Scalable Gaussian process models on matrices and tensors

Size: px
Start display at page:

Download "Scalable Gaussian process models on matrices and tensors"

Transcription

1 Scalable Gaussian process models on matrices and tensors Alan Qi CS & Statistics Purdue University Joint work with F. Yan, Z. Xu, S. Zhe, and IBM Research!

2 Models for graph and multiway data Model Algorithm 1. STGP: predict who may be your friends on facebook Data 2. DinTucker: distributed online inference for large graphs and tensors

3 Outline Nonpara. Bayesian: Gaussian processes Utilizing Kronecker product STGP Tensor data DinTucker

4 Tensor data: multiple aspects Patients Biomarkers Medicines Yi,j,k: the value of the k-th biomarker (i.e., cell population) for the j-th patient after taking the i-th medicine Predict drug response

5 Matrix data: networks people people Yi,j: 1 if node i is linked to node j, 0 otherwise. Discover communities and predict unknown interactions

6 Goals! - Predict unknown elements (e.g., drug response and network interactions) - Identify latent multi-aspect groups (communities)

7 Classical Tucker decomposition Generalization of matrix factorization 3D case: U core tensor loading matrices Y Z V Y = G 1 U 2 Z 3 V y ijk = X X X g rst u ir z js v ks r s t Sun et al. 2008

8 Assumptions Complete Continuous Multi-linear

9 Solution Gaussian process models for tensors

10 Gaussian processes: prior on scalar f Powerful Bayesian modeling tool Numerous applications: computer vision, sensor networks, computational biology, etc. f x f N(0; K x )

11 GP models on tensor F GP on a tensor: stochastic process in an infinite tensor space Tensor N (F 0; K u, K v, K z )=(2 ) 3N 2 Y s=u,v,z K s N 2 2 Projection of GP on any tensor of finite size follows a tensor-variate Gaussian distribution exp{ 1 2 (F 1 K 1 u 2 K 1 v 3 K 1 z ) F 2 }

12 Latent sparse GP on tensors Patients Biomarkers (i, j, k) u i : Medicines Element u i : z j : v k : (i, j, k) is characterized by Sparse loading vector in latent medicine groups Sparse loading vector in latent patient groups Sparse loading vector in latent biomarker groups

13 Separate covariance for each dimension Kz K v Patients Biomarkers i r K u (i, r) =k(u i, u r ) Medicines Nonlinear relationship between medicines i and r -Separate covariance/kernel function for each dimension! -The more similar loading vectors, the larger the covariance function value

14 Predict unknown tensor elements Kv Patients K v Biomarkers (i, j, k) Medicines K u w (r, s, t) Simple illustration: 1) Based on observed data, estimate loading vectors: [u i, z j, v k ] 2) Compute weights (similarities) between unknown and observed elements: w(ijk, rst) w([u i, z j, v k ], [u r, z s, v t ]) 3) Predict the unknown element: y i,j,k = P r,s,t w(ijk, rst)y r,s,t

15 Graphical model U =[u 1,...,u N ] U V Z Sparse loading vectors u i exp( u i ) Similarly, sample V and Z F Latent tensor F N(0; K u, K v, K z ) Y U Unknown data Y O Observed data Y O p(y ijk f ijk ) Y (i,j,k)2o p(y ijk f ijk ) : Gaussian for continuous data Probit for binary data Possion for count data

16 Benefits Handle binary and missing data Discover block/group structures Avoid overfitting Model prediction uncertainty Incorporate additional side information!! Yan, Xu & Qi, 2011; Xu, Yan & Qi 2011

17 Two views of the same model Infinite latent factor model / nonlinear stochastic blockmodels! Nonparametric prior on infinite exchangeable (high-dimensional) arrays

18 Infinite nonlinear tensor decomposition (z K ) Nonlinear feature mapping: infinite dimension F G (u I ) (v J ) f ijk = G 1 (u i ) 2 (v j ) 3 (z k ) G N(0, I, I, I) y ijk = f ijk + e Integrating out p(f U, V, Z) = Z G gives GP on tensor F p(f G, U, V, Z)p(G)dG = N (0, K u, K v, K z ) where k(u i, u j )= (u i ) T (u j )

19 Exchangeability and De Finetti theorem An exchangeable sequence: its joint distribution is invariant under arbitrary permutation of the indices. De Finetti (1931) and Hewitt and Savage (1955) showed: If X 1,...,X n,... is exchangeable. Then there exists a probability measure µ on probability measure, such that Z! p(x 1 A 1,...,X n A n )= Q(A 1 ) Q(A n )µ(dq) How about priors on exchangeable arrays representing graphs?

20 Prior for random graphs Aldous (1981); Diaconis & Freedman (1981) (See Lauritzen 2007) showed: Any column-row exchangeable binary matrix can be generated as follows:!!! f p U i,v j W IID Uniform(0, 1) G ij W, U i,v j Bernoulli(f(U i,v j )) Transform U i and V j to other r.v.s via suitable transformations.

21 Prior for random graphs For weakly exchangeable arrays (shared permutations over rows and columns; for undirected graphs), our model (Yan et al. 2011) is f GP U i,v j W IID Uniform(0, 1) G ij W, U i,v j Bernoulli(f(U i,u j )) Replace Bernoulli by Gaussian for continuous values Special case of Random function priors (Lloyd et al., 2012)

22 Related works Latent Eigenmodel: (Hoff 2007): parametric models on exchangeable arrays GP-LVM (Lawrence 2005): modeling only interactions in rows or columns Random function priors (Lloyd et al., 2012)

23 Algorithm: Variational EM Marginal likelihood log p(y O U, V, Z) + log p(u, V, Z) Variational approximation Iterations

24 E step: Variational-EM Maximize the variational lower bound over q(f) & q(s) : Z log p(y O U, V, Z) q(f)q(s) log p(y O, F, S U, V, Z)dF+ where model S H q (F)+H q (S) L(q(F),q(S)) are parameters or auxiliary Z variables of noise and q(f) log q(f)df. p(y ijk f ijk ) H q (F) = M step: Maximize the variational lower bound over U, Vand Z.

