Machine Learning. Inference and Learning in GM. Eric Xing , Fall Lecture 18, November 10, b r a c e

Similar documents
Exact Inference. Kayhan Batmanghelich

Probabilistic Graphical Models

Machine Learning. Graphical Models and Exact Inference. Eric Xing , Fall Lecture 17, November 7, Receptor A X 2.

Machine Learning. Recap of Basic Prob. Concepts

Why the Junction Tree Algorithm? The Junction Tree Algorithm. Clique Potential Representation. Overview. Chris Williams 1.

, each of which is a tree, and whose roots r 1. , respectively, are children of r. Data Structures & File Management

Problem 1. Solution: = show that for a constant number of particles: c and V. a) Using the definitions of P

Tangram Fractions Overview: Students will analyze standard and nonstandard

An undirected graph G = (V, E) V a set of vertices E a set of unordered edges (v,w) where v, w in V

16.unified Introduction to Computers and Programming. SOLUTIONS to Examination 4/30/04 9:05am - 10:00am

d e c b a d c b a d e c b a a c a d c c e b

Cycles and Simple Cycles. Paths and Simple Paths. Trees. Problem: There is No Completely Standard Terminology!

Math 166 Week in Review 2 Sections 1.1b, 1.2, 1.3, & 1.4

Having a glimpse of some of the possibilities for solutions of linear systems, we move to methods of finding these solutions. The basic idea we shall

5/1/2018. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees. Huffman Coding Trees

CSE 373: More on graphs; DFS and BFS. Michael Lee Wednesday, Feb 14, 2018

Bayesian belief networks

Weighted Graphs. Weighted graphs may be either directed or undirected.

CSE 373. Graphs 1: Concepts, Depth/Breadth-First Search reading: Weiss Ch. 9. slides created by Marty Stepp

Theorem 1. An undirected graph is a tree if and only if there is a unique simple path between any two of its vertices.

SAMPLE CSc 340 EXAM QUESTIONS WITH SOLUTIONS: part 2

Exam 1 Solution. CS 542 Advanced Data Structures and Algorithms 2/14/2013

Module graph.py. 1 Introduction. 2 Graph basics. 3 Module graph.py. 3.1 Objects. CS 231 Naomi Nishimura

learning objectives learn what graphs are in mathematical terms learn how to represent graphs in computers learn about typical graph algorithms

Constructive Geometric Constraint Solving

Grade 7/8 Math Circles March 4/5, Graph Theory I- Solutions

Outline. 1 Introduction. 2 Min-Cost Spanning Trees. 4 Example

Algorithmic and NP-Completeness Aspects of a Total Lict Domination Number of a Graph

1 Introduction to Modulo 7 Arithmetic

MAT3707. Tutorial letter 201/1/2017 DISCRETE MATHEMATICS: COMBINATORICS. Semester 1. Department of Mathematical Sciences MAT3707/201/1/2017

(2) If we multiplied a row of B by λ, then the value is also multiplied by λ(here lambda could be 0). namely

b. How many ternary words of length 23 with eight 0 s, nine 1 s and six 2 s?

Graph Search (6A) Young Won Lim 5/18/18

Problem solving by search

Outline. Computer Science 331. Computation of Min-Cost Spanning Trees. Costs of Spanning Trees in Weighted Graphs

CS61B Lecture #33. Administrivia: Autograder will run this evening. Today s Readings: Graph Structures: DSIJ, Chapter 12

OpenMx Matrices and Operators

EE1000 Project 4 Digital Volt Meter

ECE COMBINATIONAL BUILDING BLOCKS - INVEST 13 DECODERS AND ENCODERS

CSC Design and Analysis of Algorithms. Example: Change-Making Problem

0.1. Exercise 1: the distances between four points in a graph

Strongly connected components. Finding strongly-connected components

Paths. Connectivity. Euler and Hamilton Paths. Planar graphs.

12. Traffic engineering

Edge-Triggered D Flip-flop. Formal Analysis. Fundamental-Mode Sequential Circuits. D latch: How do flip-flops work?

Applications of trees

The University of Sydney MATH 2009

4.1 Interval Scheduling. Chapter 4. Greedy Algorithms. Interval Scheduling: Greedy Algorithms. Interval Scheduling. Interval scheduling.

