Nonlinear Eigenproblems in Data Analysis

Size: px
Start display at page:

Download "Nonlinear Eigenproblems in Data Analysis"

Transcription

1 Nonlinear Eigenproblems in Data Analysis Matthias Hein Joint work with: Thomas Bühler, Leonardo Jost, Shyam Rangapuram, Simon Setzer Department of Mathematics and Computer Science Saarland University, Saarbrücken, Germany Workshop on Unsupervised Learning and Highdimensional Statistics Ecole Normale Superieure, Paris, Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 1 / 35

2 From linear to nonlinear eigenproblems Linear eigenproblem (LEP) Ax = λ Bx x is critical point of F (x) = x,ax x,bx Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 2 / 35

3 From linear to nonlinear eigenproblems Linear eigenproblem (LEP) Ax = λ Bx x is critical point of F (x) = x,ax x,bx Generalization 0 R(x) λ S(x) = x is critical point of F (x) = R(x) S(x) Nonlinear eigenproblem (NLEP) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 2 / 35

4 From linear to nonlinear eigenproblems Linear eigenproblem (LEP) Ax = λ Bx x is critical point of F (x) = x,ax x,bx Generalization 0 R(x) λ S(x) = x is critical point of F (x) = R(x) S(x) Nonlinear eigenproblem (NLEP) Why do we need nonlinear eigenproblems? The freedom in the choice of R and S in NLEPs allows to address problems not possible before with LEPs Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 2 / 35

5 Benefits of Nonlinear Eigenproblems in Data Analysis Linear EP Nonlinear EP Modeling power low high Relaxation of loose tight combinatorial problems Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 3 / 35

6 Benefits of Nonlinear Eigenproblems in Data Analysis Linear EP Nonlinear EP Modeling power low high Relaxation of loose tight combinatorial problems Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 3 / 35

7 Robust PCA Type of eigenproblem Ratio n PCA Linear n j=1 X j 2 i i=1 w,x 1 n w 2 2 Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 4 / 35

8 Robust PCA Source of outliers noisy data adversarial manipulation Type of eigenproblem Ratio Robustness n PCA Linear n j=1 X j 2 i i=1 w,x 1 n w 2 2 no Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 4 / 35

9 Robust PCA Source of outliers noisy data adversarial manipulation PCA Type of eigenproblem Linear Nonlinear n n j=1 X j 2 Ratio i=1 w,x i 1 n w 2 2 R(w) w 2 Robustness no yes Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 4 / 35

10 Robust PCA PCA uses non-robust variance: Var(w) = n w, X i 1 n i=1 n X j 2. j=1 Robust scale measure (Rousseeuw and Croux(1993)) R(w) = Quartile ( w, X i X j, i, j = 1,..., n ) breakdown point 50% requires no location measure can be written as a difference of convex functions in w projection pursuit type technique - algorithms currently based on search over finite set resp. two-dimensional grid search Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 5 / 35

11 Challenges of Nonlinear Eigenproblems Linear EP Nonlinear EP Numerical symmetric EP almost Algorithms solved nothing Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 6 / 35

12 (Inverse) Power Method for Nonlinear Eigenproblems Inverse Power Method for Linear Eigenproblems {1 Af k+1 = B f k f k+1 = arg min u R n 2 u, Au u, Bf k } Sequence f k converges to smallest eigenvector of generalized eigenproblem. Inverse Power Method for Nonlinear Eigenproblems (H.,B.(2010)) p > 1 p = 1 g k+1 = arg min {R(u) u, s(f k ) } f k+1 = arg min {R(u) λ k u, s(f k ) } u R n f k+1 = g k+1 /S(g k+1 ) 1/p u 2 1 s(f k+1 ) S(f k+1 ) s(f k+1 ) S(f k+1 ) λ k+1 = R(f k+1) S(f k+1 ) λ k+1 = R(f k+1) S(f k+1 ) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 7 / 35

13 Properties of Nonlinear Inverse Power Method Theorem (Hein, Bühler (NIPS 2010)): It holds either λ k+1 < λ k or λ k+1 = λ k and the sequence terminates. Moreover, for every cluster point f of the sequence f k one has 0 R(f ) λ S(f ), where λ = R(f ) S(f ). Guarantees: monotonic descent method subconvergence guaranteed to some nonlinear eigenvector but not necessarily the one associated with the smallest eigenvalue. additional inclusion of proximal term allows to show f k+1 f k 0 as k. Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 8 / 35

14 Benefits of Nonlinear Eigenproblems in Data Analysis Linear EP Nonlinear EP Modeling power low high Relaxation of loose tight combinatorial problems Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 9 / 35

15 Balanced Graph Cuts - Applications Clustering/Community detection Image Segmentation Parallel Computing (Matrix Reordering) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 10 / 35

16 The Cheeger Cut Problem Cheeger cut: (C, C) is a partition of the weighted, undirected graph φ(c) = cut(c, C) min{ C,, where cut(a, B) = C } i A,j B Optimal Cheeger cut, φ = min φ(c), is NP-hard C w ij Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 11 / 35