25 Algorithm: explore model structures Example: Trace{(I + K u K v K z ) 1 } Direct computation: Matrix inversion N 3 by N 3 O(N 9 ) Kronecker product operation:

26 Properties of Kronecker product Properties: Theorem: Eigen-decomposition K = W W T If K = K u K v K z then W = W u W v W z = u v z W u : eigenvectors of K u diag{ u } : eigenvalues of K u

27 Reduced computational complexity Example: Trace{(I + K u K v K z ) 1 } Direct computation Using this theorem and trace properties O(N 9 ) NX NX NX i=1 j=1 k= u i v j z k O(N 3 ) = MX MX MX i=1 j=1 k= u i v j z k O(N 2 M) M<<N

28 2D case: GP stochastic blockmodels - Undirected networks (friend relationships and protein-protein interactions) - Represented by symmetric adjacent matrices Yan, Xu & Qi, 2011; Xu, Yan & Qi 2011

29 2D: Coauthor networks Membership produced by LEM Area Under Curve AUC values SMGB MMSB LEM Ours Number of latent groups Groups Groups Groups Nodes Membership produced by MMSB Nodes Membership produced by SMGB Nodes NIPS authors Ours Co-authorship dataset: co-authorship links from100 authors who have the largest number of co-authors from NIPS 1-17.

30 3D: Enron s Enron dataset: s from senior management of Enron before its bankruptcy in 2001.! 3D tensor representation: Sender-Recipient-Subject Area Under Curve AUC values InfTucker tp InfTucker gp CP TD HOSVD NCP WCP PTD Ours Number of Factors

31 4D: Digg Area Under Curve AUC values Ours InfTucker tp InfTucker gp CP TD HOSVD NCP WCP PTD Number of Factors Digg dataset: Social news from digg.com 4D tensor representation: user-news-keywords-category

32 4D: Digg latent groups Keyword groups learned by STGP: iphone Olympics american journey apple games australian dream internet china rothschild action *seo beijing moscow wisdom photosynth red century top *seo: search engine optimization

33 Next steps Better models for sparse graphs/ tensors? Distributed inference for big graphs/tensors

34 Outline Hierarchical Bayesian Distributed online inf. STGP NELL, fmri,... DinTucker

35 Big data (with big issues) - High computational cost: O(d 3 ) where d is the size of the largest mode. - Not enough memory to store the data in a single machine - Data are physically stored in different locations

36 Challenge for GPs on arrays Coupled array elements: a huge covariance over all them.! Obvious solutions, ADMM or stochastic gradient descent, do not apply due to the partition function.

37 Solution: Hierarchical Bayesian modeling Local training over subarrays! Global information sharing via the parent nodes

38 Solution: Hierarchical Bayesian modeling {Y Y n }. U Global memberships Ũ n Local memberships M nt Latent tensors Y nt Subarrays t=1... T n n=1... N

39 Solution: Hierarchical Bayesian modeling {Y Y n }. U Global memberships Ũ n Local memberships M nt Latent tensors Local training:! Mapper Y nt Subarrays t=1... T n n=1... N

40 Hierarchical Bayesian on MapReduce {Y Y n }. U Ũ n Global memberships Local memberships Global sharing:! Reducer M nt Latent tensors Local VB inf.:! Mapper Y nt Subarrays t=1... T n n=1... N

41 Subarray sampling strategies Uniform! Weighted! Grid

42 Three questions How does the new distributed online inference compare with the sequential inference? How does it scale in terms of number of CPUs? How does it compare with the state-ofthe-art large-scale tensor decomposition method?

43 Same accuracy as sequential inference PARAFAC NPARAFAC HOSVD Tucker InfTucker DinTucker U DinTucker W DinTucker G AUC DinTucker with! a specific sampling! strategy Number of Factors (b) digg2 AUC Number of Factors (c) enron

44 Nearly linear scalability 3.5 Scales up: Rn/R Number of Machines

45 More accurate, less time Running time (hours) DinTucker GigaTensor AUC Subarray Size (I=J=K) (a) NELL: running time 0.83 GigaTensor DinTucker G DinTucker U DinTucker W (b) NELL: prediction

46 More accurate, less time Running time (hours) DinTucker GigaTensor AUC Subarray Size (I=J=K) (c) ACC: running time 0.95 GigaTensor DinTucker G (d) ACC: prediction DinTucker U DinTucker W

47 Conclusions

48 Scalable nonparametric Bayesian methods for graphs Model Algorithm STGP: predict who may be your friends on facebook Data DinTucker: distributed online inference for large graphs and tensors

49 Acknowledgment! Funding agencies: - NSF: CAREER, Robust Intelligence, CDI, STC - NIH: R01 - Microsoft - IBM - Showalter foundation - Indiana Clinical & Translational Science Institute - Purdue

DinTucker: Scaling Up Gaussian Process Models on Large Multidimensional Arrays

DinTucker: Scaling Up Gaussian Process Models on Large Multidimensional Arrays Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI-16) DinTucker: Scaling Up Gaussian Process Models on Large Multidimensional Arrays Shandian Zhe, 1 Yuan Qi, 1 Youngja Park,

