Speech and Language Processing
|
|
- Stewart Stanley
- 5 years ago
- Views:
Transcription
1 Speech and Language rocessng Lecture 3 ayesan network and ayesan nference Informaton and ommuncatons Engneerng ourse Takahro Shnozak 08//5
2 Lecture lan (Shnozak s part) I gves the frst 6 lectures about speech recognton. Through these lectures the backbone of the latest speech recognton technques s eplaned.. 0/9 (remote) Speech recognton based on GMM HMM and N gram. 0/6 (remote) Mamum lkelhood estmaton and EM algorthm 3. /5 (@TIST) ayesan network and ayesan nference 4. /5 (@TIST) Varatonal nference and samplng 5. /6 (@TIST) Neural network based acoustc and language models 6. /6 (@TIST) Weghted fnte state transducer (WFST) and speech decodng
3 Today s Topc nswers for the prevous eercses ayesan network ayesan nference 3
4 nswers for the revous Eercses 4
5 Eercse. Show the dervaton process of obtanng 5 K k k k K k m k L log μ n k m k by mamzng 0 0 μ μ μ L L n k m k
6 Eercse. Derve the ML soluton { } of the Gaussan dstrbuton. The dervaton process must be descrbed 6 0 log 0 log n n N N n n n n
7 Eercse.3 Gven a tranng data D wth n tranng samples D={ n } obtan ML estmaton for GMM wth M mtures You can assume the varance s for smplcty 7 N M m m m N M m m m w w M M M ep log arg ma ep arg ma ˆ GMM No closed form soluton and we need EM
8 Eercse.4 ssume you have an ntal model parameter Θ 0. rove that f you take 0 0 then the lower bound 0 0 s equal to the log lkelhood log H J log H J H q H H q q J H log
9 Eercse.5 onsder the m GMM of the prevous page. Let m Θ 0 ). Obtans the followngs arg ma ˆ arg ma ˆ arg ma ˆ Q w Q Q w n w o o ˆ ˆ ˆ ˆ
10 ayesan Network 0
11 Graphs Undrected graph graph defned by nodes and undrected arcs Drected graph graph defned by nodes and a drected arcs Drected cyclc Graph: DG Drected graph that does not contan a drected cycle Eamples: Undrected graph Drected graph (Have a drected cycle) Drected acyclc graph
12 arent hld ncestor Descendant Node s a parent of node Node and are ancestors of node D Node s a chld of node D Node and D are descendant of node
13 partte When nodes of a graph are separated to two groups and there s no arc nsde the groups t s called a bpartte Eample of partte: 3
14 Drected Graph and Node Orderng drected graph s a DG Eq There s a orderng of nodes where all arcs face the same drecton (=There s a numberng of nodes where all arcs go from a lower numbered to hgher numbered nodes)
15 Outlne of the roof Statement : There s a orderng of nodes where all the arcs face the same drecton Statement : graph does not contan a drected cycle Easy Use lemma 3. (See append) 5
16 omment Feld 6
17 Eercse 3. Is the drected graph a DG? Graph Graph 7
18 ayesan Network (N) N s a graphcal model that represents a set of random varables and ther condtonal ndependence by DG D 8
19 Decomposton of Jont robablty and N y the product rule arbtrary jont probablty s decomposed to a product of condtonal probabltes ( D) D DG s made by Representng the varables as nodes onnectng the nodes by drected arcs accordng the condtonal probabltes D 9
20 ondtonal Independence and rcs ondtonal ndependence s represented by absence of arcs 0 D D ) ( ) ( ) ( ) ( ) ( D D D D ) ( D D
21 Jont robablty Defned by N roduct of condtonal probabltes assocated wth DG always satsfy the sum to one constrant roof: Snce a ayesan network s a DG wth a proper orderng of the varables the product has the followng form N { } That s does not appear n the condtonal part of. y thnkng the summaton of the followng order we have: N N N N N N N N
22 Eercse 3. Represent the followng jont probablty by a N ) ( ) ( D E D E D
23 N Representaton of a ategorcal Dstrbuton K p() μ μ μ K 3
24 N Representaton of a Gaussan Dstrbuton N ep 4
25 N Representaton of a GMM Gauss (μ σ ) Mture weght (robablty of nde) Gauss (μ σ ) Gauss3 (μ 3 σ 3 ) Gaussan dstrbuton condtoned by the nde Gauss ( ) 5
26 ayesan Network Representaton of a HMM The network has unrolled structure The length depends on the nput sequence a b Transton robablty e.g. (S t =as t- =b) HMM S S S 3 S T 3 T Emsson dstrbuton condtoned by the state e.g. ( t S t =a) ayesan network Tme 6