17 Relaxation of Cheeger Cut Problem Relaxation into semi-definite program with V 3 constraints: Best known (worst case) approximation guarantee: O( log V ). Spectral Relaxation based on graph Laplacian L L = D W, Isoperimetric inequality (Alon, Milman (1984)) (φ ) 2 2 max i d i λ 2 (L) 2φ. bipartitioning obtained by optimal thresholding of second eigenvector φ φ SPECTRAL 2 max i d i φ there are graphs known which realize upper bound Extension for graph p-laplacian (Amghibech (2003), Bühler, H.(2009)) φ φ p SPECTRAL p(max i d i ) p 1 p (φ ) 1 p. Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 12 / 35

18 1-Spectral Clustering 1-graph Laplacian: The nonlinear graph 1-Laplacian 1 induces the functional F 1 (f ), F 1 (f ) := f, 1f f 1 = 1 2 n i,j=1 w ij f i f j f 1. Theorem (Hein,Bühler (NIPS 2010)): Let G be connected, then min C cut(c, C) min{ C, C } = min F 1 (f ) = λ 2 ( 1 ), f nonconstant median(f)=0 where λ 2 ( 1 ) is the second smallest eigenvalue of 1. The second eigenvector of 1 is the indicator vector of the optimal partition. Tight relaxation of the optimal Cheeger cut! Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 13 / 35

19 Cheeger Cut: 1-Laplacian (NLEP) vs. 2-Laplacian (LEP) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 14 / 35

20 NIPM for 1-Spectral Clustering NIPM for 1-Spectral Clustering: Computing a nonconstant 1-eigenvector of the graph 1-Laplacian 1: Input: weight matrix W 2: Initialization: nonconstant f 0 with median(f 0 ) = 0 and f 0 1 = 1, accuracy ɛ 3: repeat 4: g k+1 = arg min 5: f f k+1 = g k+1 median ( g k+1) 6: v k+1 1 (f k+1 ), 7: λ k+1 = F 1 (f k+1 ) 8: until λk+1 λ k < ɛ λ k { 1 2 n i,j=1 w ij f i f j λ k f, v k } Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 15 / 35

21 Balanced Graph Cuts and Nonlinear Eigenproblems min A V cut(a, A) Ŝ(A) = min f R n 1 2 n i,j=1 w ij f i f j S(f ) Balancing set function Ŝ and their Lovasz extensions: Ŝ(A) S(f ) Cheeger cut min{ A, A } f median(f )1 1 1 Ratio cut A A n { 2 i,j=1 f i f j 1, if min{ A, A } K Hard balanced cut... 0, else. Modeling of different bias towards balanced partitions via choice of Ŝ. Tight relaxations for all balancing set functions (H., Setzer (NIPS 2011)) Earlier Work: Riemannian manifolds: Rothaus (1985) Graphs: Tillich (2000), Houdre (2001) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 16 / 35.

22 Graph Partitioning Benchmark Results Graph partitioning benchmark of Chris Walshaw (since 2000) contains results of all major methods (linear, spectral, SDP relaxation, METIS and variants, evolutionary, multi-agent methods,...). Objective is the cut subject to a strict balancing constraint. For α {0, 1, 3, 5}, { min cut(c, C) C ( max{ C, C } 1 + α 100 ) V } graphs with 2395 V and 6837 E Bipartitioning: 1-Spectral Clustering yields better cuts in 16 out of 136 possible problems 16 better than - 89 equal to - 31 worse than best reported cut Multipartitioning: Results on 34 graphs, Better cuts in 274 out of 816 problems Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 17 / 35

23 Constrained Fractional Set Programs min =C V ˆR(C) Ŝ(C) subj. to : Mi (C) k i, i = 1,..., K has a tight relaxation into a nonlinear eigenproblem if ˆR, Ŝ are non-negative set functions ˆR( ) = Ŝ( ) = 0 The constraint functions ˆM i underlie no restrictions Includes a large class of graph-based clustering/community detection problems Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 18 / 35

24 Problem classes Unconstrained Problems: balanced graph cuts (Total variation (TV) on graphs: 1 2 n i,j=1 w ij f i f j ) balanced cuts for directed graphs and hypergraphs TV directed graphs: n i,j=1 w ij max{f i f j, 0} (in preparation) TV on hypergraphs: e E w(e) max f i f j (NIPS 2013) i,j e maximum density subgraph problem Constrained Problems: cardinality constraints (ICML 2013, WWW 2013) seed constraints (ICML 2013, WWW 2013) link constraints (AISTATS, 2012) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 19 / 35

25 Definition Let f R V be ordered in increasing order f 1 f 2... f n and define C i = {j V f j > f i }.Then S : R V R given by S(f ) = n i=1 ) n 1 f i (Ŝ(Ci 1 ) Ŝ(C i ) = Ŝ(C i )(f i+1 f i ) + f 1 Ŝ(V ) i=1 is the Lovasz extension of Ŝ : 2 V R. One has S(1 A ) = Ŝ(A), A V. Definition A set function ˆF : 2 V R is submodular if for all A, B V, ˆF (A B) + ˆF (A B) ˆF (A) + ˆF (B). Lemma The Lovasz extension of a set function is convex iff the set function is submodular. Proposition Every set function can be written as difference of two submodular set functions. Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 20 / 35

