? What about ths? X X 9/30/15 49 Classfyng n 1-d Can an SVM correctly classfy ths data? And now? (extend wth polynomal bass ) X X X 9/30/15 50
26 RECAP:Polynomalregresson IntroducebassfuncHons 9/30/15 51 Dr.NandodeFretas stutoralslde NonflnearSVMs:D Theorgnalnputspace(x)canbemappedtosomehgherfdmensonal featurespace(φ(x))wherethetranngsetsseparable: x=(x 1,x ) φ(x) =(x 1,x, x 1 x ) x 1 x Φ: x φ(x) x x 1 9/30/15 5 Ths slde s courtesy of
27 NonflnearSVMs:D Theorgnalnputspace(x)canbemappedtosomehgherfdmensonal featurespace(φ(x))wherethetranngsetsseparable: x=(x 1,x ) φ(x) =(x 1,x, x 1 x ) x 1 x If data s mapped nto suffcently hgh dmenson, then samples wll n general Φ: x φ(x) be lnearly separable; N data ponts are n general separable n a space of N-1 dmensons or more!!! x x 1 9/30/15 53 Ths slde s courtesy of A lttle bt theory: Vapnk-Chervonenks (VC) dmenson If data s mapped nto suffcently hgh dmenson, then samples wll n general be lnearly separable; N data ponts are n general separable n a space of N-1 dmensons or more!!! VCdmensonofthesetoforentedlnesnR s3 ItcanbeshownthattheVCdmensonofthefamlyof orentedseparahnghyperplanesnr N satleastn+1 9/30/15 54
28 Transformaton of Inputs Possble problems - Hgh computaton burden due to hgh-dmensonalty - Many more parameters SVM solves these two ssues smultaneously Kernel trcks for effcent computaton Is ths too much computatonal work? Dual formulaton only assgns parameters to samples, not features Input space Φ(.) Φ( ) Φ( ) Φ( ) Φ ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Φ( ) Feature space 9/30/15 55 Quadratc kernels Whle workng n hgher dmensons s benefcal, t also ncreases our runnng tme because of the dot product computaton However, there s a neat trck we can use max α α α α j y y j Φ(x ) T Φ(x j ),j α y = 0 α 0 consder all quadratc terms for x 1, x x m weghts on quadratc terms wll become clear n the next slde Φ(x) = 1 x 1! x m x 1! x m m+1 lnear terms m quadratc terms m s the number of features n each vector K(x, z) := Φ(x) T Φ(z) x 1 x! m(m-1)/ parwse terms x m 1 x m 9/30/15 56
29 Dot product for quadratc kernels How many operatons do we need for the dot product? 1 1 x 1! z 1! Φ(x) T Φ(z) = x m x 1! x m. z m z 1! z m = x z + x z + x x j z z j +1 j= +1 m m m(m-1)/ =~ m x 1 x! z 1 z! K(x, z) := Φ(x) T Φ(z) x m 1 x m z m 1 z m 9/30/15 57 The kernel trck How many operatons do we need for the dot product? Φ(x) T Φ(z) = x z + x z + x x j z z j +1 j= +1 m m m(m-1)/ =~ m However, we can obtan dramatc savngs by notng that Φ(x) T Φ(z) = (x T z +1) = (x.z +1) = (x.z) + (x.z)+1 = ( x z ) + x z +1 We only need m operatons! = x z + x z + x x j z z j +1 9/30/15 58 j=+1 So,fwedefnethekernelfuncUonasfollows,theres noneedtocarryoutbassfunchon\ (.)explctly K(x, z) = (x T z +1)
30 Where we are Our dual target functon: max α α 1 α α j y y j Φ(x ) T Φ(x j ) α y = 0 α 0,j mn operatons at each teraton To evaluate a new sample x k we need to compute: w T Φ(x j )+ b = α y Φ(x ) T Φ(x j )+ b mr operatons where r are the number of support vectors (whose \alpha >0) useofkernelfunchontoavodcarryngout \ (.)explctlysknownasthekerneltrck So,fwedefnethekernelfuncUonasfollows,there snoneedtocarryout (.)representahonexplctly K(x, z) = (x T z +1) 9/30/15 59 Summary: ModfcaHonDuetoKernelFuncHon ChangeallnnerproductstokernelfuncHons Fortranng, max Orgnal α α 1 α α j y y j x T x j,j Lnear α y = 0 C >α! 0, tran 9/30/15 Wth kernel functon - nonlnear max α α 1 α α j y y j K(x,x j ) α y = 0 C >α! 0, tran,j 60
31 Summary: ModfcaHonDuetoKernelFuncHon FortesHng,thenewdatax_ts Orgnal Lnear! y = sgn α y ts x T! x ts + b tran Wth kernel functon - nonlnear! y = sgn α y ts K(x!,x ts )+ b tran 9/30/15 61 MoreexamplesofkernelfuncHons Lnearkernel(we'veseent) T K ( x, x') = x x' Polynomalkernel(wejustsawanexample) wherep=,3, Togetthefeaturevectorsweconcatenateallpthorder polynomaltermsofthecomponentsofx(weghtedapproprately) Radalbasskernel 9/30/15 ( ) d! K(x,x')= 1+ xt x' K(x,x')= exp r x x'! Inthscase.,rshyperpara.ThefeaturespaceoftheRBFkernelhasannfnte numberofdmensons Neverrepresentfeaturesexplctly Computedotproductsnclosedform VerynteresHngtheory ReproducngKernelHlbertSpaces Notcoveredndetalhere 6
32 KernelFuncHon:ImplctBass RepresentaHon Forsomekernels(e.g.RBF)themplct transformbassform\ph(x)snfntef dmensonal! ButcalculaHonswthkernelaredonenorgnalspace,so computahonalburdenandcurseofdmensonaltyaren ta problem. 9/30/15 63 Anexample:Supportvectormachneswth polynomalkernel 9/30/15 64