27 Eample of lgnment a b Feature sequence: State sequence: abaaabb S S S 3 S 7 = = 3 = 3 4 = 4 5 = 5 6 = 6 7 = S =a S =b S 3 =a S 4 =a S 5 =a S 6 =b S 7 =b 7
28 Representaton of a Repeated Structure V V t 3 T T 8
29 Representaton of arameters small crcles represent parameters Eample of GMM: 3 N S 3 N N Gauss 9
30 Eercse 3.3 Fll the blanks so that the followng HMM and the N become equvalent Intal state Intal state probablty (a) (b) (S t S t ) S t =a S t =b S t =a ( ) ( ) S t =b ( ) 0.6 a 0.4 b a a b b S S S 3 S T N a a HMM N b b N 3 T t s t s t N 30
31 Factor Graph bpartte graph where one sde of varables represent random varables and the others represent functons The arcs represent dependences of the functons to the varables factor graph defnes a jont probablty f N ssubsets of s varables s Eample: f f f 3 Factor nodes 3 4 Varable nodes f f f
32 Factor Graph Representaton of ayesan Network Each condtonal probablty can be regarded as a factor Eample D D ayesan network () () () (D) D Factor graph 3
33 robablstc Inference Margnal and condtonal probabltes are obtaned from a jont probablty by applyng the sum and product rules 33
34 Dstrbuton roperty and omputatonal ost roduct s dstrbutve over addton N af Number of products:n Number of summaton:n a N f Number of products: Number of summaton:n The same property holds for sum and ma and product and ma ma a f a f ma ma af a maf 34
35 omputatonal ost of Margnalzaton Suppose and D take 000 possble values D D # summaton = = 0 9 If the jont probablty s decomposed to: D ( ) ( ) ( ) ( D) D D D D # summaton = 30 3 Independence structure s mportant 35
36 When the Factor Graph s Lnear f f f 3 Suppose we want ( 3 ) f f f f f f f f f f f f 5 f 4
37 Message assng Vew of the Inference f f f f f 5 f 4 f f f f 3 f 3 f 3 3 f f f f 4 f f 4 f3 f 3 f 3 f 3 3 f f
38 ayesan Inference 38
39 robablstc Models and Ther arameters LM w M p Gaussan dstrbuton Multnomal dstrbuton etc. p w p w pw M Speech model consstng of language and acoustc models LM 39
40 ML Tranng and redcton N Tranng set D Test sample * arg ma D arg ma p p n n p * Mamum lkelhood (ML) tranng redcton 40
41 ayesan pproach Treat parameters as random varables 4 N Λ D p p D p p D p p D p D p D p redcton of a new sample s formulated as an evaluaton of condtonal probablty gven a tranng set D D p D
42 Defntons of Terms pror dstrbuton of parameters p robablstc model p posteror dstrbuton of parameters p D p D p D p redctve dstrbuton p D p p D 4
43 Evaluaton of osteror Dstrbuton Ecept for very smple models how to evaluate the a posteror dstrbuton s a bg ssue snce t requres ntegratons over many varables p D p D p p D 43
44 pproaches nalytcal evaluaton Ideal but only applcable for very smple models For practcal models closed form soluton s usually not obtaned. Numercal ntegraton s also not feasble when there are many varables Varatonal ayes an be appled to large models f proper analytcal appromaton s ntroduced Samplng Versatle but requres very large computatonal cost 44
45 onjugate ror For some combnatons of pror and probablstc model posteror takes the same functonal form as the pror robablstc model onjugate pror nomal dstrbuton eta dstrbuton Multnomal dstrbuton Gaussan dstrbuton Drchlet dstrbuton Mean: Gaussan dstrbuton Varance: Gamma dstrbuton 45
46 Eercse 3.4 ssumes a probablstc model (μ) a tranng sample and a pror dstrbuton of a parameter (μ) are gven as follows. Gaussan dstrbuton wth mean μ and varance ep μ ep Gaussan dstrbuton wth mean 0 and varance ) Estmate posteror dstrbuton ) Estmate predctve dstrbuton Note: ep c d c d 46
47 ppend 47
48 Lemma 3. If a graph does not contan a drected cycle then there est at least one node that has no ncomng arc? 48
49 Notaton for ondtonal Independence Let and be dsjont sets of random varables. When the followng equaton holds we say that s ndependent of gven and denote t as Note: 49
50 Graph Structure and ondtonal Independence y nvestgatng the graph structure we can read relatonshps between random varables D? D? 50
51 Tal To Tal Tal Tal In general () s not epressed as ()(). Therefor Φ does not hold. (Φ s an empty set) () s epressed as ()(). Therefore holds. 5
52 Head To Tal 5 Head Tal In general () s not epressed as ()(). Therefor Φ does not hold. () s epressed as ()(). Therefore holds.
