Unsupervised Kernel Dimension Reduction Supplemental Material

Size: px
Start display at page:

Download "Unsupervised Kernel Dimension Reduction Supplemental Material"

Transcription

1 Unsupervised Kernel Dimension Reduction Supplemental Material Meihong Wang Dept. of Computer Science U. of Southern California Los Angeles, CA Fei Sha Dept. of Computer Science U. of Southern California Los Angeles, CA Michael I. Jordan Dept. of Statistics U. of California Berkeley, CA Relationship between ĴYY X and ĴXY and Proof of Proposition We start by noting that conditional independence X Y B X does not necessarily imply the correlation between B X and Y is maximized. To see this, let X be a Gaussian random vairable with zero mean and diagonal covariance matrix. Assume B is an identity matrix and Y = X = (B X) (elementwise square for a vectorial X). The conditional independence is obviously satisfied yet the correlation between B X and Y is zero. This observation is yet another example showing the limitation of Spearson s correlation measures, which detect only linear dependence between random variables. In the following, we show that when measured in the RKHS, the two measures ĴY Y X and ĴXY are equivalent. Assume we use Gaussian RBF kernel for both ĴYY X(B X,Y) and ĴXY(B X,Y): K(x i,x ) = exp ( x i x /σ N) where σn is the bandwidth. It should be chosen to match roughly the scale of the data. While ĴXY(B X,Y) depends only on σ N, Ĵ YY X (B X,Y) depends on σ N and ǫ N, which need to be adusted in empirical studies. Since both parameters control the smoothness of the kernel matrix K B X, it is sensible to adust the relative scale of the two parameters. Particularly, the following statement establishes the link between the two measures: Proposition. Let N + and ǫ N. Additionally, assume the samples are distributed uniformly on the unit sphere. If σ N ǫ N, then up to a constant, Ĵ YY X (B X,Y) c N ǫ NĴXY(B X,Y) () Therefore, It is equivalent to minimizing ĴYY X(B X,Y) and maximizing ĴXY(B X,Y). We sketch the proof in the following. The proof consists of two main steps. In the first step, we bound the largest eigenvalue λ of K B X from above. In the second step, we show that applying the bound allows us to approximate the inverse (G B X +Nǫ N I N ) with K B X. Bound on λ. With Gaussian RBF kernel, the elements of the kernel matrix K B X is strictly positive. By Perron-Frobenius theorem, the spectral radius is λ and is bound by λ max K B i X(i,), () namely, the maximum of the row sums. The (i,)-th element of the kernel matrix is given by ( ) K B X(i,) = exp m B m x i B m x /σ N ()

2 whereb m is them-th columnofthe matrixb. Since eachcolumncontributesnonnegatively to the exponent, it is obvious λ max exp ( B m x i B m x /σ ) N () i for any m. Furthermore, since x i are assumed to be uniformly distributed on the unit sphere, we choose an arbitrary coordinate system such that B m = [ ]. This gives rise to λ exp ( (x i x ) /σn). () Note that we have removed the max i operation based on the symmetry argument. The first element x i of x i can be set as without loss of generality. Moreover, the first elements x of x can be parameterized as cosθ where θ is uniformly distributed between and π. This leads to λ exp ( ( cosθ ) /σn). () When N +, the right-hand-size tends to an integral λ N π e ( cos θ) σ N dθ. () The integral does not have a analytic closed-form. Our intention is to approximate it with a function of σ N. In particular, σ N needs to tend to zero when N +. Therefore, our next step is to identify an asymptotic expansion in σ N. After a few variable substitutions ( cosθ = cos θ/, then cosθ/ = t, the integral of eq. () is transformed (omitting constants) to I(σ N ) = t e t /σ N dt () ApplyingWatsonLemma[]tothe aboveintegralas/σn +, weobtainthe asymptotic expansion up to (and including) the first-order of / t, I(σ N ) / Γ(/) σ N /+O(σ / N ) (σ N ). () This gives us a bound on λ λ c N σ N () where c is a constant. Approximate ĴY Y X(B X,Y). We apply the Woodbury matrix inversion lemma to the independence measure. Let δ N = Nǫ N, we have Ĵ Y Y X (B X,Y) () = Trace [ G Y (HK B XH +δ N I N ) ] () = Trace [ { G Y δ N I N δ N H(K B X +Hδ N H) Hδ }] N () = δ N Trace[ G Y (K B X +δ N H) ] +const () where we have used the identities HH = H and HG Y H = G Y. We have also assumed that K B X is invertible. This is reasonable as we are using Gaussian RBF kernels. If we further assume that B X yields different B x, B x,... and B x N, then the kernel matrix is full ranked and invertible. We consider the limiting case when N +. We have shown in above that the largest eigenvalueofk B X isboundfromabovebyn σ N. Thismeansthatthesmallesteigenvalue /λ ofk B X isbound frombelowby/(n σ N ). Notethatthe largesteigenvalueofδ N H is /(Nǫ N ). Therefore, if /(N σ N ) /(Nǫ N ), ()