33 KernelFuncHons InpracHcaluseofSVM,onlythekernelfuncHon(andnot (.))sspecfed KernelfuncHoncanbethoughtofasasmlartymeasure betweenthenputobjects NotallsmlartymeasurecanbeusedaskernelfuncHon, howevermercer'scondhonstatesthatanyposhvesemf defntekernelk(x,y),.e. canbeexpressedasadotproductnahghdmensonalspace. 9/30/15 65 ChoosngtheKernelFuncHon ProbablythemosttrckypartofusngSVM. ThekernelfuncHonsmportantbecausetcreatesthekernel matrx,whchsummarzeallthedata Manyprncpleshavebeenproposed(dffusonkernel,Fsher kernel,strngkernel,treekernel,graphkernel, ) KerneltrckhashelpedNonftradHonaldatalkestrngsandtreesable tobeusedasnputtosvm,nsteadoffeaturevectors InpracHce,alowdegreepolynomalkernelorRBFkernelwth areasonablewdthsagoodnhaltryformostapplcahons. 9/30/15 66
34 Why do SVMs work? " If we are usng huge features spaces (e.g., wth kernels), how come we are not overfttng the data? # Number of parameters remans the same (and most are set to 0) # Whle we have a lot of nput values, at the end we only care about the support vectors and these are usually a small group of samples # The mnmzaton (or the maxmzng of the margn) functon acts as a sort of regularzaton term leadng to reduced overfttng 9/30/15 67 WhySVMWorks? Vapnkarguesthatthefundamentalproblemsnotthenumberofparameters tobeeshmated.rather,theproblemsabouttheflexbltyofaclassfer Vapnkarguesthattheflexbltyofaclassfershouldnotbecharacterzedby thenumberofparameters,butbythecapactyofaclassfer Thssformalzedbythe VCfdmenson ofaclassfer TheSVMobjecHvecanalsobejusHfedbystructuralrskmnmzaHon:the emprcalrsk(tranngerror),plusatermrelatedtothegeneralzahonablty oftheclassfer,smnmzed Anothervew:theSVMlossfuncHonsanalogoustordgeregresson.The term½ w shrnks theparameterstowardszerotoavodoverfång 9/30/15 68
35 Today " SupportVectorMachne(SVM) # HstoryofSVM # LargeMargnLnearClassfer # DefneMargn(M)ntermsofmodelparameter # OpHmzaHontolearnmodelparameters(w,b) # Nonlnearlyseparablecase # OpHmzaHonwthdualform # Nonlneardecsonboundary # PracHcalGude 9/30/15 69 Software A lst of SVM mplementaton can be found at Some mplementaton (such as LIBSVM) can handle mult-class classfcaton SVMLght s among one of the earlest mplementaton of SVM Several Matlab toolboxes for SVM are also avalable 9/30/15 70
36 Summary:StepsforUsngSVMnHW Preparethefeaturefdatamatrx SelectthekernelfuncHontouse SelecttheparameterofthekernelfuncHonandthe valueofc3 YoucanusethevaluessuggestedbytheSVMsoÇware,oryou cansetapartavaldahonsettodetermnethevaluesofthe parameter Executethetranngalgorthmandobtanthe Unseendatacanbeclassfedusngthe andthe supportvectors 9/30/15 71 Practcal Gude to SVM Fromauthorsofas LIBSVM: APracHcalGudetoSupportVectorClassfcaHon ChhfWeHsu,ChhfChungChang,andChhfJen Ln,003f010 hñp:// gude.pdf 9/30/15 7
37 LIBSVM hñp:// # DevelopedbyChhfJenLnetc. # ToolsforSupportVectorclassfcaHon 9/30/15 # AlsosupportmulHfclassclassfcaHon # C++/Java/Python/Matlab/Perlwrappers # Lnux/UNIX/Wndows # SMOmplementaHon,fast!!! APracHcalGudetoSupportVector ClassfcaHon 73 (a) Data fle formats for LIBSVM Tranng.dat +11: :13:14:f f11: :f14:f :1 +11: :13:f :f f11: :13:14:f : TesHng.dat 9/30/15 74
38 (b)featurepreprocessng (1)CategorcalFeature Recommendusngmnumberstorepresentanmf categoryañrbute. Onlyoneofthemnumberssone,andothersarezero. Forexample,athreefcategoryaÑrbutesuchas{red, green,blue}canberepresentedas(0,0,1),(0,1,0),and (1,0,0) 9/30/15 APracHcalGudetoSupportVector 75 ClassfcaHon FeaturePreprocessng ()ScalngbeforeapplyngSVMsvery mportant toavodañrbutesngreaternumercranges domnahngthosensmallernumercranges. toavodnumercaldffculhesdurngthecalculahon RecommendlnearlyscalngeachaÑrbutetothe range[1,+1]or[0,1]. 9/30/15 76 APracHcalGudetoSupportVector ClassfcaHon
39 FeaturePreprocessng 9/30/15 APracHcalGudetoSupportVector 77 ClassfcaHon FeaturePreprocessng (3)mssngvalue Veryverytrcky! Easyway:tosubsHtutethemssngvaluesbythe meanvalueofthevarable AlÑlebtharderway:mputaHonusngnearest neghbors Evenmorecomplex:e.g.EMbased(beyondthe scope) 9/30/15 78 APracHcalGudetoSupportVector ClassfcaHon
40 (c)modelselechon 9/30/15 79 (c)modelselechon(e.g.forlnearkernel) Selectthe rght penalty parameter C 9/30/15 80
41 (c)modelselechon Threeparametersforapolynomalkernel 9/30/15 APracHcalGudetoSupportVector ClassfcaHon 81 (d)ppelneprocedures (1)tran/test ()kffoldscrossvaldahon (3)kfCVontrantochoose hyperparameter/thentest 9/30/15 8