53 Head To Head Head Head In general () s epressed as ()(). Therefor Φ holds. () s not epressed as ()(). Therefore does not hold. 53
54 lockng a ath For a ayesan network let and be a node and be a set of nodes that does not nclude and. We say a path from to s blocked when ether of the followngs holds On the path from to there s a node n and the connecton of the arcs s tal to tal or head to tal t one of the nodes on the path from to the connecton of the arcs s head to head. In addton the node and ts all descendants are not ncluded n D E F lock lock 54
55 d separaton For a ayesan network let and be eclusve sets of nodes We say s d separated from by f all the paths startng from a node n and endng at a node n s blocked When s d separated from by holds for the jont probablty defned by the ayesan network (earl 988) 55
56 Mamzaton of Jont robablty Obtaned by replacng Σ n the sum product algorthm wth ma Ma product lgorthm 56 N N N arg ma
57 Sum roduct lgorthm (for Tree) Message passng Leaf nodes Varable node: Factor node: f f f Varable node to factor node: f f Factor node to varable node: Margnal probablty f f f M M M m m f m 57
58 EM for HMM and Effcency Q HMM K log HMM K 0 0 K K k k k T T () a b The summaton s over state sequence K The number of the sequences s eponental to the length of nput sec of feature sequence s 00 frames Drectly enumeratng all the paths s mpossble Q functon () can be effcently evaluated f posterors k t s 0 and k s' k s are obtaned where s and s are HMM state ID t t 0 Use the Sum roduct algorthm to effcently obtan the posterors 58
59 59
60 Eercse 3.3 (nswer) 60 ep 4 ep ep ep ep ep ep N d d 3 3 ep 3 N d
Maximum Likelihood Estimation of Binary Dependent Variables Models: Probit and Logit. 1. General Formulation of Binary Dependent Variables Models
ECO 452 -- OE 4: Probt and Logt Models ECO 452 -- OE 4 Mamum Lkelhood Estmaton of Bnary Dependent Varables Models: Probt and Logt hs note demonstrates how to formulate bnary dependent varables models for
More informationProbabilistic Graphical Models
School of Computer Scence robablstc Graphcal Models Appromate Inference: Markov Chan Monte Carlo 05 07 Erc Xng Lecture 7 March 9 04 X X 075 05 05 03 X 3 Erc Xng @ CMU 005-04 Recap of Monte Carlo Monte
More informationCIS526: Machine Learning Lecture 3 (Sept 16, 2003) Linear Regression. Preparation help: Xiaoying Huang. x 1 θ 1 output... θ M x M
CIS56: achne Learnng Lecture 3 (Sept 6, 003) Preparaton help: Xaoyng Huang Lnear Regresson Lnear regresson can be represented by a functonal form: f(; θ) = θ 0 0 +θ + + θ = θ = 0 ote: 0 s a dummy attrbute
More informationCourse 395: Machine Learning - Lectures
Course 395: Machne Learnng - Lectures Lecture 1-2: Concept Learnng (M. Pantc Lecture 3-4: Decson Trees & CC Intro (M. Pantc Lecture 5-6: Artfcal Neural Networks (S.Zaferou Lecture 7-8: Instance ased Learnng
More informationAn Experiment/Some Intuition (Fall 2006): Lecture 18 The EM Algorithm heads coin 1 tails coin 2 Overview Maximum Likelihood Estimation
An Experment/Some Intuton I have three cons n my pocket, 6.864 (Fall 2006): Lecture 18 The EM Algorthm Con 0 has probablty λ of heads; Con 1 has probablty p 1 of heads; Con 2 has probablty p 2 of heads
More informationHidden Markov Models
CM229S: Machne Learnng for Bonformatcs Lecture 12-05/05/2016 Hdden Markov Models Lecturer: Srram Sankararaman Scrbe: Akshay Dattatray Shnde Edted by: TBD 1 Introducton For a drected graph G we can wrte
More informationHidden Markov Models & The Multivariate Gaussian (10/26/04)
CS281A/Stat241A: Statstcal Learnng Theory Hdden Markov Models & The Multvarate Gaussan (10/26/04) Lecturer: Mchael I. Jordan Scrbes: Jonathan W. Hu 1 Hdden Markov Models As a bref revew, hdden Markov models
More informationExpectation Maximization Mixture Models HMMs
-755 Machne Learnng for Sgnal Processng Mture Models HMMs Class 9. 2 Sep 200 Learnng Dstrbutons for Data Problem: Gven a collecton of eamples from some data, estmate ts dstrbuton Basc deas of Mamum Lelhood
More informationMixture of Gaussians Expectation Maximization (EM) Part 2
Mture of Gaussans Eectaton Mamaton EM Part 2 Most of the sldes are due to Chrstoher Bsho BCS Summer School Eeter 2003. The rest of the sldes are based on lecture notes by A. Ng Lmtatons of K-means Hard