More information

Nonparametric Bayesian Matrix Factorization for Assortative Networks

Nonparametric Bayesian Matrix Factorization for Assortative Networks Nonparametric Bayesian Matrix Factorization for Assortative Networks Mingyuan Zhou IROM Department, McCombs School of Business Department of Statistics and Data Sciences The University of Texas at Austin

More information

Random function priors for exchangeable arrays with applications to graphs and relational data

Random function priors for exchangeable arrays with applications to graphs and relational data Random function priors for exchangeable arrays with applications to graphs and relational data James Robert Lloyd Department of Engineering University of Cambridge Peter Orbanz Department of Statistics

More information

Distributed Flexible Nonlinear Tensor Factorization

Distributed Flexible Nonlinear Tensor Factorization Distributed Flexible Nonlinear Tensor Factorization Shandian Zhe, Kai Zhang, Pengyuan Wang, Kuang-chih Lee, Zenglin Xu, Yuan Qi, Zoubin Gharamani Dept. Computer Science, Purdue University, NEC Laboratories

More information

Bayesian nonparametric models of sparse and exchangeable random graphs

Bayesian nonparametric models of sparse and exchangeable random graphs Bayesian nonparametric models of sparse and exchangeable random graphs F. Caron & E. Fox Technical Report Discussion led by Esther Salazar Duke University May 16, 2014 (Reading group) May 16, 2014 1 /

More information

STA 4273H: Statistical Machine Learning

STA 4273H: Statistical Machine Learning STA 4273H: Statistical Machine Learning Russ Salakhutdinov Department of Computer Science! Department of Statistical Sciences! rsalakhu@cs.toronto.edu! h0p://www.cs.utoronto.ca/~rsalakhu/ Lecture 7 Approximate

More information

Distributed Estimation, Information Loss and Exponential Families. Qiang Liu Department of Computer Science Dartmouth College

Distributed Estimation, Information Loss and Exponential Families. Qiang Liu Department of Computer Science Dartmouth College Distributed Estimation, Information Loss and Exponential Families Qiang Liu Department of Computer Science Dartmouth College Statistical Learning / Estimation Learning generative models from data Topic

More information

Tutorial on Gaussian Processes and the Gaussian Process Latent Variable Model

Tutorial on Gaussian Processes and the Gaussian Process Latent Variable Model Tutorial on Gaussian Processes and the Gaussian Process Latent Variable Model (& discussion on the GPLVM tech. report by Prof. N. Lawrence, 06) Andreas Damianou Department of Neuro- and Computer Science,

More information

STA414/2104. Lecture 11: Gaussian Processes. Department of Statistics

STA414/2104. Lecture 11: Gaussian Processes. Department of Statistics STA414/2104 Lecture 11: Gaussian Processes Department of Statistics www.utstat.utoronto.ca Delivered by Mark Ebden with thanks to Russ Salakhutdinov Outline Gaussian Processes Exam review Course evaluations

More information

Quilting Stochastic Kronecker Graphs to Generate Multiplicative Attribute Graphs

Quilting Stochastic Kronecker Graphs to Generate Multiplicative Attribute Graphs Quilting Stochastic Kronecker Graphs to Generate Multiplicative Attribute Graphs Hyokun Yun (work with S.V.N. Vishwanathan) Department of Statistics Purdue Machine Learning Seminar November 9, 2011 Overview

More information

CSC2515 Assignment #2

CSC2515 Assignment #2 CSC2515 Assignment #2 Due: Nov.4, 2pm at the START of class Worth: 18% Late assignments not accepted. 1 Pseudo-Bayesian Linear Regression (3%) In this question you will dabble in Bayesian statistics and

More information

Cheng Soon Ong & Christian Walder. Canberra February June 2018

Cheng Soon Ong & Christian Walder. Canberra February June 2018 Cheng Soon Ong & Christian Walder Research Group and College of Engineering and Computer Science Canberra February June 218 Outlines Overview Introduction Linear Algebra Probability Linear Regression 1

More information

Large-scale Collaborative Prediction Using a Nonparametric Random Effects Model

Large-scale Collaborative Prediction Using a Nonparametric Random Effects Model Large-scale Collaborative Prediction Using a Nonparametric Random Effects Model Kai Yu Joint work with John Lafferty and Shenghuo Zhu NEC Laboratories America, Carnegie Mellon University First Prev Page

More information

Gaussian Processes. Le Song. Machine Learning II: Advanced Topics CSE 8803ML, Spring 2012

Gaussian Processes. Le Song. Machine Learning II: Advanced Topics CSE 8803ML, Spring 2012 Gaussian Processes Le Song Machine Learning II: Advanced Topics CSE 8803ML, Spring 01 Pictorial view of embedding distribution Transform the entire distribution to expected features Feature space Feature

More information

Learning latent structure in complex networks

Learning latent structure in complex networks Learning latent structure in complex networks Lars Kai Hansen www.imm.dtu.dk/~lkh Current network research issues: Social Media Neuroinformatics Machine learning Joint work with Morten Mørup, Sune Lehmann

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 3 Linear

More information

Probabilistic Graphical Models

Probabilistic Graphical Models Probabilistic Graphical Models Brown University CSCI 295-P, Spring 213 Prof. Erik Sudderth Lecture 11: Inference & Learning Overview, Gaussian Graphical Models Some figures courtesy Michael Jordan s draft

More information

ECE521 week 3: 23/26 January 2017

ECE521 week 3: 23/26 January 2017 ECE521 week 3: 23/26 January 2017 Outline Probabilistic interpretation of linear regression - Maximum likelihood estimation (MLE) - Maximum a posteriori (MAP) estimation Bias-variance trade-off Linear

