Speeding Up Computation in Probabilistic Graphical Models using GPGPUs

Size: px
Start display at page:

Download "Speeding Up Computation in Probabilistic Graphical Models using GPGPUs"

Transcription

1 Speeding Up Computation in Probabilistic Graphical Models using GPGPUs Lu Zheng & Ole J. Mengshoel {lu.zheng GPU Technology Conference San Jose, CA, March 20, 2013

2 CMU in Silicon Valley Established 2002 Significant growth in the past 10 years

3 Overview Bayesian networks (BNs) Representation of joint probability distribution Computation of marginals and most probable explanations Compilation to junction trees GPU: Speed up junction tree computation Why? Parallel opportunities in junction trees Element-wise and arithmetic parallelism Thread optimization using machine learning Increasing importance of parallel computing Speed matters: Real-time requirements in applications Increasing size of data sets in machine learning Big Data Junction tree computation can be used in inner loop of EM

4 Bayesian Networks and Probability A Bayesian network is a compact representation of a joint probability distribution function P( V1,..., V n ) Representation is according to Bayesian network structure: P( V,..., V ) P( V V,..., V ) P( V V ) P( V ) 1 n n n n i 1 P( V pa( V )) i where pa(v i ) gives parents of V i in the graph. i C A B For our example: P(A, B, C, D, E) = P(E A, B, C, D) P(D A, B, C) P(C A, B) P(B A) P(A) = P(E D) P(D B, C) P(C A,) P(B A) P(A) D E

5 Bayesian Network Applications Diagnosis and monitoring Electrical power systems Spam filtering Naïve Bayes Error correction coding Biological and medical research Computer security Sensor and information fusion Probabilistic risk assessment Fault trees Natural language understanding Intelligent tutoring systems

6 Stereo Diagnosis - Structure System hidden state - underlying causes of failure System observable state sensors or tests If the CD does not spin, then there might be bad or no sound. If there s a media fault, then the CD might not spin.

7 Stereo Diagnosis - Probabilities Each BN node comes with a conditional probability table (CPT).

8 Stereo Diagnosis - Marginals Marginal: Posterior distribution P(X e), where e is evidence (input). Input: Everything is wrong. Output: Marginals suggest there is a power failure, everything else is OK. Input: LED is lighting; everything else is wrong. Output: Marginals suggest there is a media failure, everything else is OK.

9 Stereo Diagnosis - Compilation Bayesian Network LEDFault PowerFault (P) AmpFault (A) SpeakerFault (S) MediaFault (M) NoLight (G) BadSound (B) NoSpin (N) Compile Separator: Table that is the intersection of its neighboring nodes PowerFault (P) Clique: Table exponential in number of nodes PowerFault (P) NoSpin (N) NoSpin (N) P, N P SpeakerFault (S) PowerFault (P) NoLight (G) Junction Tree MediaFault (M) BadSound (B) AmpFault (A) LEDFault Root clique: A designated clique

10 Probabilistic Diagnosis Approach LEDFault PowerFault (P) AmpFault (A) SpeakerFault (S) MediaFault (M) NoLight (G) BadSound (B) NoSpin (N) PowerFault (P) PowerFault (P) NoSpin (N) PowerFault (P) Each health variable has at least two states ( healthy and faulty ), thus admitting the diagnosis of an arbitrary number of discrete faults. NoSpin (N) MediaFault (M) SpeakerFault (S) BadSound (B) AmpFault (A) NoLight (G) LEDFault OFF-LINE PHASE Offline Generation Bayesian Network (BN) Offline Compilation System Specification Junction Tree (JT) For diagnosis, we perform message passing in the junction tree to answer a probabilistic query over the health variables to understand which components and/or sensors are in non-healthy state(s). Sensor, Commands ON-LINE PHASE Online Inference Diagnosis: MLV, MPE, MAP

11 Backgound Previous Work Parallel Message Passing Tuning GPU Parameters Summary Junction Tree Algorithm Bayesian Network Inference with Junction Tree Generate junction tree (moralization and triangulation) A,B,D B D B,C D,E Propagate the belief values B =? Infer for the desired random variable B based on the posterior distribution Pr(B e). 4 / 40

12 Backgound Previous Work Parallel Message Passing Tuning GPU Parameters Summary Computational Challenge for Junction Tree Computational Difficulty Graph structure A densely connected BN can result in big cliques in junction tree. Computational complexity increases exponentially with the treewidth, which is the optimal maximal number of nodes in clique. Random variables value sets The high cardinality of source discrete BN nodes. Algorithm The inner loop of iterative algorithms like the EM algorithm. Thus Develop parallel computing techniques for junction tree inference. 5 / 40

13 Backgound Previous Work Parallel Message Passing Tuning GPU Parameters Summary 1 Backgound 2 Previous Work 3 Parallel Message Passing Techniques Experimental Results 4 Tuning GPU Parameters Techniques Experimental Results 5 Summary 6 / 40

14 Backgound Previous Work Parallel Message Passing Tuning GPU Parameters Summary Related Works Parallel Junction Tree Graph level parallelism GraphLab [Low at al, 2010] Parallel Lauritzen-Spiegelhalter algorithm [Kozlov & Singh, 1996] Node level parallelism Pointer Jumper [Namasivayam & Prasanna, 2006] Parallel node level table operation primitives [Xia & Prasanna, 2007] 7 / 40

15 1 Backgound 2 Previous Work 3 Parallel Message Passing Techniques Experimental Results 4 Tuning GPU Parameters Techniques Experimental Results 5 Summary 8 / 40

16 Message Passing Message Passing Marginalization. The potential table φ Sik of the separator is updated to φ S ik by marginalizing the potential table φ Xi : φ S ik = X i /S ik φ Xi. Scattering. The potential table of C k is updated using both the old and new table of S ik : φ X k = φ Xk φ S ik φ Sik. 9 / 40

17 Parallelism in Message Passing Contribution Identified the parallel opportunities in message passing Element-wise parallelism Arithmetic parallelism Proposed and implemented a parallel belief propagation algorithm Proposed high level technique such as clique merging to increase the parallel opportunities. B,C Parallel Message Passing B A,B,D Thread 1 Thread 2 Thread N D D,E 10 / 40

18 Element-wise Parallelism B D Clique ABD Index mapping from right clique to separator BC Clique 010 B B Separator A,B,D Index mapping from left clique to separator B,C D,E Message passing: (A, B, D) (B) (B, C). Potential table operations marginalization: φ (B) = A,D φ(a, B, D). scattering: φ (B, C) = φ(b, C) φ (B) φ(b). 11 / 40