More informationEM and Structure Learning
EM and Structure Learnng Le Song Machne Learnng II: Advanced Topcs CSE 8803ML, Sprng 2012 Partally observed graphcal models Mxture Models N(μ 1, Σ 1 ) Z X N N(μ 2, Σ 2 ) 2 Gaussan mxture model Consder
More informationHidden Markov Models
Hdden Markov Models Namrata Vaswan, Iowa State Unversty Aprl 24, 204 Hdden Markov Model Defntons and Examples Defntons:. A hdden Markov model (HMM) refers to a set of hdden states X 0, X,..., X t,...,
More informationDepartment of Computer Science Artificial Intelligence Research Laboratory. Iowa State University MACHINE LEARNING
MACHINE LEANING Vasant Honavar Bonformatcs and Computatonal Bology rogram Center for Computatonal Intellgence, Learnng, & Dscovery Iowa State Unversty honavar@cs.astate.edu www.cs.astate.edu/~honavar/
More informationWhy BP Works STAT 232B
Why BP Works STAT 232B Free Energes Helmholz & Gbbs Free Energes 1 Dstance between Probablstc Models - K-L dvergence b{ KL b{ p{ = b{ ln { } p{ Here, p{ s the eact ont prob. b{ s the appromaton, called
More informationMaximum Likelihood Estimation of Binary Dependent Variables Models: Probit and Logit. 1. General Formulation of Binary Dependent Variables Models
ECO 452 -- OE 4: Probt and Logt Models ECO 452 -- OE 4 Maxmum Lkelhood Estmaton of Bnary Dependent Varables Models: Probt and Logt hs note demonstrates how to formulate bnary dependent varables models
More informationMachine Learning for Signal Processing Linear Gaussian Models
Machne Learnng for Sgnal rocessng Lnear Gaussan Models lass 2. 2 Nov 203 Instructor: Bhsha Raj 2 Nov 203 755/8797 HW3 s up. Admnstrva rojects please send us an update 2 Nov 203 755/8797 2 Recap: MA stmators
More informationPhysical Fluctuomatics Applied Stochastic Process 9th Belief propagation
Physcal luctuomatcs ppled Stochastc Process 9th elef propagaton Kazuyuk Tanaka Graduate School of Informaton Scences Tohoku Unversty kazu@smapp.s.tohoku.ac.jp http://www.smapp.s.tohoku.ac.jp/~kazu/ Stochastc
More informationOutline. Bayesian Networks: Maximum Likelihood Estimation and Tree Structure Learning. Our Model and Data. Outline
Outlne Bayesan Networks: Maxmum Lkelhood Estmaton and Tree Structure Learnng Huzhen Yu janey.yu@cs.helsnk.f Dept. Computer Scence, Unv. of Helsnk Probablstc Models, Sprng, 200 Notces: I corrected a number
More informationProbability-Theoretic Junction Trees
Probablty-Theoretc Juncton Trees Payam Pakzad, (wth Venkat Anantharam, EECS Dept, U.C. Berkeley EPFL, ALGO/LMA Semnar 2/2/2004 Margnalzaton Problem Gven an arbtrary functon of many varables, fnd (some
More informationMATH 829: Introduction to Data Mining and Analysis The EM algorithm (part 2)
1/16 MATH 829: Introducton to Data Mnng and Analyss The EM algorthm (part 2) Domnque Gullot Departments of Mathematcal Scences Unversty of Delaware Aprl 20, 2016 Recall 2/16 We are gven ndependent observatons
More informationxp(x µ) = 0 p(x = 0 µ) + 1 p(x = 1 µ) = µ
CSE 455/555 Sprng 2013 Homework 7: Parametrc Technques Jason J. Corso Computer Scence and Engneerng SUY at Buffalo jcorso@buffalo.edu Solutons by Yngbo Zhou Ths assgnment does not need to be submtted and
More informationArtificial Intelligence Bayesian Networks
Artfcal Intellgence Bayesan Networks Adapted from sldes by Tm Fnn and Mare desjardns. Some materal borrowed from Lse Getoor. 1 Outlne Bayesan networks Network structure Condtonal probablty tables Condtonal
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 informationCS 2750 Machine Learning. Lecture 5. Density estimation. CS 2750 Machine Learning. Announcements
CS 750 Machne Learnng Lecture 5 Densty estmaton Mlos Hauskrecht mlos@cs.ptt.edu 539 Sennott Square CS 750 Machne Learnng Announcements Homework Due on Wednesday before the class Reports: hand n before
More informationComputation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models
Computaton of Hgher Order Moments from Two Multnomal Overdsperson Lkelhood Models BY J. T. NEWCOMER, N. K. NEERCHAL Department of Mathematcs and Statstcs, Unversty of Maryland, Baltmore County, Baltmore,
More informationParametric fractional imputation for missing data analysis. Jae Kwang Kim Survey Working Group Seminar March 29, 2010