3 we can approximate the inversion with (K B X +δ N H) (K B X ) = K B X. () The condition eq. () corresponds to the condition σ N ǫ N in the Proposition, which leads to (after substituting eq. () into eq.()), Ĵ Y Y X (B X,Y) = δ NĴXY(B X,Y)+const () Relation between t-sne and UKDR t-sne aims to preserve local conditional probability structure computed in the original space of X []. The structure is encoded as the random walk probability of p i = P(x i x ) = exp{ x i x /σ i } i exp{ x i x /σ i }. () The low dimensional coordinates Z are estimated such that the structure computed in the low dimensional space q i C(z i,z ) = + z i z () has the smallest KL divergence: z = argminkl( P Q) where Q s elements are q i. P = (P + P )/ is a symmetric version of P. Minimizing the KL divergence is equivalent to maximize the conditional entropy P i logq i = P i logc(z i,z ) log C(z i,z ) = Trace[P logc] log C () i i i where the logarithm of the matrix C is taken element-wisely. There is a strong analogy of this obective function to ĴXY(B X,X) with random walk kernel over X and Cauchy kernel over B X, Ĵ XY (B X,X) = Trace[PP HCH] Trace[PP C] /N C () where N is the number of data points. The approximation is valid if we assume data are well-clustered and the distances between data in the same clusters are roughly the same. We gain further insights about t-sne by comparing the two obective functions of eq. () and eq. (). Apart from using different yet related kernels (tsne s kernel is not positive semidefinite), both obective functions have two terms which function similarly. The trace terms try to match the similarities between low dimensional coordinates (encoded by C and log C respectively) with the similarities in the high dimensional coordinates. The normalization terms, to be minimized, try to push data points far away from each other. Effect of sparsity in RSN UKDR For nonlinear unsupervised kernel dimension reduction, using random and sparse features (RSF) extracted from X has significant computational advantage than using transformed X by radial basis networks. RSF features are sparse and easy to compute. In particular, the dimensionality of RSF features depends on the dimensionality of the data while RBF transformation depends on the number of data points in the training data set. Furthermore, in computing RSF features, the bias constant b can be used to yield sparser feature vectors. In Fig., we investigate the effect of the sparsity level on embeddings. The setup is similar to what is used for generating the embedding in Fig.(i)(in the main paper) where b is and leads to a feature vector with sparsity level of %. We change b to, and, obtaining sparsity levels of %, % and % in Fig.. Through visual inspection, it is clear that a sparsity level of % does not bring detrimental effects, while the higher % starts to show fragmented clustering of data. Therefore, we deem RSF features as a viable option in handling high dimension data for nonlinear UKDR.

4 (a) Sparsity: % (b) Sparsity: % (c) Sparsity: % Figure : D Embedding of USPS- dataset, computed using random sparse features with various degrees of sparsity. The higher the sparsity is, the sparser the features are. Comparison to other dimensionality reduction method We compare our UKDR method to other dimensionality reduction methods, including SNE [], t-sne, MUSHIC (Colored Maximum Variance Unfolding) [] and PCA. The first dataset is USPS, which contains digit,,,, and. The second dataset is UPSPS, which contains images of digits. The third dataset is discussion postings from Newsgroups with posting in each group. It is a subset of the Newsgroups used in [].. USPS Fig. shows the D embedding results on dataset USPS. (a) UKDR (Nonparametric embedding) Figure : D embedding results on USPS by our method UKDR and other methods. UKDR performs much better than other methods. Results on USPS Fig. shows the D embedding results on dataset USPS.. Results on Newsgroup Fig. shows the D embedding results on dataset Newsgroups

5 (a) RBN UKDR Figure : D embedding results on USPS by our method UKDR and other methods. Only UKDR and t-sne separate all classes reasonably well Group Group Group Group Group Group Group Group Group Group Group Group Group Group Group (a) UKDR Group Group Group Group Group Group x Group Group Group Group x Figure : D embedding results on Newsgroupsby our method UKDR and other methods. UKDR, t-sne and MUHSIC works well. References [] E. T. Copson. Asymptotic expansions. Cambridge University Press,. [] L. van der Maaten and G. Hinton. Visualizing data using t-sne. Journal of Machine Learning Research, :,. [] G. Hinton and S. Roweis. Stochastic neighbor embedding. Advances in Neural Information Processing Systems, pages,. [] L. Song, A. Smola, K. Borgwardt, and A. Gretton. Colored maximum variance unfolding. Advances in Neural Information Processing Systems, pages,.

Kernel methods for comparing distributions, measuring dependence

Kernel methods for comparing distributions, measuring dependence Kernel methods for comparing distributions, measuring dependence Le Song Machine Learning II: Advanced Topics CSE 8803ML, Spring 2012 Principal component analysis Given a set of M centered observations

More information

Kernel-Based Contrast Functions for Sufficient Dimension Reduction

Kernel-Based Contrast Functions for Sufficient Dimension Reduction Kernel-Based Contrast Functions for Sufficient Dimension Reduction Michael I. Jordan Departments of Statistics and EECS University of California, Berkeley Joint work with Kenji Fukumizu and Francis Bach

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

Machine Learning (BSMC-GA 4439) Wenke Liu

Machine Learning (BSMC-GA 4439) Wenke Liu Machine Learning (BSMC-GA 4439) Wenke Liu 02-01-2018 Biomedical data are usually high-dimensional Number of samples (n) is relatively small whereas number of features (p) can be large Sometimes p>>n Problems