26 Tight Relaxation of General Ratios Theorem Let ˆR, Ŝ : 2V R be non-negative set functions and R, S : R n R their Lovasz extensions. Then, If in addition ˆR(V ) = Moreover, min C V ˆR(C) Ŝ(C) = min f R n + Ŝ(V ) = 0, then min C V R(f ) S(f ) min i=1,...,n R(f ) S(f ). ˆR(C) = min R(f ) Ŝ(C) f R n S(f ). ˆR(C i ) Ŝ(C i ), where C i = {j V f j > f i }. If ˆR(V ) = Ŝ(V ) = 0, the inequality holds for all f R n. Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 21 / 35

27 Inclusion of Constraints Idea of the proof: turn the constrained combinatorial ratio problem into an unconstrained ratio problem via an exact penalty approach. use the previous theorem to get the tight relaxation Penalty Function: The penalty set function for a constraint ˆM i (C) k i is defined as { { max 0, k T i (C) = i M } i (C), C, 0, C =, Equivalence of Constrained Fractional Set Programs min M i (C) k i, i=1,...,k ˆR(C) Ŝ(C) = min C V ˆR(C) + γ k i=1 T i (C) Ŝ(C) The constant γ can be computed using a feasible set. Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 22 / 35

28 Optimization - RatioDCA Optimization Problem: R(f ) min f R n + S(f ) =: R 1(f ) R 2 (f ) S 1 (f ) S 2 (f ). Minimize non-negative ratio of one-homog. d.c functions over R n + { f l+1 = arg min R 1 (u) u, r 2 (f l ) + λ l( ) } S 2 (u) u, s 1 (f l ) u R n +, u 2 1 r 2 (f l ) R 2 (f l ), s 1 (f l ) S 1 (f l ) λ l+1 = R(f l+1) )/S(f l+1 ) sequence of convex optimization problems same convergence properties as nonlinear inverse power method finite convergence for combinatorial problems Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 23 / 35

29 Quality Guarantee Tight relaxation of combinatorial fractional program: Minimization of continuous relaxation is as hard as original combinatorial problem = non-convex and non-smooth No guarantee that one obtains optimal solution by RatioDCA! Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 24 / 35

30 Quality Guarantee Tight relaxation of combinatorial fractional program: Minimization of continuous relaxation is as hard as original combinatorial problem = non-convex and non-smooth No guarantee that one obtains optimal solution by RatioDCA! but Theorem Let A V be feasible and γ chosen such that equivalence holds. If one uses as initialization of RatioDCA, f 0 = 1 A, then either it terminates after one step or it yields an f 1 which after optimal thresholding gives a partition (B, B) which is feasible and ˆR(B) Ŝ(B) < ˆR(A) Ŝ(A). Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 24 / 35

31 Clustering in Practice Major Goal: Integration of prior knowledge in clustering/community detection problems via constraints! Prior work: Link constraints in k-means (Wagstaff et al, ICML 2001) Link constraints in spectral clustering (Eriksson et al, ICCV 2007, Wang & Davidson, KDD 2010) Communities from seed sets (Andersen, Lang (WWW 2006)) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 25 / 35

32 Constrained Normalized Cut Clustering with link constraints (Rangapuram, H. (AISTATS 2012)) must-link and cannot-link constraints a partition is called consistent if all constraints are satisfied Constrained normalized cut problem: min (C,C) consistent cut(c, C) vol(c) vol(c) first method that can guarantee that constraints are satisfied! Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 26 / 35

33 Constrained Normalized Cut - II Must-link and cannot-link constraints Unconstrained normalized cut Constrained normalized cut Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 27 / 35

34 Constrained Clustering in Image Segmentation Constrained Spectral Clustering 1-Spectral Clustering with linear equality constr. Original Image (Rangapuram, H., AISTATS 2012) (Eriksson et al, ICCV 2007) Labels Norm. Cut: Norm. Cut: Norm. Cut: Norm. Cut: Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 28 / 35

35 Constrained Normalized Cut - Results III Binary-partitioning problem (Spam dataset V = 4207): Multi-partitioning problem (extended MNIST dataset, V = ): SCLC: Eriksson, Olsson, Kahl (ICCV 2007), Xu, Li, Schuurmans, (CVPR 2009) CCSR: Li, Liu, Tang (CVPR 2009) CSP: Wang, Davidson, (KDD 2010) SL: Kamvar, Klein, Manning, (IJCAI 2003) COSC: Rangapuram, Hein (AISTATS 2012) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 29 / 35

36 Local Clustering Task of Local Clustering: find set C including seed set J which is well-clustered and small Optimization Problem: Prior work: min C V cut(c, C) Ŝ(C) subj to : vol g (C) k, and J C. Andersen, Chung, Lang (FOCS, 2006) - local PageRank (local random walk) Maji, Vishnoi, Malik (CVPR, 2011), Mahoney, Orrechia, Vishnoi (JMLR, 2012) - biased normalized cuts (local spectral method) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 30 / 35

37 Local Clustering - Results Amazon0505 Cit-HepPh CA-GrQC vertices vertices 4158 vertices edges edges edges LRW - Local Random Walk (Andersen, Lang (2006)), LS - Local Spectral (Mahoney, Orrechia, Vishnoi (2012)), CFSP - our method (Bühler, Rangapuram, Setzer, Hein (ICML, 2013)) Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 31 / 35