19 Element-wise Parallelism Details Sepset-cluster mapping table [Huang & Darwiche, 1996] - record the indices of the cells in the clique potential table that map to the same separator potential table cell, providing data separation. All sepset-cluster mapping table can be processed in parallel. 12 / 40

20 Arithmetic Parallelism Look Up Table Separator: CPT Look Up Table (d-1) 2 2 (d-1) d -1 Clique i Separator Clique k Clique k: CPT + + Clique i: CPT Sum Parallel Multiplication Parallel (d-2) (d-1) Arithmetic Parallelism Parallelize the summation in marginalization Parallelize the multiplication in scattering 13 / 40

21 Analysis Of Message Passing Time A message passing from clique i to clique k Let φ Xi, φ Xk and φ S be the potential tables of clique i, clique k and separator respectively. is the cardinality operation. p is the number of threads assigned to the multiplication parallelism. The time complexity of this message passing is log φ Xi log φ S + 2 φ X k φ S p. 14 / 40

22 Analysis Of Propagation Time Belief Propagation Let Ne(C) be the neighbors of C in the junction tree. Suppose the kernel overhead is a constant τ. The time complexity for belief propagation is 2(n 1)τ + ( log φ Xi log φ S + 2 φ ) X k. φ S p i k Ne(C i ) 15 / 40

23 Parallelism Challenges Clique Merging When φ Sik and µ Xi,S ik (n) are small, there is not enough amount of parallel opportunity to occupy the concurrency resources provided by the computing platform. Examples of Bayesian networks that consists of small cliques and separators 300 Pigs 450 Munin <2 2~5 5~10 10~20 20~5050~100 >100 Separator Table Size 0 (X10) <2 2~5 5~10 10~20 20~50 50~100 >100 Separator Table Size (X10) 16 / 40

24 Clique Merging: Handling Challenges Clique Merging A E A E C D C&D B F B F G G Theorem Two neighboring cliques C i and C j in a junction tree J with the separator S ij can be merged together into an equivalent new clique C ij with the potential function φ(x Cij ) = φ(x C i )φ(x Cj ), φ(x Sij ) while keeping all the other part of the junction tree unchanged. 17 / 40

25 Experimental Setups Data sets and Platform Implementation are tested on a number of Bayesian networks from different domains, with varying structures and state spaces. Implementation Platform NVIDIA Geforce GTX 460 # of Processing Cores 336 Shared Memory 48K per block Global Memory 785MB Memory Bandwidth 90 GB/sec peak Intel Core2 Quad CPU # of Cores 4 Processor Clock 2.5GHz Cache 8MB Memory 9 GB 18 / 40

26 CPU versus GPU: Experiments Performance Comparison Using: Element-wise parallelism, arithmetic parallelism and clique merging. The speedups for different Bayesian networks range from 1.82x to 11.94x, with an average of 5.54x (or 554%) CPU execution time GPU execution time Execution Time Pigs Munin2 Munin3 Munin4 Mildew Water Barley Diabetes 19 / 40

27 Results with AP, EP and CM (1) Dataset Mildew Water Barley Diabetes # of original JT nodes # of JT nodes after merge Ave. CPT size in original JT 341, , , Ave. CPT size after merge 447, , , Ave. SPT size in original JT 9,273 26,065 39, Ave. SPT size after merge 11,883 29,129 40, GPU time before merge [ms] GPU time after merge [ms] Merging Speedup 1.02x 1.01x 1.01x 1.02x CPU time [ms] CPU versus GPU 8.59x 7.35x 11.94x 5.81x 20 / 40

28 Results with AP, EP and CM (2) Dataset Pigs Munin2 Munin3 Munin4 # of original JT nodes # of JT nodes after merge Ave. CPT size in original JT 1,972 5,653 3,443 16,444 Ave. CPT size after merge 5,393 10,191 7,374 26,720 Ave. SPT size in original JT ,099 Ave. SPT size after merge 757 1, ,214 GPU time before merge [ms] GPU time after merge [ms] Merging Speedup 1.36x 1.14x 1.14x 1.10x CPU time [ms] CPU versus GPU 2.26x 2.43x 1.82x 3.35x 21 / 40

29 1 Backgound 2 Previous Work 3 Parallel Message Passing Techniques Experimental Results 4 Tuning GPU Parameters Techniques Experimental Results 5 Summary 22 / 40

30 Difficulty in GPU parameter tuning (PT) Tuning GPU Matching platform concurrency and parallel opportunity in message passing. Irregular topology of junction tree - there is high variability in workload and usage patterns. Limited thread availability for two dimensions of parallelism. GPU architecture evolves rapidly 23 / 40

31 GPU Performance for Large Separator Large separators great potential for many sepset-cluster mapping tables large element-wise parallel opportunity Example I Varying TBS, AP, EP 24 / 40

32 GPU Performance for Small Separator Small separators little potential for many sepset-cluster mapping tables small element-wise parallel opportunity. Example II Varying TBS, AP, EP 25 / 40

33 GPU Performance for Medium Separator Medium separators? Example III Varying TBS, AP, EP 26 / 40

34 Regression Models for GPU Goal Build regression models to emulate the GPU performance space. Use regression model to optimize parameter setting. Features Feature Description φ S Size of separator potential table φ X Size of clique potential table K Size of GPU thread block Threads for arithmetic parallelism p a Table: Features used in regression models 27 / 40

35 Polynomial Regression Polynomial Regression Polynomial model + lasso ˆβ lasso = arg min β N y i β 0 i=1 subject to Equivalent Lagrangian ˆβ lasso = arg min β p β j t, j=1 N y i β 0 i=1 p Poly j (x) j=1 p Poly j (x) j=1 2 + λ 2 p β j, j=1 28 / 40

36 Support Vector Regression (SVR) SVR SVR f (x, ω) = m ω j g j (x) + b, j=1 Loss Function: ɛ - insensitive { 0 if y f (x, ω) ɛ L ω (y, f (x, ω)) = y f (x, ω) ɛ otherwise 29 / 40

37 Metrics for Regression Model Quality Traditional Metric Residual Squared Sum (RSS) RSS = n (y i f (x i )) 2 i=1 Minimum Variance Squared Deviance (SD) SD = (T ( φ S, φ X, ˆK, ˆp a) T ) 2, φ S, φ X Miss Rate (MR) MR = φ S, φ X 1( ˆK K or ˆp a pa), N 30 / 40