More information

Unsupervised dimensionality reduction

Unsupervised dimensionality reduction Unsupervised dimensionality reduction Guillaume Obozinski Ecole des Ponts - ParisTech SOCN course 2014 Guillaume Obozinski Unsupervised dimensionality reduction 1/30 Outline 1 PCA 2 Kernel PCA 3 Multidimensional

More information

Linear and Non-Linear Dimensionality Reduction

Linear and Non-Linear Dimensionality Reduction Linear and Non-Linear Dimensionality Reduction Alexander Schulz aschulz(at)techfak.uni-bielefeld.de University of Pisa, Pisa 4.5.215 and 7.5.215 Overview Dimensionality Reduction Motivation Linear Projections

More information

Representational Power of Restricted Boltzmann Machines and Deep Belief Networks. Nicolas Le Roux and Yoshua Bengio Presented by Colin Graber

Representational Power of Restricted Boltzmann Machines and Deep Belief Networks. Nicolas Le Roux and Yoshua Bengio Presented by Colin Graber Representational Power of Restricted Boltzmann Machines and Deep Belief Networks Nicolas Le Roux and Yoshua Bengio Presented by Colin Graber Introduction Representational abilities of functions with some

More information

CSE 291. Assignment Spectral clustering versus k-means. Out: Wed May 23 Due: Wed Jun 13

CSE 291. Assignment Spectral clustering versus k-means. Out: Wed May 23 Due: Wed Jun 13 CSE 291. Assignment 3 Out: Wed May 23 Due: Wed Jun 13 3.1 Spectral clustering versus k-means Download the rings data set for this problem from the course web site. The data is stored in MATLAB format as

More information

Kernel Methods. Machine Learning A W VO

Kernel Methods. Machine Learning A W VO Kernel Methods Machine Learning A 708.063 07W VO Outline 1. Dual representation 2. The kernel concept 3. Properties of kernels 4. Examples of kernel machines Kernel PCA Support vector regression (Relevance

More information

Colored Maximum Variance Unfolding

Colored Maximum Variance Unfolding Colored Maximum Variance Unfolding Le Song, Alex Smola, Karsten Borgwardt and Arthur Gretton National ICT Australia, Canberra, Australia University of Cambridge, Cambridge, United Kingdom MPI for Biological

More information

Connection of Local Linear Embedding, ISOMAP, and Kernel Principal Component Analysis

Connection of Local Linear Embedding, ISOMAP, and Kernel Principal Component Analysis Connection of Local Linear Embedding, ISOMAP, and Kernel Principal Component Analysis Alvina Goh Vision Reading Group 13 October 2005 Connection of Local Linear Embedding, ISOMAP, and Kernel Principal

More information

Discriminative Direction for Kernel Classifiers

Discriminative Direction for Kernel Classifiers Discriminative Direction for Kernel Classifiers Polina Golland Artificial Intelligence Lab Massachusetts Institute of Technology Cambridge, MA 02139 polina@ai.mit.edu Abstract In many scientific and engineering

More information

SPECTRAL CLUSTERING AND KERNEL PRINCIPAL COMPONENT ANALYSIS ARE PURSUING GOOD PROJECTIONS

SPECTRAL CLUSTERING AND KERNEL PRINCIPAL COMPONENT ANALYSIS ARE PURSUING GOOD PROJECTIONS SPECTRAL CLUSTERING AND KERNEL PRINCIPAL COMPONENT ANALYSIS ARE PURSUING GOOD PROJECTIONS VIKAS CHANDRAKANT RAYKAR DECEMBER 5, 24 Abstract. We interpret spectral clustering algorithms in the light of unsupervised

More information

Modelling Non-linear and Non-stationary Time Series

Modelling Non-linear and Non-stationary Time Series Modelling Non-linear and Non-stationary Time Series Chapter 2: Non-parametric methods Henrik Madsen Advanced Time Series Analysis September 206 Henrik Madsen (02427 Adv. TS Analysis) Lecture Notes September

More information

Importance Reweighting Using Adversarial-Collaborative Training

Importance Reweighting Using Adversarial-Collaborative Training Importance Reweighting Using Adversarial-Collaborative Training Yifan Wu yw4@andrew.cmu.edu Tianshu Ren tren@andrew.cmu.edu Lidan Mu lmu@andrew.cmu.edu Abstract We consider the problem of reweighting a

More information

Kernel methods, kernel SVM and ridge regression

Kernel methods, kernel SVM and ridge regression Kernel methods, kernel SVM and ridge regression Le Song Machine Learning II: Advanced Topics CSE 8803ML, Spring 2012 Collaborative Filtering 2 Collaborative Filtering R: rating matrix; U: user factor;

More information

Learning Eigenfunctions: Links with Spectral Clustering and Kernel PCA

Learning Eigenfunctions: Links with Spectral Clustering and Kernel PCA Learning Eigenfunctions: Links with Spectral Clustering and Kernel PCA Yoshua Bengio Pascal Vincent Jean-François Paiement University of Montreal April 2, Snowbird Learning 2003 Learning Modal Structures

More information

Learning from Labeled and Unlabeled Data: Semi-supervised Learning and Ranking p. 1/31

Learning from Labeled and Unlabeled Data: Semi-supervised Learning and Ranking p. 1/31 Learning from Labeled and Unlabeled Data: Semi-supervised Learning and Ranking Dengyong Zhou zhou@tuebingen.mpg.de Dept. Schölkopf, Max Planck Institute for Biological Cybernetics, Germany Learning from