38 Local Minima and Critical Points Combinatorial versus continuous problem: Let ˆR, Ŝ : 2V R be non-negative set functions and R, S : R n R their Lovasz extensions. Then, min C V ˆR(C) Ŝ(C) = min f R n + R(f ) S(f ). Global minimum is the same but what about local minima? Motivation: Characterization of local minima of Cheeger cut by Bresson et al. (2012). Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 32 / 35

39 Local Minima - Continuous versus Discrete Problem Theorem Let R, S be the Lovasz extensions of ˆR, Ŝ. Then f = 1 C is a local minimum of R(f ) S(f ) with value λ = R(f ) S(f ) = ˆR(C ) if and only if for any C Ŝ(C ) with C C or C C it holds λ = ˆR(C ) Ŝ(C ) ˆR(C) Ŝ(C). Structure of local minima of continuous and combinatorial problem are (almost) the same Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 33 / 35

40 Guarantees, Local Minima and Cheeger cut Cheeger cut problem: ˆR(C) = cut(c, C) = R(f ) = 1 2 n i,j=1 w ij f i f j Ŝ(C) = min{ C, C } = S(f ) = f median(f ) 1 Proposition If NIPM terminates at f, then f = n i=1 α i1 C i where C i = {j V f j > f i } and R(f ) S(f ) = ˆR(C i ) Ŝ(Ci ) if f = 1 C, where wlog C V 2, then every C C fulfills ˆR(C ) Ŝ(C ) ˆR(C) Ŝ(C). Moreover, if C = C then f is a local minimum. Ongoing work: Global approximation guarantees. Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 34 / 35

41 Conclusion and Outlook Tight relaxation of nonlinear eigenproblems Tight relaxation of combinatorial fractional programs Integration of prior knowledge into clustering tasks RatioDCA makes computation of nonlinear eigenvectors feasible Locally optimal solutions of tight relaxations are better than globally optimal solutions of loose relaxations! Open Problems in Nonlinear Eigenproblems: What is a suitable min-max principle for nonlinear eigenvectors? Computation of higher-order eigenvectors Higher order cuts via min-max principle Approximation guarantees for tight relaxations of combinatorial problems Hein (Saarland University) Nonlinear Eigenproblems in Data Analysis 35 / 35

Inverse Power Method for Non-linear Eigenproblems

Inverse Power Method for Non-linear Eigenproblems Inverse Power Method for Non-linear Eigenproblems Matthias Hein and Thomas Bühler Anubhav Dwivedi Department of Aerospace Engineering & Mechanics 7th March, 2017 1 / 30 OUTLINE Motivation Non-Linear Eigenproblems

More information

An Inverse Power Method for Nonlinear Eigenproblems with Applications in 1-Spectral Clustering and Sparse PCA

An Inverse Power Method for Nonlinear Eigenproblems with Applications in 1-Spectral Clustering and Sparse PCA An Inverse Power Method for Nonlinear Eigenproblems with Applications in 1-Spectral Clustering and Sparse PCA Matthias Hein Thomas Bühler Saarland University, Saarbrücken, Germany {hein,tb}@cs.uni-saarland.de

More information

An Inverse Power Method for Nonlinear Eigenproblems with Applications in 1-Spectral Clustering and Sparse PCA

An Inverse Power Method for Nonlinear Eigenproblems with Applications in 1-Spectral Clustering and Sparse PCA An Inverse Power Method for Nonlinear Eigenproblems with Applications in -Spectral Clustering and Sparse PCA Matthias Hein Thomas Bühler Saarland University, Saarbrücken, Germany {hein,tb}@csuni-saarlandde

More information

Asymmetric Cheeger cut and application to multi-class unsupervised clustering

Asymmetric Cheeger cut and application to multi-class unsupervised clustering Asymmetric Cheeger cut and application to multi-class unsupervised clustering Xavier Bresson Thomas Laurent April 8, 0 Abstract Cheeger cut has recently been shown to provide excellent classification results

More information

The Total Variation on Hypergraphs - Learning on Hypergraphs Revisited

The Total Variation on Hypergraphs - Learning on Hypergraphs Revisited The Total Variation on Hypergraphs - Learning on Hypergraphs Revisited Matthias Hein, Simon Setzer, Leonardo Jost and Syama Sundar Rangapuram Department of Computer Science Saarland University Abstract

More information

Graph Partitioning Algorithms and Laplacian Eigenvalues

Graph Partitioning Algorithms and Laplacian Eigenvalues Graph Partitioning Algorithms and Laplacian Eigenvalues Luca Trevisan Stanford Based on work with Tsz Chiu Kwok, Lap Chi Lau, James Lee, Yin Tat Lee, and Shayan Oveis Gharan spectral graph theory Use linear

More information

EIGENVECTOR NORMS MATTER IN SPECTRAL GRAPH THEORY

EIGENVECTOR NORMS MATTER IN SPECTRAL GRAPH THEORY EIGENVECTOR NORMS MATTER IN SPECTRAL GRAPH THEORY Franklin H. J. Kenter Rice University ( United States Naval Academy 206) orange@rice.edu Connections in Discrete Mathematics, A Celebration of Ron Graham