More information

Lecture: Gaussian Process Regression. STAT 6474 Instructor: Hongxiao Zhu

Lecture: Gaussian Process Regression. STAT 6474 Instructor: Hongxiao Zhu Lecture: Gaussian Process Regression STAT 6474 Instructor: Hongxiao Zhu Motivation Reference: Marc Deisenroth s tutorial on Robot Learning. 2 Fast Learning for Autonomous Robots with Gaussian Processes

More information

Deep Poisson Factorization Machines: a factor analysis model for mapping behaviors in journalist ecosystem

Deep Poisson Factorization Machines: a factor analysis model for mapping behaviors in journalist ecosystem 000 001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050

More information

Nonparametric Bayesian Methods (Gaussian Processes)

Nonparametric Bayesian Methods (Gaussian Processes) [70240413 Statistical Machine Learning, Spring, 2015] Nonparametric Bayesian Methods (Gaussian Processes) Jun Zhu dcszj@mail.tsinghua.edu.cn http://bigml.cs.tsinghua.edu.cn/~jun State Key Lab of Intelligent

More information

GraphRNN: A Deep Generative Model for Graphs (24 Feb 2018)

GraphRNN: A Deep Generative Model for Graphs (24 Feb 2018) GraphRNN: A Deep Generative Model for Graphs (24 Feb 2018) Jiaxuan You, Rex Ying, Xiang Ren, William L. Hamilton, Jure Leskovec Presented by: Jesse Bettencourt and Harris Chan March 9, 2018 University

More information

Undirected Graphical Models

Undirected Graphical Models Outline Hong Chang Institute of Computing Technology, Chinese Academy of Sciences Machine Learning Methods (Fall 2012) Outline Outline I 1 Introduction 2 Properties Properties 3 Generative vs. Conditional

More information

Bayesian Methods for Machine Learning

Bayesian Methods for Machine Learning Bayesian Methods for Machine Learning CS 584: Big Data Analytics Material adapted from Radford Neal s tutorial (http://ftp.cs.utoronto.ca/pub/radford/bayes-tut.pdf), Zoubin Ghahramni (http://hunch.net/~coms-4771/zoubin_ghahramani_bayesian_learning.pdf),

More information

PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 2: PROBABILITY DISTRIBUTIONS

PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 2: PROBABILITY DISTRIBUTIONS PATTERN RECOGNITION AND MACHINE LEARNING CHAPTER 2: PROBABILITY DISTRIBUTIONS Parametric Distributions Basic building blocks: Need to determine given Representation: or? Recall Curve Fitting Binary Variables

More information

Pattern Recognition and Machine Learning

Pattern Recognition and Machine Learning Christopher M. Bishop Pattern Recognition and Machine Learning ÖSpri inger Contents Preface Mathematical notation Contents vii xi xiii 1 Introduction 1 1.1 Example: Polynomial Curve Fitting 4 1.2 Probability

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 11 Project

More information

Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, Spis treści

Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, Spis treści Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, 2017 Spis treści Website Acknowledgments Notation xiii xv xix 1 Introduction 1 1.1 Who Should Read This Book?

More information

CSci 8980: Advanced Topics in Graphical Models Gaussian Processes

CSci 8980: Advanced Topics in Graphical Models Gaussian Processes CSci 8980: Advanced Topics in Graphical Models Gaussian Processes Instructor: Arindam Banerjee November 15, 2007 Gaussian Processes Outline Gaussian Processes Outline Parametric Bayesian Regression Gaussian

More information

Computer Vision Group Prof. Daniel Cremers. 9. Gaussian Processes - Regression

Computer Vision Group Prof. Daniel Cremers. 9. Gaussian Processes - Regression Group Prof. Daniel Cremers 9. Gaussian Processes - Regression Repetition: Regularized Regression Before, we solved for w using the pseudoinverse. But: we can kernelize this problem as well! First step:

More information

Decision-making, inference, and learning theory. ECE 830 & CS 761, Spring 2016

Decision-making, inference, and learning theory. ECE 830 & CS 761, Spring 2016 Decision-making, inference, and learning theory ECE 830 & CS 761, Spring 2016 1 / 22 What do we have here? Given measurements or observations of some physical process, we ask the simple question what do

More information

Bayesian Networks: Construction, Inference, Learning and Causal Interpretation. Volker Tresp Summer 2016

Bayesian Networks: Construction, Inference, Learning and Causal Interpretation. Volker Tresp Summer 2016 Bayesian Networks: Construction, Inference, Learning and Causal Interpretation Volker Tresp Summer 2016 1 Introduction So far we were mostly concerned with supervised learning: we predicted one or several

More information

CS839: Probabilistic Graphical Models. Lecture 7: Learning Fully Observed BNs. Theo Rekatsinas

CS839: Probabilistic Graphical Models. Lecture 7: Learning Fully Observed BNs. Theo Rekatsinas CS839: Probabilistic Graphical Models Lecture 7: Learning Fully Observed BNs Theo Rekatsinas 1 Exponential family: a basic building block For a numeric random variable X p(x ) =h(x)exp T T (x) A( ) = 1

More information

Learning to Learn and Collaborative Filtering

Learning to Learn and Collaborative Filtering Appearing in NIPS 2005 workshop Inductive Transfer: Canada, December, 2005. 10 Years Later, Whistler, Learning to Learn and Collaborative Filtering Kai Yu, Volker Tresp Siemens AG, 81739 Munich, Germany