Parametrc fractonal mputaton for mssng data analyss Jae Kwang Km Survey Workng Group Semnar March 29, 2010 1 Outlne Introducton Proposed method Fractonal mputaton Approxmaton Varance estmaton Multple mputaton
More informatione i is a random error
Chapter - The Smple Lnear Regresson Model The lnear regresson equaton s: where + β + β e for,..., and are observable varables e s a random error How can an estmaton rule be constructed for the unknown
More informationGenerative and Discriminative Models. Jie Tang Department of Computer Science & Technology Tsinghua University 2012
Generatve and Dscrmnatve Models Je Tang Department o Computer Scence & Technolog Tsnghua Unverst 202 ML as Searchng Hpotheses Space ML Methodologes are ncreasngl statstcal Rule-based epert sstems beng
More informationAppendix for Causal Interaction in Factorial Experiments: Application to Conjoint Analysis
A Appendx for Causal Interacton n Factoral Experments: Applcaton to Conjont Analyss Mathematcal Appendx: Proofs of Theorems A. Lemmas Below, we descrbe all the lemmas, whch are used to prove the man theorems
More informationMarkov Chain Monte Carlo (MCMC), Gibbs Sampling, Metropolis Algorithms, and Simulated Annealing Bioinformatics Course Supplement
Markov Chan Monte Carlo MCMC, Gbbs Samplng, Metropols Algorthms, and Smulated Annealng 2001 Bonformatcs Course Supplement SNU Bontellgence Lab http://bsnuackr/ Outlne! Markov Chan Monte Carlo MCMC! Metropols-Hastngs
More informationCHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE
CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE Analytcal soluton s usually not possble when exctaton vares arbtrarly wth tme or f the system s nonlnear. Such problems can be solved by numercal tmesteppng
More informationChapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise.
Chapter - The Smple Lnear Regresson Model The lnear regresson equaton s: where y + = β + β e for =,..., y and are observable varables e s a random error How can an estmaton rule be constructed for the
More informationBayesian belief networks
CS 1571 Introducton to I Lecture 24 ayesan belef networks los Hauskrecht mlos@cs.ptt.edu 5329 Sennott Square CS 1571 Intro to I dmnstraton Homework assgnment 10 s out and due next week Fnal exam: December
More information2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification
E395 - Pattern Recognton Solutons to Introducton to Pattern Recognton, Chapter : Bayesan pattern classfcaton Preface Ths document s a soluton manual for selected exercses from Introducton to Pattern Recognton
More informationIntroduction to Hidden Markov Models
Introducton to Hdden Markov Models Alperen Degrmenc Ths document contans dervatons and algorthms for mplementng Hdden Markov Models. The content presented here s a collecton of my notes and personal nsghts
More information6. Stochastic processes (2)
Contents Markov processes Brth-death processes Lect6.ppt S-38.45 - Introducton to Teletraffc Theory Sprng 5 Markov process Consder a contnuous-tme and dscrete-state stochastc process X(t) wth state space
More information6. Stochastic processes (2)
6. Stochastc processes () Lect6.ppt S-38.45 - Introducton to Teletraffc Theory Sprng 5 6. Stochastc processes () Contents Markov processes Brth-death processes 6. Stochastc processes () Markov process
More informationMean Field / Variational Approximations
Mean Feld / Varatonal Appromatons resented by Jose Nuñez 0/24/05 Outlne Introducton Mean Feld Appromaton Structured Mean Feld Weghted Mean Feld Varatonal Methods Introducton roblem: We have dstrbuton but
More informationThe Expectation-Maximization Algorithm
The Expectaton-Maxmaton Algorthm Charles Elan elan@cs.ucsd.edu November 16, 2007 Ths chapter explans the EM algorthm at multple levels of generalty. Secton 1 gves the standard hgh-level verson of the algorthm.
More informationConjugacy and the Exponential Family
CS281B/Stat241B: Advanced Topcs n Learnng & Decson Makng Conjugacy and the Exponental Famly Lecturer: Mchael I. Jordan Scrbes: Bran Mlch 1 Conjugacy In the prevous lecture, we saw conjugate prors for the
More information4DVAR, according to the name, is a four-dimensional variational method.