42 Evaluaton Choce-I: Tran and Test Tranngdataset consstsofnputf outputpars Evaluaton f(x? ) Measure Loss on par! ( f(x? ), y? ) 9/30/15 83 Evaluaton Choce-II: CrossValdaHon Problem:don thaveenoughdatatosetasdea testset SoluHon:Eachdatapontsusedbothastran andtest Commontypes: fkffoldcrossfvaldahon(e.g.k=5,k=10) fffoldcrossfvaldahon fleavefonefoutcrossfvaldahon(loocv) 9/30/15 AgoodpracHces:torandomshuffleall tranngsamplebeforesplång
43 Why Maxmum Margn for SVM? denotes +1 denotes -1 Support Vectors are those dataponts that the margn pushes up aganst 1. Intutvely ths feels safest.. If we ve made f(x,w,b) a small = error sgn(w. n x the - b) locaton of the boundary (t s been jolted n ts perpendcular The drecton) maxmum ths gves us least chance of causng margn a lnear msclassfcaton. classfer s the lnear classfer wth the, um, maxmum margn. 3. LOOCV s easy snce the model s mmune to removal of any non-supportvector dataponts. 4. There s some theory (usng VC dmenson) that s related to (but not the same as) the proposton that ths s a good thng. Ths s the smplest knd of SVM (Called an LSVM) 5. Emprcally t works very very well. 9/30/15 85 Dr. Yanjun Q / UVA CS 6316 / f15 Evaluaton Choce-III: BascsoluHon ForHWfQ more advanced soluhon ForHWfQ 9/30/15 APracHcalGudetoSupportVectorClassfcaHon 86
44 9/30/15 APracHcalGudetoSupportVector ClassfcaHon 87 Today:Revew&PracUcalGude " SupportVectorMachne(SVM) # HstoryofSVM # LargeMargnLnearClassfer # DefneMargn(M)ntermsofmodelparameter # OpHmzaHontolearnmodelparameters(w,b) # Nonlnearlyseparablecase # OpHmzaHonwthdualform # Nonlneardecsonboundary # PracHcalGude # Fleformat/LIBSVM # Featurepreprocsssng # ModelselecHon # Ppelneprocedure 9/30/15 88
45 References ElementsofStaHsHcalLearnng,byHasHe, TbshranandFredman ssldes TutoralsldesfromDr.TefYanLu,MSRAsa APracHcalGudetoSupportVectorClassfcaHon ChhfWeHsu,ChhfChungChang,andChhfJenLn, 003f010 TutoralsldesfromStanford ConvexOpHmzaHonI Boyd&Vandenberghe 9/30/15 89
UVA CS / Introduc8on to Machine Learning and Data Mining
UVA CS 4501-001 / 6501 007 Introduc8on to Machne Learnng and Data Mnng Lecture 11: Classfca8on wth Support Vector Machne (Revew + Prac8cal Gude) Yanjun Q / Jane Unversty of Vrgna Department of Computer
More informationUVA CS / Introduc8on to Machine Learning and Data Mining. Lecture 10: Classifica8on with Support Vector Machine (cont.
UVA CS 4501-001 / 6501 007 Introduc8on to Machne Learnng and Data Mnng Lecture 10: Classfca8on wth Support Vector Machne (cont. ) Yanjun Q / Jane Unversty of Vrgna Department of Computer Scence 9/6/14
More informationSupport Vector Machines
Separatng boundary, defned by w Support Vector Machnes CISC 5800 Professor Danel Leeds Separatng hyperplane splts class 0 and class 1 Plane s defned by lne w perpendcular to plan Is data pont x n class
More informationKernel Methods and SVMs Extension
Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general
More informationSupport Vector Machines
/14/018 Separatng boundary, defned by w Support Vector Machnes CISC 5800 Professor Danel Leeds Separatng hyperplane splts class 0 and class 1 Plane s defned by lne w perpendcular to plan Is data pont x
More informationLinear Classification, SVMs and Nearest Neighbors
1 CSE 473 Lecture 25 (Chapter 18) Lnear Classfcaton, SVMs and Nearest Neghbors CSE AI faculty + Chrs Bshop, Dan Klen, Stuart Russell, Andrew Moore Motvaton: Face Detecton How do we buld a classfer to dstngush
More informationWhich Separator? Spring 1
Whch Separator? 6.034 - Sprng 1 Whch Separator? Mamze the margn to closest ponts 6.034 - Sprng Whch Separator? Mamze the margn to closest ponts 6.034 - Sprng 3 Margn of a pont " # y (w $ + b) proportonal
More informationSupport Vector Machines. Vibhav Gogate The University of Texas at dallas
Support Vector Machnes Vbhav Gogate he Unversty of exas at dallas What We have Learned So Far? 1. Decson rees. Naïve Bayes 3. Lnear Regresson 4. Logstc Regresson 5. Perceptron 6. Neural networks 7. K-Nearest
More informationSVMs: Duality and Kernel Trick. SVMs as quadratic programs
11/17/9 SVMs: Dualt and Kernel rck Machne Learnng - 161 Geoff Gordon MroslavDudík [[[partl ased on sldes of Zv-Bar Joseph] http://.cs.cmu.edu/~ggordon/161/ Novemer 18 9 SVMs as quadratc programs o optmzaton
More informationLagrange Multipliers Kernel Trick