More information

Bayesian Networks: Construction, Inference, Learning and Causal Interpretation. Volker Tresp Summer 2014

Bayesian Networks: Construction, Inference, Learning and Causal Interpretation. Volker Tresp Summer 2014 Bayesian Networks: Construction, Inference, Learning and Causal Interpretation Volker Tresp Summer 2014 1 Introduction So far we were mostly concerned with supervised learning: we predicted one or several

More information

Nonparameteric Regression:

Nonparameteric Regression: Nonparameteric Regression: Nadaraya-Watson Kernel Regression & Gaussian Process Regression Seungjin Choi Department of Computer Science and Engineering Pohang University of Science and Technology 77 Cheongam-ro,

More information

DeepCP: Flexible Nonlinear Tensor Decomposition

DeepCP: Flexible Nonlinear Tensor Decomposition DeepCP: Flexible Nonlinear Tensor Decomposition Bin Liu, Lirong He SMILELab, Sch. of Comp. Sci.& Eng., University of Electronic Science and Technology of China {liu, lirong_he}@std.uestc.edu.cn, Shandian

More information

Data Mining Techniques

Data Mining Techniques Data Mining Techniques CS 6220 - Section 3 - Fall 2016 Lecture 12 Jan-Willem van de Meent (credit: Yijun Zhao, Percy Liang) DIMENSIONALITY REDUCTION Borrowing from: Percy Liang (Stanford) Linear Dimensionality

More information

Statistical Inference on Large Contingency Tables: Convergence, Testability, Stability. COMPSTAT 2010 Paris, August 23, 2010

Statistical Inference on Large Contingency Tables: Convergence, Testability, Stability. COMPSTAT 2010 Paris, August 23, 2010 Statistical Inference on Large Contingency Tables: Convergence, Testability, Stability Marianna Bolla Institute of Mathematics Budapest University of Technology and Economics marib@math.bme.hu COMPSTAT

More information

COMP 551 Applied Machine Learning Lecture 20: Gaussian processes

COMP 551 Applied Machine Learning Lecture 20: Gaussian processes COMP 55 Applied Machine Learning Lecture 2: Gaussian processes Instructor: Ryan Lowe (ryan.lowe@cs.mcgill.ca) Slides mostly by: (herke.vanhoof@mcgill.ca) Class web page: www.cs.mcgill.ca/~hvanho2/comp55

More information

Introduction to Machine Learning CMU-10701

Introduction to Machine Learning CMU-10701 Introduction to Machine Learning CMU-10701 Markov Chain Monte Carlo Methods Barnabás Póczos & Aarti Singh Contents Markov Chain Monte Carlo Methods Goal & Motivation Sampling Rejection Importance Markov

More information

Variational Model Selection for Sparse Gaussian Process Regression

Variational Model Selection for Sparse Gaussian Process Regression Variational Model Selection for Sparse Gaussian Process Regression Michalis K. Titsias School of Computer Science University of Manchester 7 September 2008 Outline Gaussian process regression and sparse

More information

Computer Vision Group Prof. Daniel Cremers. 4. Gaussian Processes - Regression

Computer Vision Group Prof. Daniel Cremers. 4. Gaussian Processes - Regression Group Prof. Daniel Cremers 4. Gaussian Processes - Regression Definition (Rep.) Definition: A Gaussian process is a collection of random variables, any finite number of which have a joint Gaussian distribution.

More information

Probability and Information Theory. Sargur N. Srihari

Probability and Information Theory. Sargur N. Srihari Probability and Information Theory Sargur N. srihari@cedar.buffalo.edu 1 Topics in Probability and Information Theory Overview 1. Why Probability? 2. Random Variables 3. Probability Distributions 4. Marginal

More information

Deep Learning Srihari. Deep Belief Nets. Sargur N. Srihari

Deep Learning Srihari. Deep Belief Nets. Sargur N. Srihari Deep Belief Nets Sargur N. Srihari srihari@cedar.buffalo.edu Topics 1. Boltzmann machines 2. Restricted Boltzmann machines 3. Deep Belief Networks 4. Deep Boltzmann machines 5. Boltzmann machines for continuous

More information

CPSC 540: Machine Learning

CPSC 540: Machine Learning CPSC 540: Machine Learning Undirected Graphical Models Mark Schmidt University of British Columbia Winter 2016 Admin Assignment 3: 2 late days to hand it in today, Thursday is final day. Assignment 4:

More information

Fully Bayesian Deep Gaussian Processes for Uncertainty Quantification

Fully Bayesian Deep Gaussian Processes for Uncertainty Quantification Fully Bayesian Deep Gaussian Processes for Uncertainty Quantification N. Zabaras 1 S. Atkinson 1 Center for Informatics and Computational Science Department of Aerospace and Mechanical Engineering University

More information

CPSC 540: Machine Learning

CPSC 540: Machine Learning CPSC 540: Machine Learning MCMC and Non-Parametric Bayes Mark Schmidt University of British Columbia Winter 2016 Admin I went through project proposals: Some of you got a message on Piazza. No news is

More information

PROBABILITY DISTRIBUTIONS. J. Elder CSE 6390/PSYC 6225 Computational Modeling of Visual Perception

PROBABILITY DISTRIBUTIONS. J. Elder CSE 6390/PSYC 6225 Computational Modeling of Visual Perception PROBABILITY DISTRIBUTIONS Credits 2 These slides were sourced and/or modified from: Christopher Bishop, Microsoft UK Parametric Distributions 3 Basic building blocks: Need to determine given Representation:

More information

Exchangeability. Peter Orbanz. Columbia University

Exchangeability. Peter Orbanz. Columbia University Exchangeability Peter Orbanz Columbia University PARAMETERS AND PATTERNS Parameters P(X θ) = Probability[data pattern] 3 2 1 0 1 2 3 5 0 5 Inference idea data = underlying pattern + independent noise Peter

More information

Online Algorithms for Sum-Product

Online Algorithms for Sum-Product Online Algorithms for Sum-Product Networks with Continuous Variables Priyank Jaini Ph.D. Seminar Consistent/Robust Tensor Decomposition And Spectral Learning Offline Bayesian Learning ADF, EP, SGD, oem

More information

Machine Learning (CS 567) Lecture 5

Machine Learning (CS 567) Lecture 5 Machine Learning (CS 567) Lecture 5 Time: T-Th 5:00pm - 6:20pm Location: GFS 118 Instructor: Sofus A. Macskassy (macskass@usc.edu) Office: SAL 216 Office hours: by appointment Teaching assistant: Cheol

More information

Clustering using Mixture Models

Clustering using Mixture Models Clustering using Mixture Models The full posterior of the Gaussian Mixture Model is p(x, Z, µ,, ) =p(x Z, µ, )p(z )p( )p(µ, ) data likelihood (Gaussian) correspondence prob. (Multinomial) mixture prior

More information

Machine Learning. Gaussian Mixture Models. Zhiyao Duan & Bryan Pardo, Machine Learning: EECS 349 Fall

Machine Learning. Gaussian Mixture Models. Zhiyao Duan & Bryan Pardo, Machine Learning: EECS 349 Fall Machine Learning Gaussian Mixture Models Zhiyao Duan & Bryan Pardo, Machine Learning: EECS 349 Fall 2012 1 The Generative Model POV We think of the data as being generated from some process. We assume

More information

Multivariate Normal Models

Multivariate Normal Models Case Study 3: fmri Prediction Coping with Large Covariances: Latent Factor Models, Graphical Models, Graphical LASSO Machine Learning for Big Data CSE547/STAT548, University of Washington Emily Fox February

More information

Organization. I MCMC discussion. I project talks. I Lecture.

Organization. I MCMC discussion. I project talks. I Lecture. Organization I MCMC discussion I project talks. I Lecture. Content I Uncertainty Propagation Overview I Forward-Backward with an Ensemble I Model Reduction (Intro) Uncertainty Propagation in Causal Systems

More information

STA 414/2104: Machine Learning

STA 414/2104: Machine Learning STA 414/2104: Machine Learning Russ Salakhutdinov Department of Computer Science! Department of Statistics! rsalakhu@cs.toronto.edu! http://www.cs.toronto.edu/~rsalakhu/ Lecture 9 Sequential Data So far

More information

Large-scale Ordinal Collaborative Filtering

Large-scale Ordinal Collaborative Filtering Large-scale Ordinal Collaborative Filtering Ulrich Paquet, Blaise Thomson, and Ole Winther Microsoft Research Cambridge, University of Cambridge, Technical University of Denmark ulripa@microsoft.com,brmt2@cam.ac.uk,owi@imm.dtu.dk

More information

April 20th, Advanced Topics in Machine Learning California Institute of Technology. Markov Chain Monte Carlo for Machine Learning

April 20th, Advanced Topics in Machine Learning California Institute of Technology. Markov Chain Monte Carlo for Machine Learning for for Advanced Topics in California Institute of Technology April 20th, 2017 1 / 50 Table of Contents for 1 2 3 4 2 / 50 History of methods for Enrico Fermi used to calculate incredibly accurate predictions

More information

Big & Quic: Sparse Inverse Covariance Estimation for a Million Variables

Big & Quic: Sparse Inverse Covariance Estimation for a Million Variables for a Million Variables Cho-Jui Hsieh The University of Texas at Austin NIPS Lake Tahoe, Nevada Dec 8, 2013 Joint work with M. Sustik, I. Dhillon, P. Ravikumar and R. Poldrack FMRI Brain Analysis Goal:

More information

Scaling Neighbourhood Methods

Scaling Neighbourhood Methods Quick Recap Scaling Neighbourhood Methods Collaborative Filtering m = #items n = #users Complexity : m * m * n Comparative Scale of Signals ~50 M users ~25 M items Explicit Ratings ~ O(1M) (1 per billion)

More information

Introduction to the Tensor Train Decomposition and Its Applications in Machine Learning

Introduction to the Tensor Train Decomposition and Its Applications in Machine Learning Introduction to the Tensor Train Decomposition and Its Applications in Machine Learning Anton Rodomanov Higher School of Economics, Russia Bayesian methods research group (http://bayesgroup.ru) 14 March

More information

CS242: Probabilistic Graphical Models Lecture 4A: MAP Estimation & Graph Structure Learning

CS242: Probabilistic Graphical Models Lecture 4A: MAP Estimation & Graph Structure Learning CS242: Probabilistic Graphical Models Lecture 4A: MAP Estimation & Graph Structure Learning Professor Erik Sudderth Brown University Computer Science October 4, 2016 Some figures and materials courtesy

More information

Sum-Product Networks: A New Deep Architecture

Sum-Product Networks: A New Deep Architecture Sum-Product Networks: A New Deep Architecture Pedro Domingos Dept. Computer Science & Eng. University of Washington Joint work with Hoifung Poon 1 Graphical Models: Challenges Bayesian Network Markov Network

More information

A graph contains a set of nodes (vertices) connected by links (edges or arcs)

A graph contains a set of nodes (vertices) connected by links (edges or arcs) BOLTZMANN MACHINES Generative Models Graphical Models A graph contains a set of nodes (vertices) connected by links (edges or arcs) In a probabilistic graphical model, each node represents a random variable,