More information

Spectral Clustering. Zitao Liu

Spectral Clustering. Zitao Liu Spectral Clustering Zitao Liu Agenda Brief Clustering Review Similarity Graph Graph Laplacian Spectral Clustering Algorithm Graph Cut Point of View Random Walk Point of View Perturbation Theory Point of

More information

Graph Partitioning Using Random Walks

Graph Partitioning Using Random Walks Graph Partitioning Using Random Walks A Convex Optimization Perspective Lorenzo Orecchia Computer Science Why Spectral Algorithms for Graph Problems in practice? Simple to implement Can exploit very efficient

More information

Spectral Clustering. Spectral Clustering? Two Moons Data. Spectral Clustering Algorithm: Bipartioning. Spectral methods

Spectral Clustering. Spectral Clustering? Two Moons Data. Spectral Clustering Algorithm: Bipartioning. Spectral methods Spectral Clustering Seungjin Choi Department of Computer Science POSTECH, Korea seungjin@postech.ac.kr 1 Spectral methods Spectral Clustering? Methods using eigenvectors of some matrices Involve eigen-decomposition

More information

Locally-biased analytics

Locally-biased analytics Locally-biased analytics You have BIG data and want to analyze a small part of it: Solution 1: Cut out small part and use traditional methods Challenge: cutting out may be difficult a priori Solution 2:

More information

Semi-supervised Eigenvectors for Locally-biased Learning

Semi-supervised Eigenvectors for Locally-biased Learning Semi-supervised Eigenvectors for Locally-biased Learning Toke Jansen Hansen Section for Cognitive Systems DTU Informatics Technical University of Denmark tjha@imm.dtu.dk Michael W. Mahoney Department of

More information

Lecture 12: Introduction to Spectral Graph Theory, Cheeger s inequality

Lecture 12: Introduction to Spectral Graph Theory, Cheeger s inequality CSE 521: Design and Analysis of Algorithms I Spring 2016 Lecture 12: Introduction to Spectral Graph Theory, Cheeger s inequality Lecturer: Shayan Oveis Gharan May 4th Scribe: Gabriel Cadamuro Disclaimer:

More information

arxiv: v1 [stat.ml] 5 Jun 2013

arxiv: v1 [stat.ml] 5 Jun 2013 Multiclass Total Variation Clustering Xavier Bresson, Thomas Laurent, David Uminsky and James H. von Brecht arxiv:1306.1185v1 [stat.ml] 5 Jun 2013 Abstract Ideas from the image processing literature have

More information

Spectral Clustering on Handwritten Digits Database

Spectral Clustering on Handwritten Digits Database University of Maryland-College Park Advance Scientific Computing I,II Spectral Clustering on Handwritten Digits Database Author: Danielle Middlebrooks Dmiddle1@math.umd.edu Second year AMSC Student Advisor:

More information

Spectral Techniques for Clustering

Spectral Techniques for Clustering Nicola Rebagliati 1/54 Spectral Techniques for Clustering Nicola Rebagliati 29 April, 2010 Nicola Rebagliati 2/54 Thesis Outline 1 2 Data Representation for Clustering Setting Data Representation and Methods

More information

Introduction to Spectral Graph Theory and Graph Clustering

Introduction to Spectral Graph Theory and Graph Clustering Introduction to Spectral Graph Theory and Graph Clustering Chengming Jiang ECS 231 Spring 2016 University of California, Davis 1 / 40 Motivation Image partitioning in computer vision 2 / 40 Motivation

More information

Certifying the Global Optimality of Graph Cuts via Semidefinite Programming: A Theoretic Guarantee for Spectral Clustering

Certifying the Global Optimality of Graph Cuts via Semidefinite Programming: A Theoretic Guarantee for Spectral Clustering Certifying the Global Optimality of Graph Cuts via Semidefinite Programming: A Theoretic Guarantee for Spectral Clustering Shuyang Ling Courant Institute of Mathematical Sciences, NYU Aug 13, 2018 Joint

More information

Spectral Clustering. Guokun Lai 2016/10

Spectral Clustering. Guokun Lai 2016/10 Spectral Clustering Guokun Lai 2016/10 1 / 37 Organization Graph Cut Fundamental Limitations of Spectral Clustering Ng 2002 paper (if we have time) 2 / 37 Notation We define a undirected weighted graph

More information

Machine Learning for Data Science (CS4786) Lecture 11

Machine Learning for Data Science (CS4786) Lecture 11 Machine Learning for Data Science (CS4786) Lecture 11 Spectral clustering Course Webpage : http://www.cs.cornell.edu/courses/cs4786/2016sp/ ANNOUNCEMENT 1 Assignment P1 the Diagnostic assignment 1 will

More information

OWL to the rescue of LASSO

OWL to the rescue of LASSO OWL to the rescue of LASSO IISc IBM day 2018 Joint Work R. Sankaran and Francis Bach AISTATS 17 Chiranjib Bhattacharyya Professor, Department of Computer Science and Automation Indian Institute of Science,

More information

Spectral Generative Models for Graphs

Spectral Generative Models for Graphs Spectral Generative Models for Graphs David White and Richard C. Wilson Department of Computer Science University of York Heslington, York, UK wilson@cs.york.ac.uk Abstract Generative models are well known