38 Regression Model Evaluation Model Evaluation Method Data RSS SD MR Lasso(λ = 0) Reduce 12.3e % Lasso(λ = 1se) Reduce 14.9e4 11.3e % SVR Reduce 13.6e % Lasso(λ = 0) Scatter 1.31e % Lasso(λ = 1se) Scatter 1.57e % SVR Scatter 2.69e % Table: Residual Squared Sum (RSS), Squared Deviance (SD) and Miss Rate (MR) of polynomial and SVR model for GPU reduction and scattering execution time 31 / 40

39 Regression Example: Large Separator Example I - Marginalization actual lasso 30 svr 25 Execution Time p tot p e 32 / 40

40 Regression Example: Small Separator Example II - Marginalization 35 actual lasso svr 30 Execution Time p tot p e 33 / 40

41 Regression Example: Medium Separator Example III - Marginalization 0.6 actual lasso svr 0.5 Execution Time p tot p e 34 / 40

42 GPU Execution Times GPU Execution Times with EP + AP + PT threshold GPU parameters 1 thresold GPU parameters GPU parameters with lasso model (λ = 1SE) GPU parameters with lasso model (λ = 0) GPU parameters with SVR Execution Time Pigs Water Munin2 Munin3 Munin4 Mildew Barley Diabetes SVR works best in these experiments. 35 / 40

43 Results with Regression Parameter Tuning(1) Manual parameter tuning versus regression parameter tuning Pigs Water Munin2 Munin3 Munin4 # of JT nodes Ave. CPT size 1, ,297 5,653 3,443 16,444 Ave. SPT size ,099 9,273 0-threshold [ms] threshold [ms] SVR [ms] Lasso (λ = 0) [ms] Lasso (λ = 1se)[ms] GPU time [ms] CPU time [ms] SVR/CPU Speedup 3.72x 12.68x 4.80x 3.89x 7.04x 36 / 40

44 Results with Regression Parameter Tuning(2) Manual parameter tuning versus regression parameter tuning Mildew Barley Diabetes # of JT nodes Ave. CPT size 341, , Ave. SPT size 26,065 39, threshold [ms] threshold [ms] SVR [ms] Lasso (λ = 0) [ms] Lasso (λ = 1se)[ms] GPU time [ms] CPU time [ms] SVR/CPU Speedup 19.90x 21.19x 12.36x 37 / 40

45 Backgound Previous Work Parallel Message Passing Tuning GPU Parameters Summary 1 Backgound 2 Previous Work 3 Parallel Message Passing Techniques Experimental Results 4 Tuning GPU Parameters Techniques Experimental Results 5 Summary 38 / 40

46 Backgound Previous Work Parallel Message Passing Tuning GPU Parameters Summary Summary Summary Parallel belief propagation Element-wise parallelism. Arithmetic parallelism. Clique Merging. Parameter Tuning 39 / 40

47 Backgound Previous Work Parallel Message Passing Tuning GPU Parameters Summary Thanks for listening! Questions? 40 / 40

Parallel Exact Inference on the Cell Broadband Engine Processor

Parallel Exact Inference on the Cell Broadband Engine Processor Parallel Exact Inference on the Cell Broadband Engine Processor Yinglong Xia and Viktor K. Prasanna {yinglonx, prasanna}@usc.edu University of Southern California http://ceng.usc.edu/~prasanna/ SC 08 Overview

More information

Probabilistic Graphical Models

Probabilistic Graphical Models Probabilistic Graphical Models Lecture 8 Junction Trees CS/CNS/EE 155 Andreas Krause Announcements Homework 2 due next Wednesday (Nov 4) in class Start early!!! Project milestones due Monday (Nov 9) 4

More information

STAT 598L Probabilistic Graphical Models. Instructor: Sergey Kirshner. Exact Inference

STAT 598L Probabilistic Graphical Models. Instructor: Sergey Kirshner. Exact Inference STAT 598L Probabilistic Graphical Models Instructor: Sergey Kirshner Exact Inference What To Do With Bayesian/Markov Network? Compact representation of a complex model, but Goal: efficient extraction of

More information

Bayesian Networks Inference (continued) Learning

Bayesian Networks Inference (continued) Learning Learning BN tutorial: ftp://ftp.research.microsoft.com/pub/tr/tr-95-06.pdf TAN paper: http://www.cs.huji.ac.il/~nir/abstracts/frgg1.html Bayesian Networks Inference (continued) Learning Machine Learning

More information

ECE 6504: Advanced Topics in Machine Learning Probabilistic Graphical Models and Large-Scale Learning

ECE 6504: Advanced Topics in Machine Learning Probabilistic Graphical Models and Large-Scale Learning ECE 6504: Advanced Topics in Machine Learning Probabilistic Graphical Models and Large-Scale Learning Topics Bayes Nets: Inference (Finish) Variable Elimination Graph-view of VE: Fill-edges, induced width

More information

Computer Vision Group Prof. Daniel Cremers. 4. Probabilistic Graphical Models Directed Models

Computer Vision Group Prof. Daniel Cremers. 4. Probabilistic Graphical Models Directed Models Prof. Daniel Cremers 4. Probabilistic Graphical Models Directed Models The Bayes Filter (Rep.) (Bayes) (Markov) (Tot. prob.) (Markov) (Markov) 2 Graphical Representation (Rep.) We can describe the overall

More information

Probabilistic Graphical Models

Probabilistic Graphical Models Overview of Part Two Probabilistic Graphical Models Part Two: Inference and Learning Christopher M. Bishop Exact inference and the junction tree MCMC Variational methods and EM Example General variational

More information

Part II. C. M. Bishop PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 8: GRAPHICAL MODELS

Part II. C. M. Bishop PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 8: GRAPHICAL MODELS Part II C. M. Bishop PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 8: GRAPHICAL MODELS Converting Directed to Undirected Graphs (1) Converting Directed to Undirected Graphs (2) Add extra links between

More information

CS242: Probabilistic Graphical Models Lecture 3: Factor Graphs & Variable Elimination

CS242: Probabilistic Graphical Models Lecture 3: Factor Graphs & Variable Elimination CS242: Probabilistic Graphical Models Lecture 3: Factor Graphs & Variable Elimination Instructor: Erik Sudderth Brown University Computer Science September 11, 2014 Some figures and materials courtesy

More information

D-Separation. b) the arrows meet head-to-head at the node, and neither the node, nor any of its descendants, are in the set C.