More information

Cluster Kernels for Semi-Supervised Learning

Cluster Kernels for Semi-Supervised Learning Cluster Kernels for Semi-Supervised Learning Olivier Chapelle, Jason Weston, Bernhard Scholkopf Max Planck Institute for Biological Cybernetics, 72076 Tiibingen, Germany {first. last} @tuebingen.mpg.de

More information

Nonlinear Dimensionality Reduction

Nonlinear Dimensionality Reduction Outline Hong Chang Institute of Computing Technology, Chinese Academy of Sciences Machine Learning Methods (Fall 2012) Outline Outline I 1 Kernel PCA 2 Isomap 3 Locally Linear Embedding 4 Laplacian Eigenmap

More information

Linearly-solvable Markov decision problems

Linearly-solvable Markov decision problems Advances in Neural Information Processing Systems 2 Linearly-solvable Markov decision problems Emanuel Todorov Department of Cognitive Science University of California San Diego todorov@cogsci.ucsd.edu

More information

Introduction to Machine Learning

Introduction to Machine Learning Introduction to Machine Learning Kernel Methods Varun Chandola Computer Science & Engineering State University of New York at Buffalo Buffalo, NY, USA chandola@buffalo.edu Chandola@UB CSE 474/574 1 / 21

More information

Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function.

Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function. Bayesian learning: Machine learning comes from Bayesian decision theory in statistics. There we want to minimize the expected value of the loss function. Let y be the true label and y be the predicted

More information

CPSC 540: Machine Learning

CPSC 540: Machine Learning CPSC 540: Machine Learning Multivariate Gaussians Mark Schmidt University of British Columbia Winter 2019 Last Time: Multivariate Gaussian http://personal.kenyon.edu/hartlaub/mellonproject/bivariate2.html

More information

Learning features by contrasting natural images with noise

Learning features by contrasting natural images with noise Learning features by contrasting natural images with noise Michael Gutmann 1 and Aapo Hyvärinen 12 1 Dept. of Computer Science and HIIT, University of Helsinki, P.O. Box 68, FIN-00014 University of Helsinki,

More information

Statistical Convergence of Kernel CCA

Statistical Convergence of Kernel CCA Statistical Convergence of Kernel CCA Kenji Fukumizu Institute of Statistical Mathematics Tokyo 106-8569 Japan fukumizu@ism.ac.jp Francis R. Bach Centre de Morphologie Mathematique Ecole des Mines de Paris,

More information

COMS 4721: Machine Learning for Data Science Lecture 10, 2/21/2017

COMS 4721: Machine Learning for Data Science Lecture 10, 2/21/2017 COMS 4721: Machine Learning for Data Science Lecture 10, 2/21/2017 Prof. John Paisley Department of Electrical Engineering & Data Science Institute Columbia University FEATURE EXPANSIONS FEATURE EXPANSIONS

More information

Statistical Pattern Recognition

Statistical Pattern Recognition Statistical Pattern Recognition Feature Extraction Hamid R. Rabiee Jafar Muhammadi, Alireza Ghasemi, Payam Siyari Spring 2014 http://ce.sharif.edu/courses/92-93/2/ce725-2/ Agenda Dimensionality Reduction

More information

Interpreting Deep Classifiers

Interpreting Deep Classifiers Ruprecht-Karls-University Heidelberg Faculty of Mathematics and Computer Science Seminar: Explainable Machine Learning Interpreting Deep Classifiers by Visual Distillation of Dark Knowledge Author: Daniela

More information

Metric Embedding for Kernel Classification Rules

Metric Embedding for Kernel Classification Rules Metric Embedding for Kernel Classification Rules Bharath K. Sriperumbudur University of California, San Diego (Joint work with Omer Lang & Gert Lanckriet) Bharath K. Sriperumbudur (UCSD) Metric Embedding

More information

unsupervised learning

unsupervised learning unsupervised learning INF5860 Machine Learning for Image Analysis Ole-Johan Skrede 18.04.2018 University of Oslo Messages Mandatory exercise 3 is hopefully out next week. No lecture next week. But there

More information

What is semi-supervised learning?

What is semi-supervised learning? What is semi-supervised learning? In many practical learning domains, there is a large supply of unlabeled data but limited labeled data, which can be expensive to generate text processing, video-indexing,

More information

Semi-Supervised Learning through Principal Directions Estimation

Semi-Supervised Learning through Principal Directions Estimation Semi-Supervised Learning through Principal Directions Estimation Olivier Chapelle, Bernhard Schölkopf, Jason Weston Max Planck Institute for Biological Cybernetics, 72076 Tübingen, Germany {first.last}@tuebingen.mpg.de

More information

Manifold Learning: Theory and Applications to HRI

Manifold Learning: Theory and Applications to HRI Manifold Learning: Theory and Applications to HRI Seungjin Choi Department of Computer Science Pohang University of Science and Technology, Korea seungjin@postech.ac.kr August 19, 2008 1 / 46 Greek Philosopher

More information

Expectation Maximization

Expectation Maximization Expectation Maximization Machine Learning CSE546 Carlos Guestrin University of Washington November 13, 2014 1 E.M.: The General Case E.M. widely used beyond mixtures of Gaussians The recipe is the same

More information

Mixture Models & EM. Nicholas Ruozzi University of Texas at Dallas. based on the slides of Vibhav Gogate