More information

Predicting Graph Labels using Perceptron. Shuang Song

Predicting Graph Labels using Perceptron. Shuang Song Predicting Graph Labels using Perceptron Shuang Song shs037@eng.ucsd.edu Online learning over graphs M. Herbster, M. Pontil, and L. Wainer, Proc. 22nd Int. Conf. Machine Learning (ICML'05), 2005 Prediction

More information

Spectral Graph Theory and its Applications. Daniel A. Spielman Dept. of Computer Science Program in Applied Mathematics Yale Unviersity

Spectral Graph Theory and its Applications. Daniel A. Spielman Dept. of Computer Science Program in Applied Mathematics Yale Unviersity Spectral Graph Theory and its Applications Daniel A. Spielman Dept. of Computer Science Program in Applied Mathematics Yale Unviersity Outline Adjacency matrix and Laplacian Intuition, spectral graph drawing

More information

Supplemental Material for Discrete Graph Hashing

Supplemental Material for Discrete Graph Hashing Supplemental Material for Discrete Graph Hashing Wei Liu Cun Mu Sanjiv Kumar Shih-Fu Chang IM T. J. Watson Research Center Columbia University Google Research weiliu@us.ibm.com cm52@columbia.edu sfchang@ee.columbia.edu

More information

Lecture 14: Random Walks, Local Graph Clustering, Linear Programming

Lecture 14: Random Walks, Local Graph Clustering, Linear Programming CSE 521: Design and Analysis of Algorithms I Winter 2017 Lecture 14: Random Walks, Local Graph Clustering, Linear Programming Lecturer: Shayan Oveis Gharan 3/01/17 Scribe: Laura Vonessen Disclaimer: These

More information

Semidefinite and Second Order Cone Programming Seminar Fall 2001 Lecture 5

Semidefinite and Second Order Cone Programming Seminar Fall 2001 Lecture 5 Semidefinite and Second Order Cone Programming Seminar Fall 2001 Lecture 5 Instructor: Farid Alizadeh Scribe: Anton Riabov 10/08/2001 1 Overview We continue studying the maximum eigenvalue SDP, and generalize

More information

An Adaptive Total Variation Algorithm for Computing the Balanced Cut of a Graph

An Adaptive Total Variation Algorithm for Computing the Balanced Cut of a Graph Digital Commons@ Loyola Marymount University and Loyola Law School Mathematics Faculty Wors Mathematics 1-1-013 An Adaptive Total Variation Algorithm for Computing the Balanced Cut of a Graph Xavier Bresson

More information

Semidefinite Programming Basics and Applications

Semidefinite Programming Basics and Applications Semidefinite Programming Basics and Applications Ray Pörn, principal lecturer Åbo Akademi University Novia University of Applied Sciences Content What is semidefinite programming (SDP)? How to represent

More information

Submodularity in Machine Learning

Submodularity in Machine Learning Saifuddin Syed MLRG Summer 2016 1 / 39 What are submodular functions Outline 1 What are submodular functions Motivation Submodularity and Concavity Examples 2 Properties of submodular functions Submodularity

More information

8.1 Concentration inequality for Gaussian random matrix (cont d)

8.1 Concentration inequality for Gaussian random matrix (cont d) MGMT 69: Topics in High-dimensional Data Analysis Falll 26 Lecture 8: Spectral clustering and Laplacian matrices Lecturer: Jiaming Xu Scribe: Hyun-Ju Oh and Taotao He, October 4, 26 Outline Concentration

More information

Lecture 17: Cheeger s Inequality and the Sparsest Cut Problem

Lecture 17: Cheeger s Inequality and the Sparsest Cut Problem Recent Advances in Approximation Algorithms Spring 2015 Lecture 17: Cheeger s Inequality and the Sparsest Cut Problem Lecturer: Shayan Oveis Gharan June 1st Disclaimer: These notes have not been subjected

More information

First, parse the title...

First, parse the title... First, parse the title... Eigenvector localization: Eigenvectors are usually global entities But they can be localized in extremely sparse/noisy graphs/matrices Implicit regularization: Usually exactly

More information

March 13, Paper: R.R. Coifman, S. Lafon, Diffusion maps ([Coifman06]) Seminar: Learning with Graphs, Prof. Hein, Saarland University

March 13, Paper: R.R. Coifman, S. Lafon, Diffusion maps ([Coifman06]) Seminar: Learning with Graphs, Prof. Hein, Saarland University Kernels March 13, 2008 Paper: R.R. Coifman, S. Lafon, maps ([Coifman06]) Seminar: Learning with Graphs, Prof. Hein, Saarland University Kernels Figure: Example Application from [LafonWWW] meaningful geometric

More information

Submodular Functions Properties Algorithms Machine Learning

Submodular Functions Properties Algorithms Machine Learning Submodular Functions Properties Algorithms Machine Learning Rémi Gilleron Inria Lille - Nord Europe & LIFL & Univ Lille Jan. 12 revised Aug. 14 Rémi Gilleron (Mostrare) Submodular Functions Jan. 12 revised

More information

A Local Spectral Method for Graphs: With Applications to Improving Graph Partitions and Exploring Data Graphs Locally

A Local Spectral Method for Graphs: With Applications to Improving Graph Partitions and Exploring Data Graphs Locally Journal of Machine Learning Research 13 (2012) 2339-2365 Submitted 11/12; Published 9/12 A Local Spectral Method for Graphs: With Applications to Improving Graph Partitions and Exploring Data Graphs Locally