4D-Varatonal Data Assmlaton (4D-Var) 4DVAR, accordng to the name, s a four-dmensonal varatonal method. 4D-Var s actually a drect generalzaton of 3D-Var to handle observatons that are dstrbuted n tme. The
More informationEvaluation for sets of classes
Evaluaton for Tet Categorzaton Classfcaton accuracy: usual n ML, the proporton of correct decsons, Not approprate f the populaton rate of the class s low Precson, Recall and F 1 Better measures 21 Evaluaton
More informationInner Product. Euclidean Space. Orthonormal Basis. Orthogonal
Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,
More informationHidden Markov Model Cheat Sheet
Hdden Markov Model Cheat Sheet (GIT ID: dc2f391536d67ed5847290d5250d4baae103487e) Ths document s a cheat sheet on Hdden Markov Models (HMMs). It resembles lecture notes, excet that t cuts to the chase
More informationModule 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur
Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:
More informationBayesian predictive Configural Frequency Analysis
Psychologcal Test and Assessment Modelng, Volume 54, 2012 (3), 285-292 Bayesan predctve Confgural Frequency Analyss Eduardo Gutérrez-Peña 1 Abstract Confgural Frequency Analyss s a method for cell-wse
More informationRelevance Vector Machines Explained
October 19, 2010 Relevance Vector Machnes Explaned Trstan Fletcher www.cs.ucl.ac.uk/staff/t.fletcher/ Introducton Ths document has been wrtten n an attempt to make Tppng s [1] Relevance Vector Machnes
More informationMARKOV CHAIN AND HIDDEN MARKOV MODEL
MARKOV CHAIN AND HIDDEN MARKOV MODEL JIAN ZHANG JIANZHAN@STAT.PURDUE.EDU Markov chan and hdden Markov mode are probaby the smpest modes whch can be used to mode sequenta data,.e. data sampes whch are not
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 information3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X
Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number
More informationCell Biology. Lecture 1: 10-Oct-12. Marco Grzegorczyk. (Gen-)Regulatory Network. Microarray Chips. (Gen-)Regulatory Network. (Gen-)Regulatory Network
5.0.202 Genetsche Netzwerke Wntersemester 202/203 ell ology Lecture : 0-Oct-2 Marco Grzegorczyk Gen-Regulatory Network Mcroarray hps G G 2 G 3 2 3 metabolte metabolte Gen-Regulatory Network Gen-Regulatory
More informationGenerative classification models
CS 675 Intro to Machne Learnng Lecture Generatve classfcaton models Mlos Hauskrecht mlos@cs.ptt.edu 539 Sennott Square Data: D { d, d,.., dn} d, Classfcaton represents a dscrete class value Goal: learn
More informationCollege of Computer & Information Science Fall 2009 Northeastern University 20 October 2009
College of Computer & Informaton Scence Fall 2009 Northeastern Unversty 20 October 2009 CS7880: Algorthmc Power Tools Scrbe: Jan Wen and Laura Poplawsk Lecture Outlne: Prmal-dual schema Network Desgn:
More informationSTATS 306B: Unsupervised Learning Spring Lecture 10 April 30
STATS 306B: Unsupervsed Learnng Sprng 2014 Lecture 10 Aprl 30 Lecturer: Lester Mackey Scrbe: Joey Arthur, Rakesh Achanta 10.1 Factor Analyss 10.1.1 Recap Recall the factor analyss (FA) model for lnear
More informationStatistical learning
Statstcal learnng Model the data generaton process Learn the model parameters Crteron to optmze: Lkelhood of the dataset (maxmzaton) Maxmum Lkelhood (ML) Estmaton: Dataset X Statstcal model p(x;θ) (θ parameters)
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 informationNP-Completeness : Proofs
NP-Completeness : Proofs Proof Methods A method to show a decson problem Π NP-complete s as follows. (1) Show Π NP. (2) Choose an NP-complete problem Π. (3) Show Π Π. A method to show an optmzaton problem
More informationWhat Independencies does a Bayes Net Model? Bayesian Networks: Independencies and Inference. Quick proof that independence is symmetric
Bayesan Networks: Indeendences and Inference Scott Daves and ndrew Moore Note to other teachers and users of these sldes. ndrew and Scott would be delghted f you found ths source materal useful n gvng
More informationMachine learning: Density estimation
CS 70 Foundatons of AI Lecture 3 Machne learnng: ensty estmaton Mlos Hauskrecht mlos@cs.ptt.edu 539 Sennott Square ata: ensty estmaton {.. n} x a vector of attrbute values Objectve: estmate the model of
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 informationThe Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction
ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also
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 informationEstimation: Part 2. Chapter GREG estimation
Chapter 9 Estmaton: Part 2 9. GREG estmaton In Chapter 8, we have seen that the regresson estmator s an effcent estmator when there s a lnear relatonshp between y and x. In ths chapter, we generalzed the