Mixture Models & EM. Nicholas Ruozzi University of Texas at Dallas. based on the slides of Vibhav Gogate Mixture Models & EM icholas Ruozzi University of Texas at Dallas based on the slides of Vibhav Gogate Previously We looed at -means and hierarchical clustering as mechanisms for unsupervised learning -means

More information

Lecture 10: Dimension Reduction Techniques

Lecture 10: Dimension Reduction Techniques Lecture 10: Dimension Reduction Techniques Radu Balan Department of Mathematics, AMSC, CSCAMM and NWC University of Maryland, College Park, MD April 17, 2018 Input Data It is assumed that there is a set

More information

Learning gradients: prescriptive models

Learning gradients: prescriptive models Department of Statistical Science Institute for Genome Sciences & Policy Department of Computer Science Duke University May 11, 2007 Relevant papers Learning Coordinate Covariances via Gradients. Sayan

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

Mixture Models & EM. Nicholas Ruozzi University of Texas at Dallas. based on the slides of Vibhav Gogate

Mixture Models & EM. Nicholas Ruozzi University of Texas at Dallas. based on the slides of Vibhav Gogate Mixture Models & EM icholas Ruozzi University of Texas at Dallas based on the slides of Vibhav Gogate Previously We looed at -means and hierarchical clustering as mechanisms for unsupervised learning -means

More information

Series 7, May 22, 2018 (EM Convergence)

Series 7, May 22, 2018 (EM Convergence) Exercises Introduction to Machine Learning SS 2018 Series 7, May 22, 2018 (EM Convergence) Institute for Machine Learning Dept. of Computer Science, ETH Zürich Prof. Dr. Andreas Krause Web: https://las.inf.ethz.ch/teaching/introml-s18

More information

Basis Expansion and Nonlinear SVM. Kai Yu

Basis Expansion and Nonlinear SVM. Kai Yu Basis Expansion and Nonlinear SVM Kai Yu Linear Classifiers f(x) =w > x + b z(x) = sign(f(x)) Help to learn more general cases, e.g., nonlinear models 8/7/12 2 Nonlinear Classifiers via Basis Expansion

More information

Kernel-based Feature Extraction under Maximum Margin Criterion

Kernel-based Feature Extraction under Maximum Margin Criterion Kernel-based Feature Extraction under Maximum Margin Criterion Jiangping Wang, Jieyan Fan, Huanghuang Li, and Dapeng Wu 1 Department of Electrical and Computer Engineering, University of Florida, Gainesville,

More information

Multiple Similarities Based Kernel Subspace Learning for Image Classification

Multiple Similarities Based Kernel Subspace Learning for Image Classification Multiple Similarities Based Kernel Subspace Learning for Image Classification Wang Yan, Qingshan Liu, Hanqing Lu, and Songde Ma National Laboratory of Pattern Recognition, Institute of Automation, Chinese

More information

CS8803: Statistical Techniques in Robotics Byron Boots. Hilbert Space Embeddings

CS8803: Statistical Techniques in Robotics Byron Boots. Hilbert Space Embeddings CS8803: Statistical Techniques in Robotics Byron Boots Hilbert Space Embeddings 1 Motivation CS8803: STR Hilbert Space Embeddings 2 Overview Multinomial Distributions Marginal, Joint, Conditional Sum,

More information

COMP 551 Applied Machine Learning Lecture 13: Dimension reduction and feature selection

COMP 551 Applied Machine Learning Lecture 13: Dimension reduction and feature selection COMP 551 Applied Machine Learning Lecture 13: Dimension reduction and feature selection Instructor: Herke van Hoof (herke.vanhoof@cs.mcgill.ca) Based on slides by:, Jackie Chi Kit Cheung Class web page:

More information

Lecture 6: September 19

Lecture 6: September 19 36-755: Advanced Statistical Theory I Fall 2016 Lecture 6: September 19 Lecturer: Alessandro Rinaldo Scribe: YJ Choe Note: LaTeX template courtesy of UC Berkeley EECS dept. Disclaimer: These notes have

More information

Data Mining Techniques

Data Mining Techniques Data Mining Techniques CS 6220 - Section 2 - Spring 2017 Lecture 4 Jan-Willem van de Meent (credit: Yijun Zhao, Arthur Gretton Rasmussen & Williams, Percy Liang) Kernel Regression Basis function regression

More information

PCA & ICA. CE-717: Machine Learning Sharif University of Technology Spring Soleymani

PCA & ICA. CE-717: Machine Learning Sharif University of Technology Spring Soleymani PCA & ICA CE-717: Machine Learning Sharif University of Technology Spring 2015 Soleymani Dimensionality Reduction: Feature Selection vs. Feature Extraction Feature selection Select a subset of a given

More information

22 : Hilbert Space Embeddings of Distributions

22 : Hilbert Space Embeddings of Distributions 10-708: Probabilistic Graphical Models 10-708, Spring 2014 22 : Hilbert Space Embeddings of Distributions Lecturer: Eric P. Xing Scribes: Sujay Kumar Jauhar and Zhiguang Huo 1 Introduction and Motivation

More information

Bayesian Regularization

Bayesian Regularization Bayesian Regularization Aad van der Vaart Vrije Universiteit Amsterdam International Congress of Mathematicians Hyderabad, August 2010 Contents Introduction Abstract result Gaussian process priors Co-authors