Solutions for HW11. Exercise 34. (a) Use the recurrence relation t(g) = t(g e) + t(g/e) to count the number of spanning trees of v 1

CSE 373: AVL trees. Warmup: Warmup. Interlude: Exploring the balance invariant. AVL Trees: Invariants. AVL tree invariants review

In which direction do compass needles always align? Why?

Divided. diamonds. Mimic the look of facets in a bracelet that s deceptively deep RIGHT-ANGLE WEAVE. designed by Peggy Brinkman Matteliano

Using the Printable Sticker Function. Using the Edit Screen. Computer. Tablet. ScanNCutCanvas

Present state Next state Q + M N

Planar convex hulls (I)

Priority Search Trees - Part I

BASIC CAGE DETAILS SHOWN 3D MODEL: PSM ASY INNER WALL TABS ARE COINED OVER BASE AND COVER FOR RIGIDITY SPRING FINGERS CLOSED TOP

Improving Union. Implementation. Union-by-size Code. Union-by-Size Find Analysis. Path Compression! Improving Find find(e)

CS 103 BFS Alorithm. Mark Redekopp

Face Detection and Recognition. Linear Algebra and Face Recognition. Face Recognition. Face Recognition. Dimension reduction

Announcements. Not graphs. These are Graphs. Applications of Graphs. Graph Definitions. Graphs & Graph Algorithms. A6 released today: Risk

BASIC CAGE DETAILS D C SHOWN CLOSED TOP SPRING FINGERS INNER WALL TABS ARE COINED OVER BASE AND COVER FOR RIGIDITY

V={A,B,C,D,E} E={ (A,D),(A,E),(B,D), (B,E),(C,D),(C,E)}

Graph Contraction and Connectivity

Designing A Uniformly Loaded Arch Or Cable

V={A,B,C,D,E} E={ (A,D),(A,E),(B,D), (B,E),(C,D),(C,E)}

Easy Steps to build a part number... Tri-Start Series III CF P

Lecture 20: Minimum Spanning Trees (CLRS 23)

CMPS 2200 Fall Graphs. Carola Wenk. Slides courtesy of Charles Leiserson with changes and additions by Carola Wenk

Self-Adjusting Top Trees

Functions and Graphs 1. (a) (b) (c) (f) (e) (d) 2. (a) (b) (c) (d)

176 5 t h Fl oo r. 337 P o ly me r Ma te ri al s

Cooperative Triangulation in MSBNs. without Revealing Subnet Structures. Y. Xiang. Department of Computer Science, University of Regina

Math 61 : Discrete Structures Final Exam Instructor: Ciprian Manolescu. You have 180 minutes.

CS 461, Lecture 17. Today s Outline. Example Run

12/3/12. Outline. Part 10. Graphs. Circuits. Euler paths/circuits. Euler s bridge problem (Bridges of Konigsberg Problem)

5/9/13. Part 10. Graphs. Outline. Circuits. Introduction Terminology Implementing Graphs

PH427/PH527: Periodic systems Spring Overview of the PH427 website (syllabus, assignments etc.) 2. Coupled oscillations.

Nefertiti. Echoes of. Regal components evoke visions of the past MULTIPLE STITCHES. designed by Helena Tang-Lim

Designing A Concrete Arch Bridge

SOCKET WELD OR THREADED BODY TYPE (3051SFP FLOWMETER SHOWN THROUGHOUT / 3051CFP, 2051CFP AVAILABLE)

On Hamiltonian Tetrahedralizations Of Convex Polyhedra

Module 2 Motion Instructions

L.3922 M.C. L.3922 M.C. L.2996 M.C. L.3909 M.C. L.5632 M.C. L M.C. L.5632 M.C. L M.C. DRIVE STAR NORTH STAR NORTH NORTH DRIVE

Physics 222 Midterm, Form: A

Similarity Search. The Binary Branch Distance. Nikolaus Augsten.