More information

Computer Vision Group Prof. Daniel Cremers. 10a. Markov Chain Monte Carlo

Computer Vision Group Prof. Daniel Cremers. 10a. Markov Chain Monte Carlo Group Prof. Daniel Cremers 10a. Markov Chain Monte Carlo Markov Chain Monte Carlo In high-dimensional spaces, rejection sampling and importance sampling are very inefficient An alternative is Markov Chain

More information

Mathematical foundations - linear algebra

Mathematical foundations - linear algebra Mathematical foundations - linear algebra Andrea Passerini passerini@disi.unitn.it Machine Learning Vector space Definition (over reals) A set X is called a vector space over IR if addition and scalar

More information

[POLS 8500] Review of Linear Algebra, Probability and Information Theory

[POLS 8500] Review of Linear Algebra, Probability and Information Theory [POLS 8500] Review of Linear Algebra, Probability and Information Theory Professor Jason Anastasopoulos ljanastas@uga.edu January 12, 2017 For today... Basic linear algebra. Basic probability. Programming

More information

Mixed Membership Stochastic Blockmodels

Mixed Membership Stochastic Blockmodels Mixed Membership Stochastic Blockmodels (2008) Edoardo M. Airoldi, David M. Blei, Stephen E. Fienberg and Eric P. Xing Herrissa Lamothe Princeton University Herrissa Lamothe (Princeton University) Mixed

More information

Linear & nonlinear classifiers

Linear & nonlinear classifiers Linear & nonlinear classifiers Machine Learning Hamid Beigy Sharif University of Technology Fall 1394 Hamid Beigy (Sharif University of Technology) Linear & nonlinear classifiers Fall 1394 1 / 34 Table

More information

Gaussian processes for inference in stochastic differential equations

Gaussian processes for inference in stochastic differential equations Gaussian processes for inference in stochastic differential equations Manfred Opper, AI group, TU Berlin November 6, 2017 Manfred Opper, AI group, TU Berlin (TU Berlin) inference in SDE November 6, 2017

More information

Stochastic Variational Inference for Gaussian Process Latent Variable Models using Back Constraints

Stochastic Variational Inference for Gaussian Process Latent Variable Models using Back Constraints Stochastic Variational Inference for Gaussian Process Latent Variable Models using Back Constraints Thang D. Bui Richard E. Turner tdb40@cam.ac.uk ret26@cam.ac.uk Computational and Biological Learning

More information

Introduction to Machine Learning

Introduction to Machine Learning Introduction to Machine Learning 12. Gaussian Processes Alex Smola Carnegie Mellon University http://alex.smola.org/teaching/cmu2013-10-701 10-701 The Normal Distribution http://www.gaussianprocess.org/gpml/chapters/

More information

Today. Probability and Statistics. Linear Algebra. Calculus. Naïve Bayes Classification. Matrix Multiplication Matrix Inversion

Today. Probability and Statistics. Linear Algebra. Calculus. Naïve Bayes Classification. Matrix Multiplication Matrix Inversion Today Probability and Statistics Naïve Bayes Classification Linear Algebra Matrix Multiplication Matrix Inversion Calculus Vector Calculus Optimization Lagrange Multipliers 1 Classical Artificial Intelligence

More information

Gaussian Process Vine Copulas for Multivariate Dependence

Gaussian Process Vine Copulas for Multivariate Dependence Gaussian Process Vine Copulas for Multivariate Dependence José Miguel Hernández-Lobato 1,2 joint work with David López-Paz 2,3 and Zoubin Ghahramani 1 1 Department of Engineering, Cambridge University,

More information

Modeling Data with Linear Combinations of Basis Functions. Read Chapter 3 in the text by Bishop

Modeling Data with Linear Combinations of Basis Functions. Read Chapter 3 in the text by Bishop Modeling Data with Linear Combinations of Basis Functions Read Chapter 3 in the text by Bishop A Type of Supervised Learning Problem We want to model data (x 1, t 1 ),..., (x N, t N ), where x i is a vector

More information

Facebook Friends! and Matrix Functions

Facebook Friends! and Matrix Functions Facebook Friends! and Matrix Functions! Graduate Research Day Joint with David F. Gleich, (Purdue), supported by" NSF CAREER 1149756-CCF Kyle Kloster! Purdue University! Network Analysis Use linear algebra

More information

Midterm exam CS 189/289, Fall 2015

Midterm exam CS 189/289, Fall 2015 Midterm exam CS 189/289, Fall 2015 You have 80 minutes for the exam. Total 100 points: 1. True/False: 36 points (18 questions, 2 points each). 2. Multiple-choice questions: 24 points (8 questions, 3 points

More information

Intelligent Systems Statistical Machine Learning

Intelligent Systems Statistical Machine Learning Intelligent Systems Statistical Machine Learning Carsten Rother, Dmitrij Schlesinger WS2014/2015, Our tasks (recap) The model: two variables are usually present: - the first one is typically discrete k

More information

Introduction to Machine Learning Midterm Exam

Introduction to Machine Learning Midterm Exam 10-701 Introduction to Machine Learning Midterm Exam Instructors: Eric Xing, Ziv Bar-Joseph 17 November, 2015 There are 11 questions, for a total of 100 points. This exam is open book, open notes, but

More information

Naïve Bayes classification

Naïve Bayes classification Naïve Bayes classification 1 Probability theory Random variable: a variable whose possible values are numerical outcomes of a random phenomenon. Examples: A person s height, the outcome of a coin toss

More information

Learning Bayesian network : Given structure and completely observed data

Learning Bayesian network : Given structure and completely observed data Learning Bayesian network : Given structure and completely observed data Probabilistic Graphical Models Sharif University of Technology Spring 2017 Soleymani Learning problem Target: true distribution

More information

Gaussian Processes for Big Data. James Hensman

Gaussian Processes for Big Data. James Hensman Gaussian Processes for Big Data James Hensman Overview Motivation Sparse Gaussian Processes Stochastic Variational Inference Examples Overview Motivation Sparse Gaussian Processes Stochastic Variational

More information

Computer Vision Group Prof. Daniel Cremers. 2. Regression (cont.)