Lagrange Multplers Kernel Trck Ncholas Ruozz Unversty of Texas at Dallas Based roughly on the sldes of Davd Sontag General Optmzaton A mathematcal detour, we ll come back to SVMs soon! subject to: f x
More informationSVMs: Duality and Kernel Trick. SVMs as quadratic programs
/8/9 SVMs: Dualt and Kernel rck Machne Learnng - 6 Geoff Gordon MroslavDudík [[[partl ased on sldes of Zv-Bar Joseph] http://.cs.cmu.edu/~ggordon/6/ Novemer 8 9 SVMs as quadratc programs o optmzaton prolems:
More informationLecture 10 Support Vector Machines II
Lecture 10 Support Vector Machnes II 22 February 2016 Taylor B. Arnold Yale Statstcs STAT 365/665 1/28 Notes: Problem 3 s posted and due ths upcomng Frday There was an early bug n the fake-test data; fxed
More informationLecture 10 Support Vector Machines. Oct
Lecture 10 Support Vector Machnes Oct - 20-2008 Lnear Separators Whch of the lnear separators s optmal? Concept of Margn Recall that n Perceptron, we learned that the convergence rate of the Perceptron
More informationSupport Vector Machines CS434
Support Vector Machnes CS434 Lnear Separators Many lnear separators exst that perfectly classfy all tranng examples Whch of the lnear separators s the best? Intuton of Margn Consder ponts A, B, and C We
More informationCS 3710: Visual Recognition Classification and Detection. Adriana Kovashka Department of Computer Science January 13, 2015
CS 3710: Vsual Recognton Classfcaton and Detecton Adrana Kovashka Department of Computer Scence January 13, 2015 Plan for Today Vsual recognton bascs part 2: Classfcaton and detecton Adrana s research
More informationSupport Vector Machines
CS 2750: Machne Learnng Support Vector Machnes Prof. Adrana Kovashka Unversty of Pttsburgh February 17, 2016 Announcement Homework 2 deadlne s now 2/29 We ll have covered everythng you need today or at
More informationSupport Vector Machines CS434
Support Vector Machnes CS434 Lnear Separators Many lnear separators exst that perfectly classfy all tranng examples Whch of the lnear separators s the best? + + + + + + + + + Intuton of Margn Consder ponts
More informationChapter 6 Support vector machine. Séparateurs à vaste marge
Chapter 6 Support vector machne Séparateurs à vaste marge Méthode de classfcaton bnare par apprentssage Introdute par Vladmr Vapnk en 1995 Repose sur l exstence d un classfcateur lnéare Apprentssage supervsé
More informationUVA$CS$6316$$ $Fall$2015$Graduate:$$ Machine$Learning$$ $ $Lecture$15:$LogisAc$Regression$/$ GeneraAve$vs.$DiscriminaAve$$
Dr.YanjunQ/UVACS6316/f15 UVACS6316 Fall2015Graduate: MachneLearnng Lecture15:LogsAcRegresson/ GeneraAvevs.DscrmnaAve 10/21/15 Dr.YanjunQ UnverstyofVrgna Departmentof ComputerScence 1 Wherearewe?! FvemajorsecHonsofthscourse
More informationLecture 3: Dual problems and Kernels
Lecture 3: Dual problems and Kernels C4B Machne Learnng Hlary 211 A. Zsserman Prmal and dual forms Lnear separablty revsted Feature mappng Kernels for SVMs Kernel trck requrements radal bass functons SVM
More informationSupport Vector Machines
Support Vector Machnes Konstantn Tretyakov (kt@ut.ee) MTAT.03.227 Machne Learnng So far Supervsed machne learnng Lnear models Least squares regresson Fsher s dscrmnant, Perceptron, Logstc model Non-lnear
More informationSupport Vector Machines
Support Vector Machnes Konstantn Tretyakov (kt@ut.ee) MTAT.03.227 Machne Learnng So far So far Supervsed machne learnng Lnear models Non-lnear models Unsupervsed machne learnng Generc scaffoldng So far
More informationIntro to Visual Recognition
CS 2770: Computer Vson Intro to Vsual Recognton Prof. Adrana Kovashka Unversty of Pttsburgh February 13, 2018 Plan for today What s recognton? a.k.a. classfcaton, categorzaton Support vector machnes Separable
More informationNatural Language Processing and Information Retrieval
Natural Language Processng and Informaton Retreval Support Vector Machnes Alessandro Moschtt Department of nformaton and communcaton technology Unversty of Trento Emal: moschtt@ds.untn.t Summary Support
More informationKristin P. Bennett. Rensselaer Polytechnic Institute
Support Vector Machnes and Other Kernel Methods Krstn P. Bennett Mathematcal Scences Department Rensselaer Polytechnc Insttute Support Vector Machnes (SVM) A methodology for nference based on Statstcal
More informationImage classification. Given the bag-of-features representations of images from different classes, how do we learn a model for distinguishing i them?