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

Lecture: Local Spectral Methods (3 of 4) 20 An optimization perspective on local spectral methods

Lecture: Local Spectral Methods (3 of 4) 20 An optimization perspective on local spectral methods Stat260/CS294: Spectral Graph Methods Lecture 20-04/07/205 Lecture: Local Spectral Methods (3 of 4) Lecturer: Michael Mahoney Scribe: Michael Mahoney Warning: these notes are still very rough. They provide

More information

MATH 567: Mathematical Techniques in Data Science Clustering II

MATH 567: Mathematical Techniques in Data Science Clustering II This lecture is based on U. von Luxburg, A Tutorial on Spectral Clustering, Statistics and Computing, 17 (4), 2007. MATH 567: Mathematical Techniques in Data Science Clustering II Dominique Guillot Departments

More information

Machine Learning And Applications: Supervised Learning-SVM

Machine Learning And Applications: Supervised Learning-SVM Machine Learning And Applications: Supervised Learning-SVM Raphaël Bournhonesque École Normale Supérieure de Lyon, Lyon, France raphael.bournhonesque@ens-lyon.fr 1 Supervised vs unsupervised learning Machine

More information

9. Submodular function optimization

9. Submodular function optimization Submodular function maximization 9-9. Submodular function optimization Submodular function maximization Greedy algorithm for monotone case Influence maximization Greedy algorithm for non-monotone case

More information

Learning Spectral Graph Segmentation

Learning Spectral Graph Segmentation Learning Spectral Graph Segmentation AISTATS 2005 Timothée Cour Jianbo Shi Nicolas Gogin Computer and Information Science Department University of Pennsylvania Computer Science Ecole Polytechnique Graph-based

More information

Data Analysis and Manifold Learning Lecture 7: Spectral Clustering

Data Analysis and Manifold Learning Lecture 7: Spectral Clustering Data Analysis and Manifold Learning Lecture 7: Spectral Clustering Radu Horaud INRIA Grenoble Rhone-Alpes, France Radu.Horaud@inrialpes.fr http://perception.inrialpes.fr/ Outline of Lecture 7 What is spectral

More information

Learning with Submodular Functions: A Convex Optimization Perspective

Learning with Submodular Functions: A Convex Optimization Perspective Foundations and Trends R in Machine Learning Vol. 6, No. 2-3 (2013) 145 373 c 2013 F. Bach DOI: 10.1561/2200000039 Learning with Submodular Functions: A Convex Optimization Perspective Francis Bach INRIA

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

Diffuse interface methods on graphs: Data clustering and Gamma-limits

Diffuse interface methods on graphs: Data clustering and Gamma-limits Diffuse interface methods on graphs: Data clustering and Gamma-limits Yves van Gennip joint work with Andrea Bertozzi, Jeff Brantingham, Blake Hunter Department of Mathematics, UCLA Research made possible

More information

MLCC Clustering. Lorenzo Rosasco UNIGE-MIT-IIT

MLCC Clustering. Lorenzo Rosasco UNIGE-MIT-IIT MLCC 2018 - Clustering Lorenzo Rosasco UNIGE-MIT-IIT About this class We will consider an unsupervised setting, and in particular the problem of clustering unlabeled data into coherent groups. MLCC 2018

More information

Lecture 12 : Graph Laplacians and Cheeger s Inequality

Lecture 12 : Graph Laplacians and Cheeger s Inequality CPS290: Algorithmic Foundations of Data Science March 7, 2017 Lecture 12 : Graph Laplacians and Cheeger s Inequality Lecturer: Kamesh Munagala Scribe: Kamesh Munagala Graph Laplacian Maybe the most beautiful

More information

Graph Functional Methods for Climate Partitioning

Graph Functional Methods for Climate Partitioning Graph Functional Methods for Climate Partitioning Mathilde Mougeot - with D. Picard, V. Lefieux*, M. Marchand* Université Paris Diderot, France *Réseau Transport Electrique (RTE) Buenos Aires, 2015 Mathilde

More information

Spectral Clustering based on the graph p-laplacian

Spectral Clustering based on the graph p-laplacian Sectral Clustering based on the grah -Lalacian Thomas Bühler tb@cs.uni-sb.de Matthias Hein hein@cs.uni-sb.de Saarland University Comuter Science Deartment Camus E 663 Saarbrücken Germany Abstract We resent

More information

Graphs in Machine Learning

Graphs in Machine Learning Graphs in Machine Learning Michal Valko INRIA Lille - Nord Europe, France Partially based on material by: Ulrike von Luxburg, Gary Miller, Doyle & Schnell, Daniel Spielman January 27, 2015 MVA 2014/2015

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

Lecture: Local Spectral Methods (1 of 4)

Lecture: Local Spectral Methods (1 of 4) Stat260/CS294: Spectral Graph Methods Lecture 18-03/31/2015 Lecture: Local Spectral Methods (1 of 4) Lecturer: Michael Mahoney Scribe: Michael Mahoney Warning: these notes are still very rough. They provide

More information

Four graph partitioning algorithms. Fan Chung University of California, San Diego