CS September 2018

I N A C O M P L E X W O R L D

TCI SERIES, 3 CELL THE COOLING TOWER CO., L.C. TCI 3 CELL SERIES GA

Graphs Depth First Search

GREEDY TECHNIQUE. Greedy method vs. Dynamic programming method:

DEVELOPING COMPUTER PROGRAM FOR COMPUTING EIGENPAIRS OF 2 2 MATRICES AND 3 3 UPPER TRIANGULAR MATRICES USING THE SIMPLE ALGORITHM

MULTIPLE-LEVEL LOGIC OPTIMIZATION II

The University of Sydney MATH2969/2069. Graph Theory Tutorial 5 (Week 12) Solutions 2008

Spanning Trees. BFS, DFS spanning tree Minimum spanning tree. March 28, 2018 Cinda Heeren / Geoffrey Tien 1

Outline. Binary Tree

Graphs. Graphs. Graphs: Basic Terminology. Directed Graphs. Dr Papalaskari 1

Rectangular Waveguides

Compression. Compression. Compression. This part of the course... Ifi, UiO Norsk Regnesentral Vårsemester 2005 Wolfgang Leister

7 ACM FOR FRAME 2SET 6 FRAME 2SET 5 ACM FOR MAIN FRAME 2SET 4 MAIN FRAME 2SET 3 POLE ASSLY 1 2 CROWN STRUCTURE ASSLY 1 1 CROWN ASSLY 1

Transcription:

Min Lrning 10-701 Fll 2015 Inrn n Lrning in GM ri Xing r Ltur 18 Novr 10 2015 Ring: p. 8.B ook ri Xing @ MU 2006-2015 1

Inrn n Lrning W now v opt rprsnttions o proility istriutions: BN BN M sris uniqu proility istriution Typil tsks: Tsk 1: How o w nswr quris out? W us inrn s n or t pross o oputing nswrs to su quris Tsk 2: How o w stit plusil ol M ro t D? i. W us lrning s n or t pross o otining point stit o M. ii. But or Bysin ty sk pm D wi is tully n inrn prol. iii. Wn not ll vrils r osrvl vn oputing point stit o M n to o inrn to iput t issing t. ri Xing @ MU 2006-2015 2

ppros to inrn xt inrn lgorits T liintion lgorit Bli propgtion T juntion tr lgorits ut will not ovr in til r pproxit inrn tniqus Vritionl lgorits Stosti siultion / spling tos Mrkov in Mont rlo tos ri Xing @ MU 2006-2015 3

oo w: Qury: By in oposition w gt Mrginliztion n liintion g g B D F G H nïv sution ns to nurt ovr n xponntil nur o trs Wt is t proility tt wks r lving givn tt t grss onition is poor? g g 4 ri Xing @ MU 2006-2015

Qury: N to liint: BDFGH Initil tors: oos n liintion orr: HGFDB Stp 1: onitioning ix t vin no i.. on its osrv vlu i.. : Tis stp is isoorpi to rginliztion stp: B D F G H g ~ p ~ p ~ B D F G Vril liintion 5 ri Xing @ MU 2006-2015

Qury: B N to liint: BDFG Initil tors: Stp 2: liint G oput B D F G H g g 1 g g g p B D F g xpl: Vril liintion 6 ri Xing @ MU 2006-2015

Qury: B N to liint: BDF Initil tors: Stp 3: liint F oput B D F G H xpl: Vril liintion g g p B D 7 ri Xing @ MU 2006-2015

B D Qury: B N to liint: BD Initil tors: Stp 4: liint oput B D F G H xpl: Vril liintion g g p B D 8 ri Xing @ MU 2006-2015

Qury: B N to liint: BD Initil tors: Stp 5: liint D oput B D F G H xpl: Vril liintion g g p B 9 ri Xing @ MU 2006-2015

Qury: B N to liint: B Initil tors: Stp 6: liint oput B D F G H xpl: Vril liintion p g g B 10 ri Xing @ MU 2006-2015

Qury: B N to liint: B Initil tors: Stp 7: liint B oput B D F G H xpl: Vril liintion g g p 11 ri Xing @ MU 2006-2015

Qury: B N to liint: B Initil tors: Stp 8: Wrp-up B D F G H xpl: Vril liintion g g ~ p p p p ~ p p ~ 12 ri Xing @ MU 2006-2015

oplxity o vril liintion Suppos in on liintion stp w oput Tis rquirs x y1 yk ' x x y1 yk ' x x x k 1 yk i x y i i1 k Vl X Vl Y ultiplitions i y i For vlu o x y 1 y k w o k ultiplitions Vl X Vl Y i itions i For vlu o y 1 y k w o VlX itions oplxity is xponntil in nur o vrils in t intrit tor ri Xing @ MU 2006-2015 13

liintion liqu Inu pnny uring rginliztion is ptur in liintion liqus Sution <-> liintion Intrit tr <-> liintion liqu B B D D F n tis l to n gnri inrn lgorit? G H F ri Xing @ MU 2006-2015 14