Image classfcaton Gven te bag-of-features representatons of mages from dfferent classes ow do we learn a model for dstngusng tem? Classfers Learn a decson rule assgnng bag-offeatures representatons of
More information10-701/ Machine Learning, Fall 2005 Homework 3
10-701/15-781 Machne Learnng, Fall 2005 Homework 3 Out: 10/20/05 Due: begnnng of the class 11/01/05 Instructons Contact questons-10701@autonlaborg for queston Problem 1 Regresson and Cross-valdaton [40
More information1 Convex Optimization
Convex Optmzaton We wll consder convex optmzaton problems. Namely, mnmzaton problems where the objectve s convex (we assume no constrants for now). Such problems often arse n machne learnng. For example,
More informationKernels in Support Vector Machines. Based on lectures of Martin Law, University of Michigan
Kernels n Support Vector Machnes Based on lectures of Martn Law, Unversty of Mchgan Non Lnear separable problems AND OR NOT() The XOR problem cannot be solved wth a perceptron. XOR Per Lug Martell - Systems
More informationNonlinear Classifiers II
Nonlnear Classfers II Nonlnear Classfers: Introducton Classfers Supervsed Classfers Lnear Classfers Perceptron Least Squares Methods Lnear Support Vector Machne Nonlnear Classfers Part I: Mult Layer Neural
More information18-660: Numerical Methods for Engineering Design and Optimization
8-66: Numercal Methods for Engneerng Desgn and Optmzaton n L Department of EE arnege Mellon Unversty Pttsburgh, PA 53 Slde Overve lassfcaton Support vector machne Regularzaton Slde lassfcaton Predct categorcal
More informationADVANCED MACHINE LEARNING ADVANCED MACHINE LEARNING
1 ADVANCED ACHINE LEARNING ADVANCED ACHINE LEARNING Non-lnear regresson technques 2 ADVANCED ACHINE LEARNING Regresson: Prncple N ap N-dm. nput x to a contnuous output y. Learn a functon of the type: N
More informationEnsemble Methods: Boosting
Ensemble Methods: Boostng Ncholas Ruozz Unversty of Texas at Dallas Based on the sldes of Vbhav Gogate and Rob Schapre Last Tme Varance reducton va baggng Generate new tranng data sets by samplng wth replacement
More informationSupport Vector Machines. Jie Tang Knowledge Engineering Group Department of Computer Science and Technology Tsinghua University 2012
Support Vector Machnes Je Tang Knowledge Engneerng Group Department of Computer Scence and Technology Tsnghua Unversty 2012 1 Outlne What s a Support Vector Machne? Solvng SVMs Kernel Trcks 2 What s a
More informationCSE 252C: Computer Vision III
CSE 252C: Computer Vson III Lecturer: Serge Belonge Scrbe: Catherne Wah LECTURE 15 Kernel Machnes 15.1. Kernels We wll study two methods based on a specal knd of functon k(x, y) called a kernel: Kernel
More information17 Support Vector Machines
17 We now dscuss an nfluental and effectve classfcaton algorthm called (SVMs). In addton to ther successes n many classfcaton problems, SVMs are responsble for ntroducng and/or popularzng several mportant
More informationNeural networks. Nuno Vasconcelos ECE Department, UCSD
Neural networs Nuno Vasconcelos ECE Department, UCSD Classfcaton a classfcaton problem has two types of varables e.g. X - vector of observatons (features) n the world Y - state (class) of the world x X
More informationMLE and Bayesian Estimation. Jie Tang Department of Computer Science & Technology Tsinghua University 2012
MLE and Bayesan Estmaton Je Tang Department of Computer Scence & Technology Tsnghua Unversty 01 1 Lnear Regresson? As the frst step, we need to decde how we re gong to represent the functon f. One example:
More informationLecture 6: Support Vector Machines
Lecture 6: Support Vector Machnes Marna Melă mmp@stat.washngton.edu Department of Statstcs Unversty of Washngton November, 2018 Lnear SVM s The margn and the expected classfcaton error Maxmum Margn Lnear
More informationKernel Methods and SVMs
Statstcal Machne Learnng Notes 7 Instructor: Justn Domke Kernel Methods and SVMs Contents 1 Introducton 2 2 Kernel Rdge Regresson 2 3 The Kernel Trck 5 4 Support Vector Machnes 7 5 Examples 1 6 Kernel
More informationINF 5860 Machine learning for image classification. Lecture 3 : Image classification and regression part II Anne Solberg January 31, 2018
INF 5860 Machne learnng for mage classfcaton Lecture 3 : Image classfcaton and regresson part II Anne Solberg January 3, 08 Today s topcs Multclass logstc regresson and softma Regularzaton Image classfcaton
More informationBoostrapaggregating (Bagging)
Boostrapaggregatng (Baggng) An ensemble meta-algorthm desgned to mprove the stablty and accuracy of machne learnng algorthms Can be used n both regresson and classfcaton Reduces varance and helps to avod
More informationEvaluation of classifiers MLPs
Lecture Evaluaton of classfers MLPs Mlos Hausrecht mlos@cs.ptt.edu 539 Sennott Square Evaluaton For any data set e use to test the model e can buld a confuson matrx: Counts of examples th: class label
More informationFMA901F: Machine Learning Lecture 5: Support Vector Machines. Cristian Sminchisescu
FMA901F: Machne Learnng Lecture 5: Support Vector Machnes Crstan Smnchsescu Back to Bnary Classfcaton Setup We are gven a fnte, possbly nosy, set of tranng data:,, 1,..,. Each nput s pared wth a bnary
More informationAdvanced Introduction to Machine Learning