More informationStanford University CS254: Computational Complexity Notes 7 Luca Trevisan January 29, Notes for Lecture 7
Stanford Unversty CS54: Computatonal Complexty Notes 7 Luca Trevsan January 9, 014 Notes for Lecture 7 1 Approxmate Countng wt an N oracle We complete te proof of te followng result: Teorem 1 For every
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 informationChapter 3. Two-Variable Regression Model: The Problem of Estimation
Chapter 3. Two-Varable Regresson Model: The Problem of Estmaton Ordnary Least Squares Method (OLS) Recall that, PRF: Y = β 1 + β X + u Thus, snce PRF s not drectly observable, t s estmated by SRF; that
More informationRELIABILITY ASSESSMENT
CHAPTER Rsk Analyss n Engneerng and Economcs RELIABILITY ASSESSMENT A. J. Clark School of Engneerng Department of Cvl and Envronmental Engneerng 4a CHAPMAN HALL/CRC Rsk Analyss for Engneerng Department
More informationLimited Dependent Variables
Lmted Dependent Varables. What f the left-hand sde varable s not a contnuous thng spread from mnus nfnty to plus nfnty? That s, gven a model = f (, β, ε, where a. s bounded below at zero, such as wages
More informationEconomics 130. Lecture 4 Simple Linear Regression Continued
Economcs 130 Lecture 4 Contnued Readngs for Week 4 Text, Chapter and 3. We contnue wth addressng our second ssue + add n how we evaluate these relatonshps: Where do we get data to do ths analyss? How do
More information6 Supplementary Materials
6 Supplementar Materals 61 Proof of Theorem 31 Proof Let m Xt z 1:T : l m Xt X,z 1:t Wethenhave mxt z1:t ˆm HX Xt z 1:T mxt z1:t m HX Xt z 1:T + mxt z 1:T HX We consder each of the two terms n equaton
More informationGaussian process classification: a message-passing viewpoint
Gaussan process classfcaton: a message-passng vewpont Flpe Rodrgues fmpr@de.uc.pt November 014 Abstract The goal of ths short paper s to provde a message-passng vewpont of the Expectaton Propagaton EP
More informationA new Approach for Solving Linear Ordinary Differential Equations
, ISSN 974-57X (Onlne), ISSN 974-5718 (Prnt), Vol. ; Issue No. 1; Year 14, Copyrght 13-14 by CESER PUBLICATIONS A new Approach for Solvng Lnear Ordnary Dfferental Equatons Fawz Abdelwahd Department of
More informationDigital Signal Processing
Dgtal Sgnal Processng Dscrete-tme System Analyss Manar Mohasen Offce: F8 Emal: manar.subh@ut.ac.r School of IT Engneerng Revew of Precedent Class Contnuous Sgnal The value of the sgnal s avalable over
More informationConditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data
Condtonal Random Felds: Probablstc Models for Segmentng and Labelng Sequence Data Paper by John Lafferty, Andrew McCallum, and Fernando Perera ICML 2001 Presentaton by Joe Drsh May 9, 2002 Man Goals Present
More informationGaussian Mixture Models
Lab Gaussan Mxture Models Lab Objectve: Understand the formulaton of Gaussan Mxture Models (GMMs) and how to estmate GMM parameters. You ve already seen GMMs as the observaton dstrbuton n certan contnuous
More informationOPTIMISATION. Introduction Single Variable Unconstrained Optimisation Multivariable Unconstrained Optimisation Linear Programming
OPTIMIATION Introducton ngle Varable Unconstraned Optmsaton Multvarable Unconstraned Optmsaton Lnear Programmng Chapter Optmsaton /. Introducton In an engneerng analss, sometmes etremtes, ether mnmum or
More informationRetrieval Models: Language models
CS-590I Informaton Retreval Retreval Models: Language models Luo S Department of Computer Scence Purdue Unversty Introducton to language model Ungram language model Document language model estmaton Maxmum
More informationAPPROXIMATE PRICES OF BASKET AND ASIAN OPTIONS DUPONT OLIVIER. Premia 14
APPROXIMAE PRICES OF BASKE AND ASIAN OPIONS DUPON OLIVIER Prema 14 Contents Introducton 1 1. Framewor 1 1.1. Baset optons 1.. Asan optons. Computng the prce 3. Lower bound 3.1. Closed formula for the prce
More informationOverview. Hidden Markov Models and Gaussian Mixture Models. Acoustic Modelling. Fundamental Equation of Statistical Speech Recognition
Overvew Hdden Marov Models and Gaussan Mxture Models Steve Renals and Peter Bell Automatc Speech Recognton ASR Lectures &5 8/3 January 3 HMMs and GMMs Key models and algorthms for HMM acoustc models Gaussans
More informationME 501A Seminar in Engineering Analysis Page 1
umercal Solutons of oundary-value Problems n Os ovember 7, 7 umercal Solutons of oundary- Value Problems n Os Larry aretto Mechancal ngneerng 5 Semnar n ngneerng nalyss ovember 7, 7 Outlne Revew stff equaton
More informationHidden Markov Models
Note to other teachers and users of these sldes. Andrew would be delghted f you found ths source materal useful n gvng your own lectures. Feel free to use these sldes verbatm, or to modfy them to ft your