Fro liintion to Mssg ssing liintion ssg pssing on liqu tr B B D p g G g D H F F Mssgs n rus ri Xing @ MU 2006-2015 15

Fro liintion to Mssg ssing liintion ssg pssing on liqu tr notr qury... B B D D F G g H F Mssgs n r rus otrs n to roput ri Xing @ MU 2006-2015 16

Fro liintion to ssg pssing Rll LIMINTION lgorit: oos n orring Z in wi qury no is t inl no l ll potntils on n tiv list liint no i y roving ll potntils ontining i tk su/prout ovr x i. l t rsultnt tor k on t list For TR grp: oos qury no s t root o t tr Viw tr s irt tr wit gs pointing towrs ro liintion orring s on pt-irst trvrsl liintion o no n onsir s ssg-pssing or Bli ropgtion irtly long tr rns rtr tn on so trnsor grps tus w n us t tr itsl s t-strutur to o gnrl inrn!! ri Xing @ MU 2006-2015 17

Mssg pssing or trs Lt ij x i not t tor rsulting ro liinting vrils ro llow up to i wi is untion o x i : i Tis is rinisnt o ssg snt ro j to i. j k l ij x i rprsnts "li" o x i ro x j! ri Xing @ MU 2006-2015 18

liintion on trs is quivlnt to ssg pssing long tr rns! i j k l ri Xing @ MU 2006-2015 19

T ssg pssing protool: two-pss lgorit: X 1 21 X 1 12 X 2 32 X 2 X 2 42 X 2 X 3 X 4 24 X 4 23 X 3 ri Xing @ MU 2006-2015 20

Bli ropgtion S-lgorit: Squntil iplnttion ri Xing @ MU 2006-2015 21

Inrn on gnrl GM Now wt i t GM is not tr-lik grp? n w still irtly run ssg ssg-pssing protool long its gs? For non-trs w o not v t gurnt tt ssg-pssing will onsistnt! Tn wt? onstrut grp t-strutur ro tt s tr strutur n run ssg-pssing on it! Juntion tr lgorit ri Xing @ MU 2006-2015 22

Sury: xt Inrn T sipl liint lgorit pturs t ky lgoriti Oprtion unrlying proilisti inrn: --- Tt o tking su ovr prout o potntil untions T oputtionl oplxity o t liint lgorit n ru to purly grp-torti onsirtions. Tis grp intrprttion will lso provi ints out ow to sign iprov inrn lgorits Wt n w sy out t ovrll oputtionl oplxity o t lgorit? In prtiulr ow n w ontrol t "siz" o t suns tt ppr in t squn o sution oprtion. ri Xing @ MU 2006-2015 23

ppros to inrn xt inrn lgorits T liintion lgorit Bli propgtion T juntion tr lgorits ut will not ovr in til r pproxit inrn tniqus Vritionl lgorits Stosti siultion / spling tos Mrkov in Mont rlo tos ri Xing @ MU 2006-2015 24

Mont rlo tos Drw rno spls ro t sir istriution Yil stosti rprsnttion o oplx istriution rginls n otr xptions n pproxit using spl-s vrgs N 1 t [ x ] x N t 1 syptotilly xt n sy to pply to ritrry ols llngs: ow to rw spls ro givn ist. not ll istriutions n trivilly spl? ow to k ttr us o t spls not ll spl r usul or qlly usul s n xpl ltr? ow to know w'v spl noug? ri Xing @ MU 2006-2015

xpl: niv spling onstrut spls oring to proilitis givn in BN. lr xpl: oos t rigt spling squn 1 Spling:B=<0.001 0.999> suppos it is ls B0. S or 0. B0 0=<0.001 0.999> suppos it is ls... 2 Frquny ounting: In t spls rigt J0=J0/0=<1/9 8/9>. ri Xing @ MU 2006-2015 0 B0 0 M0 J0 0 B0 0 M0 J0 0 B0 0 M0 J1 0 B0 0 M0 J0 0 B0 0 M0 J0 0 B0 0 M0 J0 1 B0 1 M1 J1 0 B0 0 M0 J0 0 B0 0 M0 J0 0 B0 0 M0 J0

xpl: niv spling onstrut spls oring to proilitis givn in BN. lr xpl: oos t rigt spling squn 3 wt i w wnt to oput J1? w v only on spl... J1=J1/1=<0 1>. 4 wt i w wnt to oput JB1? No su spl vill! J1=JB1/B1 n not in. For ol wit unrs or or vrils rr vnts will vry r to grnr voug spls vn tr long ti or spling... 0 B0 0 M0 J0 0 B0 0 M0 J0 0 B0 0 M0 J1 0 B0 0 M0 J0 0 B0 0 M0 J0 0 B0 0 M0 J0 1 B0 1 M1 J1 0 B0 0 M0 J0 0 B0 0 M0 J0 0 B0 0 M0 J0 ri Xing @ MU 2006-2015