Advanced Introducton to Machne Learnng 10715, Fall 2014 The Kernel Trck, Reproducng Kernel Hlbert Space, and the Representer Theorem Erc Xng Lecture 6, September 24, 2014 Readng: Erc Xng @ CMU, 2014 1
More informationGeneralized Linear Methods
Generalzed Lnear Methods 1 Introducton In the Ensemble Methods the general dea s that usng a combnaton of several weak learner one could make a better learner. More formally, assume that we have a set
More informationLecture Notes on Linear Regression
Lecture Notes on Lnear Regresson Feng L fl@sdueducn Shandong Unversty, Chna Lnear Regresson Problem In regresson problem, we am at predct a contnuous target value gven an nput feature vector We assume
More informationPattern Classification
Pattern Classfcaton All materals n these sldes ere taken from Pattern Classfcaton (nd ed) by R. O. Duda, P. E. Hart and D. G. Stork, John Wley & Sons, 000 th the permsson of the authors and the publsher
More informationOnline Classification: Perceptron and Winnow
E0 370 Statstcal Learnng Theory Lecture 18 Nov 8, 011 Onlne Classfcaton: Perceptron and Wnnow Lecturer: Shvan Agarwal Scrbe: Shvan Agarwal 1 Introducton In ths lecture we wll start to study the onlne learnng
More informationMaximal Margin Classifier
CS81B/Stat41B: Advanced Topcs n Learnng & Decson Makng Mamal Margn Classfer Lecturer: Mchael Jordan Scrbes: Jana van Greunen Corrected verson - /1/004 1 References/Recommended Readng 1.1 Webstes www.kernel-machnes.org
More informationLogistic Classifier CISC 5800 Professor Daniel Leeds
lon 9/7/8 Logstc Classfer CISC 58 Professor Danel Leeds Classfcaton strategy: generatve vs. dscrmnatve Generatve, e.g., Bayes/Naïve Bayes: 5 5 Identfy probablty dstrbuton for each class Determne class
More informationFeature Selection: Part 1
CSE 546: Machne Learnng Lecture 5 Feature Selecton: Part 1 Instructor: Sham Kakade 1 Regresson n the hgh dmensonal settng How do we learn when the number of features d s greater than the sample sze n?
More informationC4B Machine Learning Answers II. = σ(z) (1 σ(z)) 1 1 e z. e z = σ(1 σ) (1 + e z )
C4B Machne Learnng Answers II.(a) Show that for the logstc sgmod functon dσ(z) dz = σ(z) ( σ(z)) A. Zsserman, Hlary Term 20 Start from the defnton of σ(z) Note that Then σ(z) = σ = dσ(z) dz = + e z e z
More informationFor now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.
Neural Networks : Dervaton compled by Alvn Wan from Professor Jtendra Malk s lecture Ths type of computaton s called deep learnng and s the most popular method for many problems, such as computer vson
More informationSELECTED SOLUTIONS, SECTION (Weak duality) Prove that the primal and dual values p and d defined by equations (4.3.2) and (4.3.3) satisfy p d.
SELECTED SOLUTIONS, SECTION 4.3 1. Weak dualty Prove that the prmal and dual values p and d defned by equatons 4.3. and 4.3.3 satsfy p d. We consder an optmzaton problem of the form The Lagrangan for ths
More informationLinear Approximation with Regularization and Moving Least Squares
Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...
More informationInstance-Based Learning (a.k.a. memory-based learning) Part I: Nearest Neighbor Classification
Instance-Based earnng (a.k.a. memory-based learnng) Part I: Nearest Neghbor Classfcaton Note to other teachers and users of these sldes. Andrew would be delghted f you found ths source materal useful n
More informationPairwise Multi-classification Support Vector Machines: Quadratic Programming (QP-P A MSVM) formulations
Proceedngs of the 6th WSEAS Int. Conf. on NEURAL NEWORKS, Lsbon, Portugal, June 6-8, 005 (pp05-0) Parwse Mult-classfcaton Support Vector Machnes: Quadratc Programmng (QP-P A MSVM) formulatons HEODORE B.
More informationThe exam is closed book, closed notes except your one-page cheat sheet.
CS 89 Fall 206 Introducton to Machne Learnng Fnal Do not open the exam before you are nstructed to do so The exam s closed book, closed notes except your one-page cheat sheet Usage of electronc devces
More informationMACHINE APPLIED MACHINE LEARNING LEARNING. Gaussian Mixture Regression
11 MACHINE APPLIED MACHINE LEARNING LEARNING MACHINE LEARNING Gaussan Mture Regresson 22 MACHINE APPLIED MACHINE LEARNING LEARNING Bref summary of last week s lecture 33 MACHINE APPLIED MACHINE LEARNING
More informationCS246: Mining Massive Datasets Jure Leskovec, Stanford University
CS246: Mnng Massve Datasets Jure Leskovec, Stanford Unversty http://cs246.stanford.edu 2/19/18 Jure Leskovec, Stanford CS246: Mnng Massve Datasets, http://cs246.stanford.edu 2 Hgh dm. data Graph data Infnte
More informationLogistic Regression. CAP 5610: Machine Learning Instructor: Guo-Jun QI
Logstc Regresson CAP 561: achne Learnng Instructor: Guo-Jun QI Bayes Classfer: A Generatve model odel the posteror dstrbuton P(Y X) Estmate class-condtonal dstrbuton P(X Y) for each Y Estmate pror dstrbuton
More informationSingular Value Decomposition: Theory and Applications
Sngular Value Decomposton: Theory and Applcatons Danel Khashab Sprng 2015 Last Update: March 2, 2015 1 Introducton A = UDV where columns of U and V are orthonormal and matrx D s dagonal wth postve real
More informationConvex Optimization. Optimality conditions. (EE227BT: UC Berkeley) Lecture 9 (Optimality; Conic duality) 9/25/14. Laurent El Ghaoui.