D-Separation. b) the arrows meet head-to-head at the node, and neither the node, nor any of its descendants, are in the set C. D-Separation Say: A, B, and C are non-intersecting subsets of nodes in a directed graph. A path from A to B is blocked by C if it contains a node such that either a) the arrows on the path meet either

More information

OSU CS 536 Probabilistic Graphical Models. Loopy Belief Propagation and Clique Trees / Join Trees

OSU CS 536 Probabilistic Graphical Models. Loopy Belief Propagation and Clique Trees / Join Trees OSU CS 536 Probabilistic Graphical Models Loopy Belief Propagation and Clique Trees / Join Trees Slides from Kevin Murphy s Graphical Model Tutorial (with minor changes) Reading: Koller and Friedman Ch

More information

Computer Vision Group Prof. Daniel Cremers. 4a. Inference in Graphical Models

Computer Vision Group Prof. Daniel Cremers. 4a. Inference in Graphical Models Group Prof. Daniel Cremers 4a. Inference in Graphical Models Inference on a Chain (Rep.) The first values of µ α and µ β are: The partition function can be computed at any node: Overall, we have O(NK 2

More information

Modeling and Reasoning with Bayesian Networks. Adnan Darwiche University of California Los Angeles, CA

Modeling and Reasoning with Bayesian Networks. Adnan Darwiche University of California Los Angeles, CA Modeling and Reasoning with Bayesian Networks Adnan Darwiche University of California Los Angeles, CA darwiche@cs.ucla.edu June 24, 2008 Contents Preface 1 1 Introduction 1 1.1 Automated Reasoning........................

More information

Lecture 5: Exact inference. Queries. Complexity of inference. Queries (continued) Bayesian networks can answer questions about the underlying

Lecture 5: Exact inference. Queries. Complexity of inference. Queries (continued) Bayesian networks can answer questions about the underlying given that Maximum a posteriori (MAP query: given evidence 2 which has the highest probability: instantiation of all other variables in the network,, Most probable evidence (MPE: given evidence, find an

More information

Junction tree propagation - BNDG 4-4.6

Junction tree propagation - BNDG 4-4.6 Junction tree propagation - BNDG 4-4. Finn V. Jensen and Thomas D. Nielsen Junction tree propagation p. 1/2 Exact Inference Message Passing in Join Trees More sophisticated inference technique; used in

More information

Computer Vision Group Prof. Daniel Cremers. 4. Probabilistic Graphical Models Directed Models

Computer Vision Group Prof. Daniel Cremers. 4. Probabilistic Graphical Models Directed Models Prof. Daniel Cremers 4. Probabilistic Graphical Models Directed Models The Bayes Filter (Rep.) (Bayes) (Markov) (Tot. prob.) (Markov) (Markov) 2 Graphical Representation (Rep.) We can describe the overall

More information

Lecture 9: Undirected Graphical Models Machine Learning

Lecture 9: Undirected Graphical Models Machine Learning Lecture 9: Undirected Graphical Models Machine Learning Andrew Rosenberg March 5, 2010 1/1 Today Graphical Models Probabilities in Undirected Graphs 2/1 Undirected Graphs What if we allow undirected graphs?

More information

STA 4273H: Statistical Machine Learning

STA 4273H: Statistical Machine Learning STA 4273H: Statistical Machine Learning Russ Salakhutdinov Department of Statistics! rsalakhu@utstat.toronto.edu! http://www.utstat.utoronto.ca/~rsalakhu/ Sidney Smith Hall, Room 6002 Lecture 5 Inference

More information

CS242: Probabilistic Graphical Models Lecture 2B: Loopy Belief Propagation & Junction Trees

CS242: Probabilistic Graphical Models Lecture 2B: Loopy Belief Propagation & Junction Trees CS242: Probabilistic Graphical Models Lecture 2B: Loopy Belief Propagation & Junction Trees Professor Erik Sudderth Brown University Computer Science September 22, 2016 Some figures and materials courtesy

More information

Probabilistic Graphical Models

Probabilistic Graphical Models Probabilistic Graphical Models Raquel Urtasun and Tamir Hazan TTI Chicago April 25, 2011 Raquel Urtasun and Tamir Hazan (TTI-C) Graphical Models April 25, 2011 1 / 17 Clique Trees Today we are going to

More information

Bayesian Networks Inference

Bayesian Networks Inference Bayesian Networks Inference Machine Learning 10701/15781 Carlos Guestrin Carnegie Mellon University November 5 th, 2007 2005-2007 Carlos Guestrin 1 General probabilistic inference Flu Allergy Query: Sinus

More information

Scalable Parallel Implementation of Bayesian Network to Junction Tree Conversion for Exact Inference 1

Scalable Parallel Implementation of Bayesian Network to Junction Tree Conversion for Exact Inference 1 Scalable Parallel Implementation of Bayesian Network to Junction Tree Conversion for Exact Inference Vasanth Krishna Namasivayam, Animesh Pathak and Viktor K. Prasanna Department of Electrical Engineering

More information

Belief propagation in a bucket-tree. Handouts, 275B Fall Rina Dechter. November 1, 2000

Belief propagation in a bucket-tree. Handouts, 275B Fall Rina Dechter. November 1, 2000 Belief propagation in a bucket-tree Handouts, 275B Fall-2000 Rina Dechter November 1, 2000 1 From bucket-elimination to tree-propagation The bucket-elimination algorithm, elim-bel, for belief updating

More information

Mean Field and Variational Methods finishing off

Mean Field and Variational Methods finishing off Readings: K&F: 10.1, 10.5 Mean Field and Variational Methods finishing off Graphical Models 10708 Carlos Guestrin Carnegie Mellon University November 5 th, 2008 10-708 Carlos Guestrin 2006-2008 1 10-708

More information

Lecture 5: Exact inference

Lecture 5: Exact inference Lecture 5: Exact inference Queries Inference in chains Variable elimination Without evidence With evidence Complexity of variable elimination which has the highest probability: instantiation of all other

More information

Machine Learning / Jan 27, 2010

Machine Learning / Jan 27, 2010 Revisiting Logistic Regression & Naïve Bayes Aarti Singh Machine Learning 10-701/15-781 Jan 27, 2010 Generative and Discriminative Classifiers Training classifiers involves learning a mapping f: X -> Y,

More information

Part I: Sum Product Algorithm and (Loopy) Belief Propagation. What s wrong with VarElim. Forwards algorithm (filtering) Forwards-backwards algorithm

Part I: Sum Product Algorithm and (Loopy) Belief Propagation. What s wrong with VarElim. Forwards algorithm (filtering) Forwards-backwards algorithm OU 56 Probabilistic Graphical Models Loopy Belief Propagation and lique Trees / Join Trees lides from Kevin Murphy s Graphical Model Tutorial (with minor changes) eading: Koller and Friedman h 0 Part I:

More information

ECE521 W17 Tutorial 10

ECE521 W17 Tutorial 10 ECE521 W17 Tutorial 10 Shenlong Wang and Renjie Liao *Some of materials are credited to Jimmy Ba, Eric Sudderth, Chris Bishop Introduction to A4 1, Graphical Models 2, Message Passing 3, HMM Introduction