Mrkov in Mont rlo MM onstrut Mrkov in wos sttionry istriution is t trgt nsity = X. Run or T spls urn-in ti until t in onvrgs/ixs/rs sttionry istriution. Tn ollt M orrlt spls x. Ky issus: Dsigning proposls so tt t in ixs rpily. Dignosing onvrgn. ri Xing @ MU 2006-2015

Mrkov ins Dinition: Givn n n-insionl stt sp Rno vtor X = x 1 x n x t = x t ti-stp t x t trnsitions to x t+1 wit pro x t+1 x t x 1 = Tx t+1 x t = Tx t x t+1 Hoognous: in trin y stt x 0 ix trnsition krnl Q rows su to 1 quiliriu: x is sttionry quiliriu istriution i x' = x x Qxx'. i.. is lt ignvtor o t trnsition trix T = T Q. 0.25 0.7 0. 2 0. 5 0. 3 0. 2 0. 5 0. 3 0. 25 0 0. 5 0 0. 7 0. 5 0. 75 0. 3 0 ri Xing @ MU 2006-2015 X 1 X 2 0.75 0.5 0.5 0.3 X 3

Gis spling T trnsition trix upts no on t ti using t ollowing proposl: Q x x x ' x p x ' x i i i i i i ' It is iint sin p x i x i only pns on t vlus in X i s Mrkov lnkt ri Xing @ MU 2006-2015

Gis spling Gis spling is n MM lgorit tt is spilly pproprit or inrn in grpil ols. T prou w v vril st X={x 1 x 2 x 3... x N } or GM t stp on o t vrils X i is slt t rno or oring to so ix squns not t rining vrils s X -i n its urrnt vlu s x -i t-1 Using t "lr ntwork" s n xpl sy t ti t w oos X n w not t urrnt vlu ssignnts o t rining vrils X - otin ro prvious spls s t 1 t 1 t 1 t 1 t 1 x x x x x t onitonl istriution px i x -i t-1 is oput vlu x i t is spl ro tis istriution t spl x i t rpls t prvious spl vlu o X i in X. B J M i.. x x x t t 1 t ri Xing @ MU 2006-2015

Mrkov Blnkt Mrkov Blnkt in BN vril is inpnnt ro otrs givn its prnts ilrn n ilrn s prnts sprtion. MB in MRF vril is inpnnt ll its non-nigors givn ll its irt nigors. px i X -i = px i MBX i Gis spling vry stp oos on vril n spl it y XMBX s on prvious spl. ri Xing @ MU 2006-2015

Gis spling o t lr ntwork MB={B J M} MB={ B} To lult JB1M1 oos B101M1J1 s strt vins r B1 M1 vrils r J. oos nxt vril s Spl y MB=B1 0 M1 J1 suppos to ls. B1 0 0 M1 J1 oos nxt rno vril s spl ~B10... ri Xing @ MU 2006-2015

xpl ri Xing @ MU 2006-2015

xpl: ri Xing @ MU 2006-2015

xpl ri Xing @ MU 2006-2015

xpl J1 B1M1 = 0.90 J1 1M0 = 0.14 1 J1 = 0.01 1 M1 = 0.04 1 M1J1 = 0.17 ri Xing @ MU 2006-2015

T o siultion Run svrl ins Strt t ovr-isprs points Monitor t log lik. Monitor t sril orrltions Monitor ptn rtios R-prtriz to gt pprox. inp. R-lok Gis ollps int. ovr otr prs. Run wit troul prs. ix t rsonl vls. ri Xing @ MU 2006-2015

Lrning Grpil Mols T gol: Givn st o inpnnt spls ssignnts o rno vrils in t st t ost likly? Bysin Ntwork ot DG n Ds B B R R BR=TFFTF BR=TFTTF.. BR=FTTTF ri Xing @ MU 2006-2015 B B 0.9 0.1 0.2 0.8 0.9 0.1 0.01 0.99 39