Convex Optmzaton (EE227BT: UC Berkeley) Lecture 9 (Optmalty; Conc dualty) 9/25/14 Laurent El Ghaou Organsatonal Mdterm: 10/7/14 (1.5 hours, n class, double-sded cheat sheet allowed) Project: Intal proposal
More informationStructural Extensions of Support Vector Machines. Mark Schmidt March 30, 2009
Structural Extensons of Support Vector Machnes Mark Schmdt March 30, 2009 Formulaton: Bnary SVMs Multclass SVMs Structural SVMs Tranng: Subgradents Cuttng Planes Margnal Formulatons Mn-Max Formulatons
More informationp 1 c 2 + p 2 c 2 + p 3 c p m c 2
Where to put a faclty? Gven locatons p 1,..., p m n R n of m houses, want to choose a locaton c n R n for the fre staton. Want c to be as close as possble to all the house. We know how to measure dstance
More informationSolutions to exam in SF1811 Optimization, Jan 14, 2015
Solutons to exam n SF8 Optmzaton, Jan 4, 25 3 3 O------O -4 \ / \ / The network: \/ where all lnks go from left to rght. /\ / \ / \ 6 O------O -5 2 4.(a) Let x = ( x 3, x 4, x 23, x 24 ) T, where the varable
More informationCSE 546 Midterm Exam, Fall 2014(with Solution)
CSE 546 Mdterm Exam, Fall 014(wth Soluton) 1. Personal nfo: Name: UW NetID: Student ID:. There should be 14 numbered pages n ths exam (ncludng ths cover sheet). 3. You can use any materal you brought:
More informationMMA and GCMMA two methods for nonlinear optimization
MMA and GCMMA two methods for nonlnear optmzaton Krster Svanberg Optmzaton and Systems Theory, KTH, Stockholm, Sweden. krlle@math.kth.se Ths note descrbes the algorthms used n the author s 2007 mplementatons
More information10) Activity analysis
3C3 Mathematcal Methods for Economsts (6 cr) 1) Actvty analyss Abolfazl Keshvar Ph.D. Aalto Unversty School of Busness Sldes orgnally by: Tmo Kuosmanen Updated by: Abolfazl Keshvar 1 Outlne Hstorcal development
More informationWeek 5: Neural Networks
Week 5: Neural Networks Instructor: Sergey Levne Neural Networks Summary In the prevous lecture, we saw how we can construct neural networks by extendng logstc regresson. Neural networks consst of multple
More informationU.C. Berkeley CS294: Beyond Worst-Case Analysis Luca Trevisan September 5, 2017
U.C. Berkeley CS94: Beyond Worst-Case Analyss Handout 4s Luca Trevsan September 5, 07 Summary of Lecture 4 In whch we ntroduce semdefnte programmng and apply t to Max Cut. Semdefnte Programmng Recall that
More informationCSC 411 / CSC D11 / CSC C11
18 Boostng s a general strategy for learnng classfers by combnng smpler ones. The dea of boostng s to take a weak classfer that s, any classfer that wll do at least slghtly better than chance and use t
More informationprinceton univ. F 17 cos 521: Advanced Algorithm Design Lecture 7: LP Duality Lecturer: Matt Weinberg
prnceton unv. F 17 cos 521: Advanced Algorthm Desgn Lecture 7: LP Dualty Lecturer: Matt Wenberg Scrbe: LP Dualty s an extremely useful tool for analyzng structural propertes of lnear programs. Whle there
More informationStructured Perceptrons & Structural SVMs
Structured Perceptrons Structural SVMs 4/6/27 CS 59: Advanced Topcs n Machne Learnng Recall: Sequence Predcton Input: x = (x,,x M ) Predct: y = (y,,y M ) Each y one of L labels. x = Fsh Sleep y = (N, V)
More informationLecture 12: Classification
Lecture : Classfcaton g Dscrmnant functons g The optmal Bayes classfer g Quadratc classfers g Eucldean and Mahalanobs metrcs g K Nearest Neghbor Classfers Intellgent Sensor Systems Rcardo Guterrez-Osuna
More informationVQ widely used in coding speech, image, and video
at Scalar quantzers are specal cases of vector quantzers (VQ): they are constraned to look at one sample at a tme (memoryless) VQ does not have such constrant better RD perfomance expected Source codng
More informationHomework Assignment 3 Due in class, Thursday October 15
Homework Assgnment 3 Due n class, Thursday October 15 SDS 383C Statstcal Modelng I 1 Rdge regresson and Lasso 1. Get the Prostrate cancer data from http://statweb.stanford.edu/~tbs/elemstatlearn/ datasets/prostate.data.
More informationCOS 521: Advanced Algorithms Game Theory and Linear Programming
COS 521: Advanced Algorthms Game Theory and Lnear Programmng Moses Charkar February 27, 2013 In these notes, we ntroduce some basc concepts n game theory and lnear programmng (LP). We show a connecton
More informationA Bayes Algorithm for the Multitask Pattern Recognition Problem Direct Approach
A Bayes Algorthm for the Multtask Pattern Recognton Problem Drect Approach Edward Puchala Wroclaw Unversty of Technology, Char of Systems and Computer etworks, Wybrzeze Wyspanskego 7, 50-370 Wroclaw, Poland
More informationMaxent Models & Deep Learning
Maxent Models & Deep Learnng 1. Last bts of maxent (sequence) models 1.MEMMs vs. CRFs 2.Smoothng/regularzaton n maxent models 2. Deep Learnng 1. What s t? Why s t good? (Part 1) 2. From logstc regresson
More informationThe conjugate prior to a Bernoulli is. A) Bernoulli B) Gaussian C) Beta D) none of the above
The conjugate pror to a Bernoull s A) Bernoull B) Gaussan C) Beta D) none of the above The conjugate pror to a Gaussan s A) Bernoull B) Gaussan C) Beta D) none of the above MAP estmates A) argmax θ p(θ
More informationLinear discriminants. Nuno Vasconcelos ECE Department, UCSD
Lnear dscrmnants Nuno Vasconcelos ECE Department UCSD Classfcaton a classfcaton problem as to tpes of varables e.g. X - vector of observatons features n te orld Y - state class of te orld X R 2 fever blood
More informationAssortment Optimization under MNL
Assortment Optmzaton under MNL Haotan Song Aprl 30, 2017 1 Introducton The assortment optmzaton problem ams to fnd the revenue-maxmzng assortment of products to offer when the prces of products are fxed.