More informationENG 8801/ Special Topics in Computer Engineering: Pattern Recognition. Memorial University of Newfoundland Pattern Recognition
EG 880/988 - Specal opcs n Computer Engneerng: Pattern Recognton Memoral Unversty of ewfoundland Pattern Recognton Lecture 7 May 3, 006 http://wwwengrmunca/~charlesr Offce Hours: uesdays hursdays 8:30-9:30
More informationDiscriminative classifier: Logistic Regression. CS534-Machine Learning
Dscrmnatve classfer: Logstc Regresson CS534-Machne Learnng robablstc Classfer Gven an nstance, hat does a probablstc classfer do dfferentl compared to, sa, perceptron? It does not drectl predct Instead,
More informationLecture 3. Ax x i a i. i i
18.409 The Behavor of Algorthms n Practce 2/14/2 Lecturer: Dan Spelman Lecture 3 Scrbe: Arvnd Sankar 1 Largest sngular value In order to bound the condton number, we need an upper bound on the largest
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 informationThe Feynman path integral
The Feynman path ntegral Aprl 3, 205 Hesenberg and Schrödnger pctures The Schrödnger wave functon places the tme dependence of a physcal system n the state, ψ, t, where the state s a vector n Hlbert space
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 informationA quantum-statistical-mechanical extension of Gaussian mixture model
A quantum-statstcal-mechancal extenson of Gaussan mxture model Kazuyuk Tanaka, and Koj Tsuda 2 Graduate School of Informaton Scences, Tohoku Unversty, 6-3-09 Aramak-aza-aoba, Aoba-ku, Senda 980-8579, Japan
More informationProbability Theory. The nth coefficient of the Taylor series of f(k), expanded around k = 0, gives the nth moment of x as ( ik) n n!
8333: Statstcal Mechancs I Problem Set # 3 Solutons Fall 3 Characterstc Functons: Probablty Theory The characterstc functon s defned by fk ep k = ep kpd The nth coeffcent of the Taylor seres of fk epanded
More informationChapter 1. Probability
Chapter. Probablty Mcroscopc propertes of matter: quantum mechancs, atomc and molecular propertes Macroscopc propertes of matter: thermodynamcs, E, H, C V, C p, S, A, G How do we relate these two propertes?
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 informationLINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity
LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have
More informationApproximate Inference: Mean Field Methods
School of Comuter Scence Aromate Inference: Mean Feld Methods Probablstc Grahcal Models 10-708 Lecture 17 Nov 12 2007 Recetor A Knase C Gene G Recetor B X 1 X 2 Knase D Knase X 3 X 4 X 5 TF F X 6 Gene
More informationβ0 + β1xi and want to estimate the unknown
SLR Models Estmaton Those OLS Estmates Estmators (e ante) v. estmates (e post) The Smple Lnear Regresson (SLR) Condtons -4 An Asde: The Populaton Regresson Functon B and B are Lnear Estmators (condtonal
More informationMIMA Group. Chapter 2 Bayesian Decision Theory. School of Computer Science and Technology, Shandong University. Xin-Shun SDU
Group M D L M Chapter Bayesan Decson heory Xn-Shun Xu @ SDU School of Computer Scence and echnology, Shandong Unversty Bayesan Decson heory Bayesan decson theory s a statstcal approach to data mnng/pattern
More informationLECTURE :FACTOR ANALYSIS
LCUR :FACOR ANALYSIS Rta Osadchy Based on Lecture Notes by A. Ng Motvaton Dstrbuton coes fro MoG Have suffcent aount of data: >>n denson Use M to ft Mture of Gaussans nu. of tranng ponts If
More information8 : Learning in Fully Observed Markov Networks. 1 Why We Need to Learn Undirected Graphical Models. 2 Structural Learning for Completely Observed MRF
10-708: Probablstc Graphcal Models 10-708, Sprng 2014 8 : Learnng n Fully Observed Markov Networks Lecturer: Erc P. Xng Scrbes: Meng Song, L Zhou 1 Why We Need to Learn Undrected Graphcal Models In the
More informationGames of Threats. Elon Kohlberg Abraham Neyman. Working Paper
Games of Threats Elon Kohlberg Abraham Neyman Workng Paper 18-023 Games of Threats Elon Kohlberg Harvard Busness School Abraham Neyman The Hebrew Unversty of Jerusalem Workng Paper 18-023 Copyrght 2017
More informationTime-Varying Systems and Computations Lecture 6
Tme-Varyng Systems and Computatons Lecture 6 Klaus Depold 14. Januar 2014 The Kalman Flter The Kalman estmaton flter attempts to estmate the actual state of an unknown dscrete dynamcal system, gven nosy
More informationCHAPTER 3: BAYESIAN DECISION THEORY
HATER 3: BAYESIAN DEISION THEORY Decson mang under uncertanty 3 Data comes from a process that s completely not nown The lac of nowledge can be compensated by modelng t as a random process May be the underlyng
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 information