More information

Density Estimation (II)

Density Estimation (II) Density Estimation (II) Yesterday Overview & Issues Histogram Kernel estimators Ideogram Today Further development of optimization Estimating variance and bias Adaptive kernels Multivariate kernel estimation

More information

Machine Learning Basics: Maximum Likelihood Estimation

Machine Learning Basics: Maximum Likelihood Estimation Machine Learning Basics: Maximum Likelihood Estimation Sargur N. srihari@cedar.buffalo.edu This is part of lecture slides on Deep Learning: http://www.cedar.buffalo.edu/~srihari/cse676 1 Topics 1. Learning

More information

Discrete Mathematics and Probability Theory Fall 2015 Lecture 21

Discrete Mathematics and Probability Theory Fall 2015 Lecture 21 CS 70 Discrete Mathematics and Probability Theory Fall 205 Lecture 2 Inference In this note we revisit the problem of inference: Given some data or observations from the world, what can we infer about

More information

UNIVERSITY of PENNSYLVANIA CIS 520: Machine Learning Final, Fall 2013

UNIVERSITY of PENNSYLVANIA CIS 520: Machine Learning Final, Fall 2013 UNIVERSITY of PENNSYLVANIA CIS 520: Machine Learning Final, Fall 2013 Exam policy: This exam allows two one-page, two-sided cheat sheets; No other materials. Time: 2 hours. Be sure to write your name and

More information

CPSC 540: Machine Learning

CPSC 540: Machine Learning CPSC 540: Machine Learning Mixture Models, Density Estimation, Factor Analysis Mark Schmidt University of British Columbia Winter 2016 Admin Assignment 2: 1 late day to hand it in now. Assignment 3: Posted,

More information

Dimensionality Reduction Using the Sparse Linear Model: Supplementary Material

Dimensionality Reduction Using the Sparse Linear Model: Supplementary Material Dimensionality Reduction Using the Sparse Linear Model: Supplementary Material Ioannis Gkioulekas arvard SEAS Cambridge, MA 038 igkiou@seas.harvard.edu Todd Zickler arvard SEAS Cambridge, MA 038 zickler@seas.harvard.edu

More information

Supplementary Material for: Spectral Unsupervised Parsing with Additive Tree Metrics

Supplementary Material for: Spectral Unsupervised Parsing with Additive Tree Metrics Supplementary Material for: Spectral Unsupervised Parsing with Additive Tree Metrics Ankur P. Parikh School of Computer Science Carnegie Mellon University apparikh@cs.cmu.edu Shay B. Cohen School of Informatics

More information

CSC 411 Lecture 12: Principal Component Analysis

CSC 411 Lecture 12: Principal Component Analysis CSC 411 Lecture 12: Principal Component Analysis Roger Grosse, Amir-massoud Farahmand, and Juan Carrasquilla University of Toronto UofT CSC 411: 12-PCA 1 / 23 Overview Today we ll cover the first unsupervised

More information

LECTURE NOTE #11 PROF. ALAN YUILLE

LECTURE NOTE #11 PROF. ALAN YUILLE LECTURE NOTE #11 PROF. ALAN YUILLE 1. NonLinear Dimension Reduction Spectral Methods. The basic idea is to assume that the data lies on a manifold/surface in D-dimensional space, see figure (1) Perform

More information

Classification for High Dimensional Problems Using Bayesian Neural Networks and Dirichlet Diffusion Trees

Classification for High Dimensional Problems Using Bayesian Neural Networks and Dirichlet Diffusion Trees Classification for High Dimensional Problems Using Bayesian Neural Networks and Dirichlet Diffusion Trees Rafdord M. Neal and Jianguo Zhang Presented by Jiwen Li Feb 2, 2006 Outline Bayesian view of feature

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

PCA with random noise. Van Ha Vu. Department of Mathematics Yale University

PCA with random noise. Van Ha Vu. Department of Mathematics Yale University PCA with random noise Van Ha Vu Department of Mathematics Yale University An important problem that appears in various areas of applied mathematics (in particular statistics, computer science and numerical

More information

Statistical Machine Learning

Statistical Machine Learning Statistical Machine Learning Christoph Lampert Spring Semester 2015/2016 // Lecture 12 1 / 36 Unsupervised Learning Dimensionality Reduction 2 / 36 Dimensionality Reduction Given: data X = {x 1,..., x

More information

Dimension Reduction Techniques. Presented by Jie (Jerry) Yu

Dimension Reduction Techniques. Presented by Jie (Jerry) Yu Dimension Reduction Techniques Presented by Jie (Jerry) Yu Outline Problem Modeling Review of PCA and MDS Isomap Local Linear Embedding (LLE) Charting Background Advances in data collection and storage

More information

Notes on Regularized Least Squares Ryan M. Rifkin and Ross A. Lippert

Notes on Regularized Least Squares Ryan M. Rifkin and Ross A. Lippert Computer Science and Artificial Intelligence Laboratory Technical Report MIT-CSAIL-TR-2007-025 CBCL-268 May 1, 2007 Notes on Regularized Least Squares Ryan M. Rifkin and Ross A. Lippert massachusetts institute