Computer Vision Group Prof. Daniel Cremers. 2. Regression (cont.) Prof. Daniel Cremers 2. Regression (cont.) Regression with MLE (Rep.) Assume that y is affected by Gaussian noise : t = f(x, w)+ where Thus, we have p(t x, w, )=N (t; f(x, w), 2 ) 2 Maximum A-Posteriori

More information

The Distribution of Partially Exchangeable Random Variables

The Distribution of Partially Exchangeable Random Variables The Distribution of Partially Exchangeable Random Variables Hanxiang Peng a,1,, Xin Dang b,1, Xueqin Wang c,2 a Department of Mathematical Sciences, Indiana University-Purdue University at Indianapolis,

More information

A brief introduction to Conditional Random Fields

A brief introduction to Conditional Random Fields A brief introduction to Conditional Random Fields Mark Johnson Macquarie University April, 2005, updated October 2010 1 Talk outline Graphical models Maximum likelihood and maximum conditional likelihood

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 7 Approximate

More information

Lecture 21: Spectral Learning for Graphical Models

Lecture 21: Spectral Learning for Graphical Models 10-708: Probabilistic Graphical Models 10-708, Spring 2016 Lecture 21: Spectral Learning for Graphical Models Lecturer: Eric P. Xing Scribes: Maruan Al-Shedivat, Wei-Cheng Chang, Frederick Liu 1 Motivation

More information

Linear Models for Regression CS534

Linear Models for Regression CS534 Linear Models for Regression CS534 Example Regression Problems Predict housing price based on House size, lot size, Location, # of rooms Predict stock price based on Price history of the past month Predict

More information

p L yi z n m x N n xi

p L yi z n m x N n xi y i z n x n N x i Overview Directed and undirected graphs Conditional independence Exact inference Latent variables and EM Variational inference Books statistical perspective Graphical Models, S. Lauritzen

More information

Mathematical Formulation of Our Example

Mathematical Formulation of Our Example Mathematical Formulation of Our Example We define two binary random variables: open and, where is light on or light off. Our question is: What is? Computer Vision 1 Combining Evidence Suppose our robot

More information

Overfitting, Bias / Variance Analysis

Overfitting, Bias / Variance Analysis Overfitting, Bias / Variance Analysis Professor Ameet Talwalkar Professor Ameet Talwalkar CS260 Machine Learning Algorithms February 8, 207 / 40 Outline Administration 2 Review of last lecture 3 Basic

More information

Gaussian Process Regression

Gaussian Process Regression Gaussian Process Regression 4F1 Pattern Recognition, 21 Carl Edward Rasmussen Department of Engineering, University of Cambridge November 11th - 16th, 21 Rasmussen (Engineering, Cambridge) Gaussian Process

More information

Probabilistic numerics for deep learning

Probabilistic numerics for deep learning Presenter: Shijia Wang Department of Engineering Science, University of Oxford rning (RLSS) Summer School, Montreal 2017 Outline 1 Introduction Probabilistic Numerics 2 Components Probabilistic modeling

More information

Conquering the Complexity of Time: Machine Learning for Big Time Series Data

Conquering the Complexity of Time: Machine Learning for Big Time Series Data Conquering the Complexity of Time: Machine Learning for Big Time Series Data Yan Liu Computer Science Department University of Southern California Mini-Workshop on Theoretical Foundations of Cyber-Physical

More information

Linear Dynamical Systems

Linear Dynamical Systems Linear Dynamical Systems Sargur N. srihari@cedar.buffalo.edu Machine Learning Course: http://www.cedar.buffalo.edu/~srihari/cse574/index.html Two Models Described by Same Graph Latent variables Observations

More information

Unsupervised Machine Learning and Data Mining. DS 5230 / DS Fall Lecture 7. Jan-Willem van de Meent

Unsupervised Machine Learning and Data Mining. DS 5230 / DS Fall Lecture 7. Jan-Willem van de Meent Unsupervised Machine Learning and Data Mining DS 5230 / DS 4420 - Fall 2018 Lecture 7 Jan-Willem van de Meent DIMENSIONALITY REDUCTION Borrowing from: Percy Liang (Stanford) Dimensionality Reduction Goal:

More information

Dynamic Data Modeling, Recognition, and Synthesis. Rui Zhao Thesis Defense Advisor: Professor Qiang Ji

Dynamic Data Modeling, Recognition, and Synthesis. Rui Zhao Thesis Defense Advisor: Professor Qiang Ji Dynamic Data Modeling, Recognition, and Synthesis Rui Zhao Thesis Defense Advisor: Professor Qiang Ji Contents Introduction Related Work Dynamic Data Modeling & Analysis Temporal localization Insufficient

More information

Data Mining Techniques

Data Mining Techniques Data Mining Techniques CS 622 - Section 2 - Spring 27 Pre-final Review Jan-Willem van de Meent Feedback Feedback https://goo.gl/er7eo8 (also posted on Piazza) Also, please fill out your TRACE evaluations!

More information