More informationDepartment of Computer Science Artificial Intelligence Research Laboratory. Iowa State University MACHINE LEARNING
Iowa State Unversty Department of Computer Scence Artfcal Intellgence Research Laboratory MACHINE LEARNING Vasant Honavar Artfcal Intellgence Research Laboratory Department of Computer Scence Bonformatcs
More informationSolutions HW #2. minimize. Ax = b. Give the dual problem, and make the implicit equality constraints explicit. Solution.
Solutons HW #2 Dual of general LP. Fnd the dual functon of the LP mnmze subject to c T x Gx h Ax = b. Gve the dual problem, and make the mplct equalty constrants explct. Soluton. 1. The Lagrangan s L(x,
More informationWe present the algorithm first, then derive it later. Assume access to a dataset {(x i, y i )} n i=1, where x i R d and y i { 1, 1}.
CS 189 Introducton to Machne Learnng Sprng 2018 Note 26 1 Boostng We have seen that n the case of random forests, combnng many mperfect models can produce a snglodel that works very well. Ths s the dea
More informationLearning Theory: Lecture Notes
Learnng Theory: Lecture Notes Lecturer: Kamalka Chaudhur Scrbe: Qush Wang October 27, 2012 1 The Agnostc PAC Model Recall that one of the constrants of the PAC model s that the data dstrbuton has to be
More informationFisher Linear Discriminant Analysis
Fsher Lnear Dscrmnant Analyss Max Wellng Department of Computer Scence Unversty of Toronto 10 Kng s College Road Toronto, M5S 3G5 Canada wellng@cs.toronto.edu Abstract Ths s a note to explan Fsher lnear
More informationLearning from Data 1 Naive Bayes
Learnng from Data 1 Nave Bayes Davd Barber dbarber@anc.ed.ac.uk course page : http://anc.ed.ac.uk/ dbarber/lfd1/lfd1.html c Davd Barber 2001, 2002 1 Learnng from Data 1 : c Davd Barber 2001,2002 2 1 Why
More informationClassification as a Regression Problem
Target varable y C C, C,, ; Classfcaton as a Regresson Problem { }, 3 L C K To treat classfcaton as a regresson problem we should transform the target y nto numercal values; The choce of numercal class
More informationThe Gaussian classifier. Nuno Vasconcelos ECE Department, UCSD
he Gaussan classfer Nuno Vasconcelos ECE Department, UCSD Bayesan decson theory recall that we have state of the world X observatons g decson functon L[g,y] loss of predctng y wth g Bayes decson rule s
More informationMachine Learning: and 15781, 2003 Assignment 4
ahne Learnng: 070 and 578, 003 Assgnment 4. VC Dmenson 30 onts Consder the spae of nstane X orrespondng to all ponts n the D x, plane. Gve the VC dmenson of the followng hpothess spaes. No explanaton requred.
More informationLinear Feature Engineering 11
Lnear Feature Engneerng 11 2 Least-Squares 2.1 Smple least-squares Consder the followng dataset. We have a bunch of nputs x and correspondng outputs y. The partcular values n ths dataset are x y 0.23 0.19
More informationLecture 20: November 7
0-725/36-725: Convex Optmzaton Fall 205 Lecturer: Ryan Tbshran Lecture 20: November 7 Scrbes: Varsha Chnnaobreddy, Joon Sk Km, Lngyao Zhang Note: LaTeX template courtesy of UC Berkeley EECS dept. Dsclamer:
More informationEEE 241: Linear Systems
EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they
More informationProbabilistic Classification: Bayes Classifiers. Lecture 6:
Probablstc Classfcaton: Bayes Classfers Lecture : Classfcaton Models Sam Rowes January, Generatve model: p(x, y) = p(y)p(x y). p(y) are called class prors. p(x y) are called class condtonal feature dstrbutons.
More informationChapter 10 The Support-Vector-Machine (SVM) A statistical approach of learning theory for designing an optimal classifier
Chapter 0 The Support-Vector-Machne (SVM) A statstcal approach of learnng theory for desgnng an optmal classfer Content:. Problem 2. VC-Dmenson and mnmzaton of overall error 3. Lnear SVM Separable classes
More informationStatistical machine learning and its application to neonatal seizure detection
19/Oct/2009 Statstcal machne learnng and ts applcaton to neonatal sezure detecton Presented by Andry Temko Department of Electrcal and Electronc Engneerng Page 2 of 42 A. Temko, Statstcal Machne Learnng
More informationAPPENDIX A Some Linear Algebra
APPENDIX A Some Lnear Algebra The collecton of m, n matrces A.1 Matrces a 1,1,..., a 1,n A = a m,1,..., a m,n wth real elements a,j s denoted by R m,n. If n = 1 then A s called a column vector. Smlarly,
More information