More information

Inference Complexity As Learning Bias. The Goal Outline. Don t use model complexity as your learning bias

Inference Complexity As Learning Bias. The Goal Outline. Don t use model complexity as your learning bias Inference Complexity s Learning Bias Daniel Lowd Dept. of Computer and Information Science University of Oregon Don t use model complexity as your learning bias Use inference complexity. Joint work with

More information

10-701/15-781, Fall 2006, Final

10-701/15-781, Fall 2006, Final -7/-78, Fall 6, Final Dec, :pm-8:pm There are 9 questions in this exam ( pages including this cover sheet). If you need more room to work out your answer to a question, use the back of the page and clearly

More information

Introduction to Graphical Models

Introduction to Graphical Models Robert Collins CSE586 Introduction to Graphical Models Readings in Prince textbook: Chapters 10 and 11 but mainly only on directed graphs at this time Credits: Several slides are from: Review: Probability

More information

Compiling Probabilistic Graphical Models using Sentential Decision Diagrams

Compiling Probabilistic Graphical Models using Sentential Decision Diagrams Compiling Probabilistic Graphical Models using Sentential Decision Diagrams Arthur Choi, Doga Kisa, and Adnan Darwiche University of California, Los Angeles, California 90095, USA {aychoi,doga,darwiche}@cs.ucla.edu

More information

Automatic Scaling Iterative Computations. Aug. 7 th, 2012

Automatic Scaling Iterative Computations. Aug. 7 th, 2012 Automatic Scaling Iterative Computations Guozhang Wang Cornell University Aug. 7 th, 2012 1 What are Non-Iterative Computations? Non-iterative computation flow Directed Acyclic Examples Batch style analytics

More information

V,T C3: S,L,B T C4: A,L,T A,L C5: A,L,B A,B C6: C2: X,A A

V,T C3: S,L,B T C4: A,L,T A,L C5: A,L,B A,B C6: C2: X,A A Inference II Daphne Koller Stanford University CS228 Handout #13 In the previous chapter, we showed how efficient inference can be done in a BN using an algorithm called Variable Elimination, that sums

More information

ECE521 Lecture 21 HMM cont. Message Passing Algorithms

ECE521 Lecture 21 HMM cont. Message Passing Algorithms ECE521 Lecture 21 HMM cont Message Passing Algorithms Outline Hidden Markov models Numerical example of figuring out marginal of the observed sequence Numerical example of figuring out the most probable

More information

Probabilistic Graphical Models

Probabilistic Graphical Models Probabilistic Graphical Models Raquel Urtasun and Tamir Hazan TTI Chicago April 8, 2011 Raquel Urtasun and Tamir Hazan (TTI-C) Graphical Models April 8, 2011 1 / 19 Factor Graphs H does not reveal the

More information

Performance potential for simulating spin models on GPU

Performance potential for simulating spin models on GPU Performance potential for simulating spin models on GPU Martin Weigel Institut für Physik, Johannes-Gutenberg-Universität Mainz, Germany 11th International NTZ-Workshop on New Developments in Computational

More information

Parallel Methods for Verifying the Consistency of Weakly-Ordered Architectures. Adam McLaughlin, Duane Merrill, Michael Garland, and David A.

Parallel Methods for Verifying the Consistency of Weakly-Ordered Architectures. Adam McLaughlin, Duane Merrill, Michael Garland, and David A. Parallel Methods for Verifying the Consistency of Weakly-Ordered Architectures Adam McLaughlin, Duane Merrill, Michael Garland, and David A. Bader Challenges of Design Verification Contemporary hardware

More information

Mean Field and Variational Methods finishing off

Mean Field and Variational Methods finishing off Readings: K&F: 10.1, 10.5 Mean Field and Variational Methods finishing off Graphical Models 10708 Carlos Guestrin Carnegie Mellon University November 5 th, 2008 10-708 Carlos Guestrin 2006-2008 1 10-708

More information

Search Algorithms for Solving Queries on Graphical Models & the Importance of Pseudo-trees in their Complexity.

Search Algorithms for Solving Queries on Graphical Models & the Importance of Pseudo-trees in their Complexity. Search Algorithms for Solving Queries on Graphical Models & the Importance of Pseudo-trees in their Complexity. University of California, Irvine CS199: Individual Study with Rina Dechter Héctor Otero Mediero

More information

2. Graphical Models. Undirected pairwise graphical models. Factor graphs. Bayesian networks. Conversion between graphical models. Graphical Models 2-1

2. Graphical Models. Undirected pairwise graphical models. Factor graphs. Bayesian networks. Conversion between graphical models. Graphical Models 2-1 Graphical Models 2-1 2. Graphical Models Undirected pairwise graphical models Factor graphs Bayesian networks Conversion between graphical models Graphical Models 2-2 Graphical models Families of graphical

More information

The exam is closed book, closed notes except your one-page cheat sheet.

The exam is closed book, closed notes except your one-page cheat sheet. CS 189 Fall 2015 Introduction to Machine Learning Final Please do not turn over the page before you are instructed to do so. You have 2 hours and 50 minutes. Please write your initials on the top-right

More information

Lecture 13: May 10, 2002

Lecture 13: May 10, 2002 EE96 Pat. Recog. II: Introduction to Graphical Models University of Washington Spring 00 Dept. of Electrical Engineering Lecture : May 0, 00 Lecturer: Jeff Bilmes Scribe: Arindam Mandal, David Palmer(000).

More information

Machine Learning. Sourangshu Bhattacharya

Machine Learning. Sourangshu Bhattacharya Machine Learning Sourangshu Bhattacharya Bayesian Networks Directed Acyclic Graph (DAG) Bayesian Networks General Factorization Curve Fitting Re-visited Maximum Likelihood Determine by minimizing sum-of-squares

More information

Using Machine Learning to Optimize Storage Systems

Using Machine Learning to Optimize Storage Systems Using Machine Learning to Optimize Storage Systems Dr. Kiran Gunnam 1 Outline 1. Overview 2. Building Flash Models using Logistic Regression. 3. Storage Object classification 4. Storage Allocation recommendation

More information

A New Approach For Convert Multiply-Connected Trees in Bayesian networks

A New Approach For Convert Multiply-Connected Trees in Bayesian networks A New Approach For Convert Multiply-Connected Trees in Bayesian networks 1 Hussein Baloochian, Alireza khantimoory, 2 Saeed Balochian 1 Islamic Azad university branch of zanjan 2 Islamic Azad university