Lrning Grpil Mols ont. Snrios: opltly osrv GMs irt unirt prtilly osrv GMs irt unirt n opn rsr topi stition prinipls: Mxil liklioo stition ML Bysin stition Mxil onitionl liklioo Mxil "Mrgin" W us lrning s n or t pross o stiting t prtrs n in so ss t topology o t ntwork ro t. ri Xing @ MU 2006-2015 40

ML or gnrl BN prtrs I w ssu t prtrs or D r glolly inpnnt n ll nos r ully osrv tn t logliklioo untion oposs into su o lol trs on pr no: l ; D log p D log i p xn i xn log p x i n i xn i i n i i n X 2 =1 X 5 =0 X 2 =0 X 5 =1 ri Xing @ MU 2006-2015 41

xpl: oposl liklioo o irt ol onsir t istriution in y t irt yli GM: p x p x1 1 p x2 x1 2 p x3 x1 3 p x4 x2 x3 4 Tis is xtly lik lrning our sprt sll BNs o wi onsists o no n its prnts. X 1 X 1 X 1 X 1 X 2 X 3 X 2 X 2 X 3 X 3 X 4 X 4 ri Xing @ MU 2006-2015 42

.g.: ML or BNs wit tulr Ds ssu D is rprsnt s tl ultinoil wr p X j X k Not tt in s o ultipl prnts will v oposit stt n t D will ig-insionl tl T suiint sttistis r ounts o ily onigurtions T log-liklioo is Using Lgrng ultiplir j to nor 1 w gt: ijk ijk n ijk i n x j n i i X i x k n i ijk l ; D log n log i j k ML ijk n ijk n i j' k i j k ijk n ij' k ijk ijk ri Xing @ MU 2006-2015 43

Dinition o HMM Trnsition proilitis twn ny two stts y 1 y 2 y 3... y T j i p yt 1 yt 1 1 i j x 1 x 2 x 3... x T or i. i p yt yt 1 1 ~ Multinoil i 1 i 2 i M I Strt proilitis p ~ Multinoil. y 1 1 2 M ission proilitis ssoit wit stt i. i p xt yt 1 ~ Multinoil i 1 i 2 i K I or in gnrl: i. i p x y 1 ~ I t t i ri Xing @ MU 2006-2015 44

Suprvis ML stition Givn x = x 1 x N or wi t tru stt pt y = y 1 y N is known T T l θ; x y log p x y log p y p y y p x x Din: ij B ik = # tis stt trnsition ij ours in y = # tis stt i in y its k in x W n sow tt t xiu liklioo prtrs r: ML ij ML ik # i j # i # i k # i n T n t n T n t n i j 2 n t 1 n t T i y t 2 n t 1 1 T I y is ontinuous w n trt xn t yn t : t 1 : T n 1 : N s NT osrvtions o.g. Gussin n pply lrning ruls or Gussin y y t 1 i n t y x i n t y k n t k ' ij j ' Bik B ij ' ik ' n 1 n t n t1 n t n t t2 t1 ri Xing @ MU 2006-2015 45

onsir t istriution in y t irt yli GM: N to oput px H x V inrn Wt i so nos r not osrv? 1 3 2 4 3 1 3 1 1 2 1 1 x x x p x x p x x p x p x p X 1 X 4 X 2 X 3 46 ri Xing @ MU 2006-2015

ML or BNs wit tulr Ds ssu D is rprsnt s tl ultinoil wr p X j X k Not tt in s o ultipl prnts will v oposit stt n t D will ig-insionl tl T suiint sttistis r ounts o ily onigurtions T log-liklioo is Using Lgrng ultiplir j to nor 1 w gt: ijk ijk n ijk i n x j n i i X i x k n i ijk l ; D log n log i j k ML ijk n ijk n i j' k i j k ijk n ij ' k ijk ijk ri Xing @ MU 2006-2015 47

Sury GMM sris uniqu proility istriution Typil tsks: Tsk 1: How o w nswr quris out? W us inrn s n or t pross o oputing nswrs to su quris Tsk 2: How o w stit plusil ol M ro t D? i. W us lrning s n or t pross o otining point stit o M. ii. But or Bysin ty sk pm D wi is tully n inrn prol. iii. Wn not ll vrils r osrvl vn oputing point stit o M n to o inrn to iput t issing t. ri Xing @ MU 2006-2015 48