More information

Lecture 11: Unsupervised Machine Learning

Lecture 11: Unsupervised Machine Learning CSE517A Machine Learning Spring 2018 Lecture 11: Unsupervised Machine Learning Instructor: Marion Neumann Scribe: Jingyu Xin Reading: fcml Ch6 (Intro), 6.2 (k-means), 6.3 (Mixture Models); [optional]:

More information

Entropic Affinities: Properties and Efficient Numerical Computation

Entropic Affinities: Properties and Efficient Numerical Computation Entropic Affinities: Properties and Efficient Numerical Computation Max Vladymyrov and Miguel Carreira-Perpiñán Electrical Engineering and Computer Science University of California, Merced http://eecs.ucmerced.edu

More information

Fast Learning with Noise in Deep Neural Nets

Fast Learning with Noise in Deep Neural Nets Fast Learning with Noise in Deep Neural Nets Zhiyun Lu U. of Southern California Los Angeles, CA 90089 zhiyunlu@usc.edu Zi Wang Massachusetts Institute of Technology Cambridge, MA 02139 ziwang.thu@gmail.com

More information

Singular Value Decomposition

Singular Value Decomposition Chapter 6 Singular Value Decomposition In Chapter 5, we derived a number of algorithms for computing the eigenvalues and eigenvectors of matrices A R n n. Having developed this machinery, we complete our

More information

Regression on Manifolds Using Kernel Dimension Reduction

Regression on Manifolds Using Kernel Dimension Reduction Jens Nilsson JENSN@MATHS.LTH.SE Centre for Mathematical Sciences, Lund University, Box 118, SE-221 00 Lund, Sweden Fei Sha FEISHA@CS.BERKELEY.EDU Computer Science Division, University of California, Berkeley,

More information

Distinguishing Causes from Effects using Nonlinear Acyclic Causal Models

Distinguishing Causes from Effects using Nonlinear Acyclic Causal Models Distinguishing Causes from Effects using Nonlinear Acyclic Causal Models Kun Zhang Dept of Computer Science and HIIT University of Helsinki 14 Helsinki, Finland kun.zhang@cs.helsinki.fi Aapo Hyvärinen

More information

Elements of Positive Definite Kernel and Reproducing Kernel Hilbert Space

Elements of Positive Definite Kernel and Reproducing Kernel Hilbert Space Elements of Positive Definite Kernel and Reproducing Kernel Hilbert Space Statistical Inference with Reproducing Kernel Hilbert Space Kenji Fukumizu Institute of Statistical Mathematics, ROIS Department

More information

PCA, Kernel PCA, ICA

PCA, Kernel PCA, ICA PCA, Kernel PCA, ICA Learning Representations. Dimensionality Reduction. Maria-Florina Balcan 04/08/2015 Big & High-Dimensional Data High-Dimensions = Lot of Features Document classification Features per

More information

Mathematical Foundations of the Generalization of t-sne and SNE for Arbitrary Divergences

Mathematical Foundations of the Generalization of t-sne and SNE for Arbitrary Divergences MACHINE LEARNING REPORTS Mathematical Foundations of the Generalization of t-sne and SNE for Arbitrary Report 02/2010 Submitted: 01.04.2010 Published:26.04.2010 T. Villmann and S. Haase University of Applied

More information

Beyond the Point Cloud: From Transductive to Semi-Supervised Learning

Beyond the Point Cloud: From Transductive to Semi-Supervised Learning Beyond the Point Cloud: From Transductive to Semi-Supervised Learning Vikas Sindhwani, Partha Niyogi, Mikhail Belkin Andrew B. Goldberg goldberg@cs.wisc.edu Department of Computer Sciences University of

More information

Data Preprocessing Tasks

Data Preprocessing Tasks Data Tasks 1 2 3 Data Reduction 4 We re here. 1 Dimensionality Reduction Dimensionality reduction is a commonly used approach for generating fewer features. Typically used because too many features can

More information

Self-Tuning Spectral Clustering

Self-Tuning Spectral Clustering Self-Tuning Spectral Clustering Lihi Zelnik-Manor Pietro Perona Department of Electrical Engineering Department of Electrical Engineering California Institute of Technology California Institute of Technology

More information

COMS 4721: Machine Learning for Data Science Lecture 19, 4/6/2017

COMS 4721: Machine Learning for Data Science Lecture 19, 4/6/2017 COMS 4721: Machine Learning for Data Science Lecture 19, 4/6/2017 Prof. John Paisley Department of Electrical Engineering & Data Science Institute Columbia University PRINCIPAL COMPONENT ANALYSIS DIMENSIONALITY

More information

Table of Contents. Multivariate methods. Introduction II. Introduction I

Table of Contents. Multivariate methods. Introduction II. Introduction I Table of Contents Introduction Antti Penttilä Department of Physics University of Helsinki Exactum summer school, 04 Construction of multinormal distribution Test of multinormality with 3 Interpretation

More information

Chemometrics: Classification of spectra

Chemometrics: Classification of spectra Chemometrics: Classification of spectra Vladimir Bochko Jarmo Alander University of Vaasa November 1, 2010 Vladimir Bochko Chemometrics: Classification 1/36 Contents Terminology Introduction Big picture

More information

Learning Binary Classifiers for Multi-Class Problem

Learning Binary Classifiers for Multi-Class Problem Research Memorandum No. 1010 September 28, 2006 Learning Binary Classifiers for Multi-Class Problem Shiro Ikeda The Institute of Statistical Mathematics 4-6-7 Minami-Azabu, Minato-ku, Tokyo, 106-8569,