More information

Bayesian Classification Using Probabilistic Graphical Models

Bayesian Classification Using Probabilistic Graphical Models San Jose State University SJSU ScholarWorks Master's Projects Master's Theses and Graduate Research Spring 2014 Bayesian Classification Using Probabilistic Graphical Models Mehal Patel San Jose State University

More information

Belief Updating in Bayes Networks (1)

Belief Updating in Bayes Networks (1) Belief Updating in Bayes Networks (1) Mark Fishel, Mihkel Pajusalu, Roland Pihlakas Bayes Networks p. 1 Some Concepts Belief: probability Belief updating: probability calculating based on a BN (model,

More information

2. Graphical Models. Undirected graphical models. Factor graphs. Bayesian networks. Conversion between graphical models. Graphical Models 2-1

2. Graphical Models. Undirected graphical models. Factor graphs. Bayesian networks. Conversion between graphical models. Graphical Models 2-1 Graphical Models 2-1 2. Graphical Models Undirected graphical models Factor graphs Bayesian networks Conversion between graphical models Graphical Models 2-2 Graphical models There are three families of

More information

Improving performances of an embedded RDBMS with a hybrid CPU/GPU processing engine

Improving performances of an embedded RDBMS with a hybrid CPU/GPU processing engine Improving performances of an embedded RDBMS with a hybrid CPU/GPU processing engine Samuel Cremer 1,2, Michel Bagein 1, Saïd Mahmoudi 1, Pierre Manneback 1 1 UMONS, University of Mons Computer Science

More information

CS 188: Artificial Intelligence Fall Machine Learning

CS 188: Artificial Intelligence Fall Machine Learning CS 188: Artificial Intelligence Fall 2007 Lecture 23: Naïve Bayes 11/15/2007 Dan Klein UC Berkeley Machine Learning Up till now: how to reason or make decisions using a model Machine learning: how to select

More information

Sequence Labeling: The Problem

Sequence Labeling: The Problem Sequence Labeling: The Problem Given a sequence (in NLP, words), assign appropriate labels to each word. For example, POS tagging: DT NN VBD IN DT NN. The cat sat on the mat. 36 part-of-speech tags used

More information

DIFFERENTIAL. Tomáš Oberhuber, Atsushi Suzuki, Jan Vacata, Vítězslav Žabka

DIFFERENTIAL. Tomáš Oberhuber, Atsushi Suzuki, Jan Vacata, Vítězslav Žabka USE OF FOR Tomáš Oberhuber, Atsushi Suzuki, Jan Vacata, Vítězslav Žabka Faculty of Nuclear Sciences and Physical Engineering Czech Technical University in Prague Mini workshop on advanced numerical methods

More information

LOCAL STRUCTURE AND DETERMINISM IN PROBABILISTIC DATABASES. Theodoros Rekatsinas, Amol Deshpande, Lise Getoor

LOCAL STRUCTURE AND DETERMINISM IN PROBABILISTIC DATABASES. Theodoros Rekatsinas, Amol Deshpande, Lise Getoor LOCAL STRUCTURE AND DETERMINISM IN PROBABILISTIC DATABASES Theodoros Rekatsinas, Amol Deshpande, Lise Getoor Motivation Probabilistic databases store, manage and query uncertain data Numerous applications

More information

Machine Learning. Supervised Learning. Manfred Huber

Machine Learning. Supervised Learning. Manfred Huber Machine Learning Supervised Learning Manfred Huber 2015 1 Supervised Learning Supervised learning is learning where the training data contains the target output of the learning system. Training data D

More information

Machine Learning: Think Big and Parallel

Machine Learning: Think Big and Parallel Day 1 Inderjit S. Dhillon Dept of Computer Science UT Austin CS395T: Topics in Multicore Programming Oct 1, 2013 Outline Scikit-learn: Machine Learning in Python Supervised Learning day1 Regression: Least

More information

Finding Non-overlapping Clusters for Generalized Inference Over Graphical Models

Finding Non-overlapping Clusters for Generalized Inference Over Graphical Models 1 Finding Non-overlapping Clusters for Generalized Inference Over Graphical Models Divyanshu Vats and José M. F. Moura arxiv:1107.4067v2 [stat.ml] 18 Mar 2012 Abstract Graphical models use graphs to compactly

More information

Escola Politécnica, University of São Paulo Av. Prof. Mello Moraes, 2231, , São Paulo, SP - Brazil

Escola Politécnica, University of São Paulo Av. Prof. Mello Moraes, 2231, , São Paulo, SP - Brazil Generalizing Variable Elimination in Bayesian Networks FABIO GAGLIARDI COZMAN Escola Politécnica, University of São Paulo Av. Prof. Mello Moraes, 2231, 05508-900, São Paulo, SP - Brazil fgcozman@usp.br

More information

Probabilistic Graphical Models

Probabilistic Graphical Models Overview of Part One Probabilistic Graphical Models Part One: Graphs and Markov Properties Christopher M. Bishop Graphs and probabilities Directed graphs Markov properties Undirected graphs Examples Microsoft

More information

Collective classification in network data

Collective classification in network data 1 / 50 Collective classification in network data Seminar on graphs, UCSB 2009 Outline 2 / 50 1 Problem 2 Methods Local methods Global methods 3 Experiments Outline 3 / 50 1 Problem 2 Methods Local methods

More information

CS 343: Artificial Intelligence

CS 343: Artificial Intelligence CS 343: Artificial Intelligence Bayes Nets: Inference Prof. Scott Niekum The University of Texas at Austin [These slides based on those of Dan Klein and Pieter Abbeel for CS188 Intro to AI at UC Berkeley.

More information

CS839: Probabilistic Graphical Models. Lecture 10: Learning with Partially Observed Data. Theo Rekatsinas

CS839: Probabilistic Graphical Models. Lecture 10: Learning with Partially Observed Data. Theo Rekatsinas CS839: Probabilistic Graphical Models Lecture 10: Learning with Partially Observed Data Theo Rekatsinas 1 Partially Observed GMs Speech recognition 2 Partially Observed GMs Evolution 3 Partially Observed

More information

How to Optimize Geometric Multigrid Methods on GPUs

How to Optimize Geometric Multigrid Methods on GPUs How to Optimize Geometric Multigrid Methods on GPUs Markus Stürmer, Harald Köstler, Ulrich Rüde System Simulation Group University Erlangen March 31st 2011 at Copper Schedule motivation imaging in gradient

More information

Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching

Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching Correlated Correspondences [ASP*04] A Complete Registration System [HAW*08] In this session Advanced Global Matching Some practical

More information

Cluster Analysis. Jia Li Department of Statistics Penn State University. Summer School in Statistics for Astronomers IV June 9-14, 2008

Cluster Analysis. Jia Li Department of Statistics Penn State University. Summer School in Statistics for Astronomers IV June 9-14, 2008 Cluster Analysis Jia Li Department of Statistics Penn State University Summer School in Statistics for Astronomers IV June 9-1, 8 1 Clustering A basic tool in data mining/pattern recognition: Divide a

More information

Parallel Gibbs Sampling From Colored Fields to Thin Junction Trees

Parallel Gibbs Sampling From Colored Fields to Thin Junction Trees Parallel Gibbs Sampling From Colored Fields to Thin Junction Trees Joseph Gonzalez Yucheng Low Arthur Gretton Carlos Guestrin Draw Samples Sampling as an Inference Procedure Suppose we wanted to know the

More information

Compiling Bayesian Networks by Symbolic Probability Calculation Based on Zero-suppressed BDDs

Compiling Bayesian Networks by Symbolic Probability Calculation Based on Zero-suppressed BDDs Compiling Bayesian Networks by Symbolic Probability Calculation Based on Zero-suppressed BDDs Shin-ichi Minato Div. of Computer Science Hokkaido University Sapporo 6 814, Japan Ken Satoh National Institute

More information

PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 8: GRAPHICAL MODELS

PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 8: GRAPHICAL MODELS PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 8: GRAPHICAL MODELS Bayesian Networks Directed Acyclic Graph (DAG) Bayesian Networks General Factorization Bayesian Curve Fitting (1) Polynomial Bayesian

More information

Av. Prof. Mello Moraes, 2231, , São Paulo, SP - Brazil

Av. Prof. Mello Moraes, 2231, , São Paulo, SP - Brazil " Generalizing Variable Elimination in Bayesian Networks FABIO GAGLIARDI COZMAN Escola Politécnica, University of São Paulo Av Prof Mello Moraes, 31, 05508-900, São Paulo, SP - Brazil fgcozman@uspbr Abstract

More information

Machine Learning A W 1sst KU. b) [1 P] For the probability distribution P (A, B, C, D) with the factorization

Machine Learning A W 1sst KU. b) [1 P] For the probability distribution P (A, B, C, D) with the factorization Machine Learning A 708.064 13W 1sst KU Exercises Problems marked with * are optional. 1 Conditional Independence a) [1 P] For the probability distribution P (A, B, C, D) with the factorization P (A, B,

More information

Building Classifiers using Bayesian Networks

Building Classifiers using Bayesian Networks Building Classifiers using Bayesian Networks Nir Friedman and Moises Goldszmidt 1997 Presented by Brian Collins and Lukas Seitlinger Paper Summary The Naive Bayes classifier has reasonable performance

More information

K-Means and Gaussian Mixture Models

K-Means and Gaussian Mixture Models K-Means and Gaussian Mixture Models David Rosenberg New York University June 15, 2015 David Rosenberg (New York University) DS-GA 1003 June 15, 2015 1 / 43 K-Means Clustering Example: Old Faithful Geyser

More information

ECE 6504: Advanced Topics in Machine Learning Probabilistic Graphical Models and Large-Scale Learning

ECE 6504: Advanced Topics in Machine Learning Probabilistic Graphical Models and Large-Scale Learning ECE 6504: Advanced Topics in Machine Learning Probabilistic Graphical Models and Large-Scale Learning Topics Markov Random Fields: Inference Exact: VE Exact+Approximate: BP Readings: Barber 5 Dhruv Batra

More information

Computational Intelligence

Computational Intelligence Computational Intelligence A Logical Approach Problems for Chapter 10 Here are some problems to help you understand the material in Computational Intelligence: A Logical Approach. They are designed to

More information

6 : Factor Graphs, Message Passing and Junction Trees

6 : Factor Graphs, Message Passing and Junction Trees 10-708: Probabilistic Graphical Models 10-708, Spring 2018 6 : Factor Graphs, Message Passing and Junction Trees Lecturer: Kayhan Batmanghelich Scribes: Sarthak Garg 1 Factor Graphs Factor Graphs are graphical

More information

Visual Analysis of Lagrangian Particle Data from Combustion Simulations

Visual Analysis of Lagrangian Particle Data from Combustion Simulations Visual Analysis of Lagrangian Particle Data from Combustion Simulations Hongfeng Yu Sandia National Laboratories, CA Ultrascale Visualization Workshop, SC11 Nov 13 2011, Seattle, WA Joint work with Jishang

More information

Machine Learning. Chao Lan

Machine Learning. Chao Lan Machine Learning Chao Lan Machine Learning Prediction Models Regression Model - linear regression (least square, ridge regression, Lasso) Classification Model - naive Bayes, logistic regression, Gaussian

More information

Scene Grammars, Factor Graphs, and Belief Propagation

Scene Grammars, Factor Graphs, and Belief Propagation Scene Grammars, Factor Graphs, and Belief Propagation Pedro Felzenszwalb Brown University Joint work with Jeroen Chua Probabilistic Scene Grammars General purpose framework for image understanding and

More information

Machine Learning A WS15/16 1sst KU Version: January 11, b) [1 P] For the probability distribution P (A, B, C, D) with the factorization

Machine Learning A WS15/16 1sst KU Version: January 11, b) [1 P] For the probability distribution P (A, B, C, D) with the factorization Machine Learning A 708.064 WS15/16 1sst KU Version: January 11, 2016 Exercises Problems marked with * are optional. 1 Conditional Independence I [3 P] a) [1 P] For the probability distribution P (A, B,

More information

Challenges Simulating Real Fuel Combustion Kinetics: The Role of GPUs

Challenges Simulating Real Fuel Combustion Kinetics: The Role of GPUs Challenges Simulating Real Fuel Combustion Kinetics: The Role of GPUs M. J. McNenly and R. A. Whitesides GPU Technology Conference March 27, 2014 San Jose, CA LLNL-PRES-652254! This work performed under

More information

Treewidth: Preprocessing and Kernelization

Treewidth: Preprocessing and Kernelization Treewidth: Preprocessing and Kernelization Hans L. Bodlaender Joint work with Arie Koster, Frank van den Eijkhof, Bart Jansen, Stefan Kratsch, Vincent Kreuzen 1 This talk Survey of work on preprocessing

More information

CS 188: Artificial Intelligence

CS 188: Artificial Intelligence CS 188: Artificial Intelligence Bayes Nets: Inference Instructors: Dan Klein and Pieter Abbeel --- University of California, Berkeley [These slides were created by Dan Klein and Pieter Abbeel for CS188