More information

Statistics: Learning models from data

Statistics: Learning models from data DS-GA 1002 Lecture notes 5 October 19, 2015 Statistics: Learning models from data Learning models from data that are assumed to be generated probabilistically from a certain unknown distribution is a crucial

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

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 8 Continuous Latent Variable

More information

2 Tikhonov Regularization and ERM

2 Tikhonov Regularization and ERM Introduction Here we discusses how a class of regularization methods originally designed to solve ill-posed inverse problems give rise to regularized learning algorithms. These algorithms are kernel methods

More information

Learning a Kernel Matrix for Nonlinear Dimensionality Reduction

Learning a Kernel Matrix for Nonlinear Dimensionality Reduction Learning a Kernel Matrix for Nonlinear Dimensionality Reduction Kilian Q. Weinberger kilianw@cis.upenn.edu Fei Sha feisha@cis.upenn.edu Lawrence K. Saul lsaul@cis.upenn.edu Department of Computer and Information

More information

Lecture 13. Principal Component Analysis. Brett Bernstein. April 25, CDS at NYU. Brett Bernstein (CDS at NYU) Lecture 13 April 25, / 26

Lecture 13. Principal Component Analysis. Brett Bernstein. April 25, CDS at NYU. Brett Bernstein (CDS at NYU) Lecture 13 April 25, / 26 Principal Component Analysis Brett Bernstein CDS at NYU April 25, 2017 Brett Bernstein (CDS at NYU) Lecture 13 April 25, 2017 1 / 26 Initial Question Intro Question Question Let S R n n be symmetric. 1

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

Nonparametric Inference via Bootstrapping the Debiased Estimator

Nonparametric Inference via Bootstrapping the Debiased Estimator Nonparametric Inference via Bootstrapping the Debiased Estimator Yen-Chi Chen Department of Statistics, University of Washington ICSA-Canada Chapter Symposium 2017 1 / 21 Problem Setup Let X 1,, X n be

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

Kernel Principal Component Analysis

Kernel Principal Component Analysis Kernel Principal Component Analysis Seungjin Choi Department of Computer Science and Engineering Pohang University of Science and Technology 77 Cheongam-ro, Nam-gu, Pohang 37673, Korea seungjin@postech.ac.kr

More information

Latent Variable Models and EM algorithm

Latent Variable Models and EM algorithm Latent Variable Models and EM algorithm SC4/SM4 Data Mining and Machine Learning, Hilary Term 2017 Dino Sejdinovic 3.1 Clustering and Mixture Modelling K-means and hierarchical clustering are non-probabilistic

More information

L 2,1 Norm and its Applications

L 2,1 Norm and its Applications L 2, Norm and its Applications Yale Chang Introduction According to the structure of the constraints, the sparsity can be obtained from three types of regularizers for different purposes.. Flat Sparsity.

More information

Analysis of Spectral Kernel Design based Semi-supervised Learning

Analysis of Spectral Kernel Design based Semi-supervised Learning Analysis of Spectral Kernel Design based Semi-supervised Learning Tong Zhang IBM T. J. Watson Research Center Yorktown Heights, NY 10598 Rie Kubota Ando IBM T. J. Watson Research Center Yorktown Heights,

More information

arxiv: v1 [astro-ph.im] 5 May 2015

arxiv: v1 [astro-ph.im] 5 May 2015 Autoencoding Time Series for Visualisation arxiv:155.936v1 [astro-ph.im] 5 May 215 Nikolaos Gianniotis 1, Dennis Kügler 1, Peter Tiňo 2, Kai Polsterer 1 and Ranjeev Misra 3 1- Astroinformatics - Heidelberg

More information

The Laplacian PDF Distance: A Cost Function for Clustering in a Kernel Feature Space

The Laplacian PDF Distance: A Cost Function for Clustering in a Kernel Feature Space The Laplacian PDF Distance: A Cost Function for Clustering in a Kernel Feature Space Robert Jenssen, Deniz Erdogmus 2, Jose Principe 2, Torbjørn Eltoft Department of Physics, University of Tromsø, Norway

More information

Kernel Learning via Random Fourier Representations

Kernel Learning via Random Fourier Representations Kernel Learning via Random Fourier Representations L. Law, M. Mider, X. Miscouridou, S. Ip, A. Wang Module 5: Machine Learning L. Law, M. Mider, X. Miscouridou, S. Ip, A. Wang Kernel Learning via Random

More information

Classification. The goal: map from input X to a label Y. Y has a discrete set of possible values. We focused on binary Y (values 0 or 1).

Classification. The goal: map from input X to a label Y. Y has a discrete set of possible values. We focused on binary Y (values 0 or 1). Regression and PCA Classification The goal: map from input X to a label Y. Y has a discrete set of possible values We focused on binary Y (values 0 or 1). But we also discussed larger number of classes

More information

Learning a kernel matrix for nonlinear dimensionality reduction

Learning a kernel matrix for nonlinear dimensionality reduction University of Pennsylvania ScholarlyCommons Departmental Papers (CIS) Department of Computer & Information Science 7-4-2004 Learning a kernel matrix for nonlinear dimensionality reduction Kilian Q. Weinberger

More information