More information

Machine Learning

Machine Learning Machine Learning 10-601 Tom M. Mitchell Machine Learning Department Carnegie Mellon University April 1, 2019 Today: Inference in graphical models Learning graphical models Readings: Bishop chapter 8 Bayesian

More information

CSE 591: GPU Programming. Introduction. Entertainment Graphics: Virtual Realism for the Masses. Computer games need to have: Klaus Mueller

CSE 591: GPU Programming. Introduction. Entertainment Graphics: Virtual Realism for the Masses. Computer games need to have: Klaus Mueller Entertainment Graphics: Virtual Realism for the Masses CSE 591: GPU Programming Introduction Computer games need to have: realistic appearance of characters and objects believable and creative shading,

More information

Scene Grammars, Factor Graphs, and Belief Propagation

Scene Grammars, Factor Graphs, and Belief Propagation Scene Grammars, Factor Graphs, and Belief Propagation Pedro Felzenszwalb Brown University Joint work with Jeroen Chua Probabilistic Scene Grammars General purpose framework for image understanding and

More information

Unsupervised Learning. CS 3793/5233 Artificial Intelligence Unsupervised Learning 1

Unsupervised Learning. CS 3793/5233 Artificial Intelligence Unsupervised Learning 1 Unsupervised CS 3793/5233 Artificial Intelligence Unsupervised 1 EM k-means Procedure Data Random Assignment Assign 1 Assign 2 Soft k-means In clustering, the target feature is not given. Goal: Construct

More information

Learning the Structure of Sum-Product Networks. Robert Gens Pedro Domingos

Learning the Structure of Sum-Product Networks. Robert Gens Pedro Domingos Learning the Structure of Sum-Product Networks Robert Gens Pedro Domingos w 20 10x O(n) X Y LL PLL CLL CMLL Motivation SPN Structure Experiments Review Learning Graphical Models Representation Inference

More information

Scalable Multi Agent Simulation on the GPU. Avi Bleiweiss NVIDIA Corporation San Jose, 2009

Scalable Multi Agent Simulation on the GPU. Avi Bleiweiss NVIDIA Corporation San Jose, 2009 Scalable Multi Agent Simulation on the GPU Avi Bleiweiss NVIDIA Corporation San Jose, 2009 Reasoning Explicit State machine, serial Implicit Compute intensive Fits SIMT well Collision avoidance Motivation

More information

Utilizing Device Behavior in Structure-Based Diagnosis

Utilizing Device Behavior in Structure-Based Diagnosis Utilizing Device Behavior in Structure-Based Diagnosis Adnan Darwiche Cognitive Systems Laboratory Department of Computer Science University of California Los Angeles, CA 90024 darwiche @cs. ucla. edu

More information

Machine Learning

Machine Learning Machine Learning 10-601 Tom M. Mitchell Machine Learning Department Carnegie Mellon University October 2, 2012 Today: Graphical models Bayes Nets: Representing distributions Conditional independencies

More information

CS 664 Flexible Templates. Daniel Huttenlocher

CS 664 Flexible Templates. Daniel Huttenlocher CS 664 Flexible Templates Daniel Huttenlocher Flexible Template Matching Pictorial structures Parts connected by springs and appearance models for each part Used for human bodies, faces Fischler&Elschlager,

More information

TUNING CUDA APPLICATIONS FOR MAXWELL

TUNING CUDA APPLICATIONS FOR MAXWELL TUNING CUDA APPLICATIONS FOR MAXWELL DA-07173-001_v7.0 March 2015 Application Note TABLE OF CONTENTS Chapter 1. Maxwell Tuning Guide... 1 1.1. NVIDIA Maxwell Compute Architecture... 1 1.2. CUDA Best Practices...2

More information

Efficient Caching in Elimination Trees

Efficient Caching in Elimination Trees Efficient Caching in Elimination Trees Kevin Grant and Michael C. Horsch epartment Of Computer Science University of Saskatchewan 176 Thorvaldson Building, 110 Science Place Saskatoon, Saskatchewan S7N

More information

Parallel Evidence Propagation on Multicore Processors

Parallel Evidence Propagation on Multicore Processors Parallel Evidence Propagation on Multicore Processors Yinglong Xia 1, Xiaojun Feng 3 and Viktor K. Prasanna 2,1 Computer Science Department 1 and Department of Electrical Engineering 2 University of Southern

More information

Machine Learning

Machine Learning Machine Learning 10-601 Tom M. Mitchell Machine Learning Department Carnegie Mellon University February 18, 2015 Today: Graphical models Bayes Nets: Representing distributions Conditional independencies

More information

A Comparison of Lauritzen-Spiegelhalter, Hugin, and Shenoy-Shafer Architectures for Computing Marginals of Probability Distributions

A Comparison of Lauritzen-Spiegelhalter, Hugin, and Shenoy-Shafer Architectures for Computing Marginals of Probability Distributions Appeared in: G. F. Cooper & S. Moral (eds.), Uncertainty in Artificial Intelligence, Vol. 14, 1999, pp. 328--337, Morgan Kaufmann, San Francisco, CA. A Comparison of Lauritzen-Spiegelhalter, Hugin, and

More information

Algorithms for Markov Random Fields in Computer Vision

Algorithms for Markov Random Fields in Computer Vision Algorithms for Markov Random Fields in Computer Vision Dan Huttenlocher November, 2003 (Joint work with Pedro Felzenszwalb) Random Field Broadly applicable stochastic model Collection of n sites S Hidden

More information

Preface to the Second Edition. Preface to the First Edition. 1 Introduction 1

Preface to the Second Edition. Preface to the First Edition. 1 Introduction 1 Preface to the Second Edition Preface to the First Edition vii xi 1 Introduction 1 2 Overview of Supervised Learning 9 2.1 Introduction... 9 2.2 Variable Types and Terminology... 9 2.3 Two Simple Approaches

More information

The exam is closed book, closed notes except your one-page (two-sided) cheat sheet.

The exam is closed book, closed notes except your one-page (two-sided) cheat sheet. CS 189 Spring 2015 Introduction to Machine Learning Final You have 2 hours 50 minutes for the exam. The exam is closed book, closed notes except your one-page (two-sided) cheat sheet. No calculators or

More information

THE estimated cyber-security market is expected to grow

THE estimated cyber-security market is expected to grow 1 Exact Inference Techniques for the Analysis of Bayesian Attack Graphs Luis Muñoz-González, Daniele Sgandurra, Martín Barrère, and Emil C. Lupu arxiv:1510.02427v2 [cs.cr] 4 Nov 2016 Abstract Attack graphs

More information