Four graph partitioning algorithms. Fan Chung University of California, San Diego Four graph partitioning algorithms Fan Chung University of California, San Diego History of graph partitioning NP-hard approximation algorithms Spectral method, Fiedler 73, Folklore Multicommunity flow,

More information

Randomized Coordinate Descent Methods on Optimization Problems with Linearly Coupled Constraints

Randomized Coordinate Descent Methods on Optimization Problems with Linearly Coupled Constraints Randomized Coordinate Descent Methods on Optimization Problems with Linearly Coupled Constraints By I. Necoara, Y. Nesterov, and F. Glineur Lijun Xu Optimization Group Meeting November 27, 2012 Outline

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

Spectral clustering. Two ideal clusters, with two points each. Spectral clustering algorithms

Spectral clustering. Two ideal clusters, with two points each. Spectral clustering algorithms A simple example Two ideal clusters, with two points each Spectral clustering Lecture 2 Spectral clustering algorithms 4 2 3 A = Ideally permuted Ideal affinities 2 Indicator vectors Each cluster has an

More information

Graph Metrics and Dimension Reduction

Graph Metrics and Dimension Reduction Graph Metrics and Dimension Reduction Minh Tang 1 Michael Trosset 2 1 Applied Mathematics and Statistics The Johns Hopkins University 2 Department of Statistics Indiana University, Bloomington November

More information

Graph Clustering Algorithms

Graph Clustering Algorithms PhD Course on Graph Mining Algorithms, Università di Pisa February, 2018 Clustering: Intuition to Formalization Task Partition a graph into natural groups so that the nodes in the same cluster are more

More information

Cholesky Decomposition Rectification for Non-negative Matrix Factorization

Cholesky Decomposition Rectification for Non-negative Matrix Factorization Cholesky Decomposition Rectification for Non-negative Matrix Factorization Tetsuya Yoshida Graduate School of Information Science and Technology, Hokkaido University N-14 W-9, Sapporo 060-0814, Japan yoshida@meme.hokudai.ac.jp

More information

Global vs. Multiscale Approaches

Global vs. Multiscale Approaches Harmonic Analysis on Graphs Global vs. Multiscale Approaches Weizmann Institute of Science, Rehovot, Israel July 2011 Joint work with Matan Gavish (WIS/Stanford), Ronald Coifman (Yale), ICML 10' Challenge:

More information

EECS 275 Matrix Computation

EECS 275 Matrix Computation EECS 275 Matrix Computation Ming-Hsuan Yang Electrical Engineering and Computer Science University of California at Merced Merced, CA 95344 http://faculty.ucmerced.edu/mhyang Lecture 23 1 / 27 Overview

More information

arxiv: v1 [math.oc] 1 Jul 2016

arxiv: v1 [math.oc] 1 Jul 2016 Convergence Rate of Frank-Wolfe for Non-Convex Objectives Simon Lacoste-Julien INRIA - SIERRA team ENS, Paris June 8, 016 Abstract arxiv:1607.00345v1 [math.oc] 1 Jul 016 We give a simple proof that the

More information

U.C. Berkeley CS294: Spectral Methods and Expanders Handout 11 Luca Trevisan February 29, 2016

U.C. Berkeley CS294: Spectral Methods and Expanders Handout 11 Luca Trevisan February 29, 2016 U.C. Berkeley CS294: Spectral Methods and Expanders Handout Luca Trevisan February 29, 206 Lecture : ARV In which we introduce semi-definite programming and a semi-definite programming relaxation of sparsest

More information

Networks and Their Spectra

Networks and Their Spectra Networks and Their Spectra Victor Amelkin University of California, Santa Barbara Department of Computer Science victor@cs.ucsb.edu December 4, 2017 1 / 18 Introduction Networks (= graphs) are everywhere.

More information

Convex optimization COMS 4771

Convex optimization COMS 4771 Convex optimization COMS 4771 1. Recap: learning via optimization Soft-margin SVMs Soft-margin SVM optimization problem defined by training data: w R d λ 2 w 2 2 + 1 n n [ ] 1 y ix T i w. + 1 / 15 Soft-margin

More information

Linear Spectral Hashing

Linear Spectral Hashing Linear Spectral Hashing Zalán Bodó and Lehel Csató Babeş Bolyai University - Faculty of Mathematics and Computer Science Kogălniceanu 1., 484 Cluj-Napoca - Romania Abstract. assigns binary hash keys to

More information

Using Local Spectral Methods in Theory and in Practice

Using Local Spectral Methods in Theory and in Practice Using Local Spectral Methods in Theory and in Practice Michael W. Mahoney ICSI and Dept of Statistics, UC Berkeley ( For more info, see: http: // www. stat. berkeley. edu/ ~ mmahoney or Google on Michael

More information

Convex relaxation for Combinatorial Penalties

Convex relaxation for Combinatorial Penalties Convex relaxation for Combinatorial Penalties Guillaume Obozinski Equipe Imagine Laboratoire d Informatique Gaspard Monge Ecole des Ponts - ParisTech Joint work with Francis Bach Fête Parisienne in Computation,

More information

Space-Variant Computer Vision: A Graph Theoretic Approach

Space-Variant Computer Vision: A Graph Theoretic Approach p.1/65 Space-Variant Computer Vision: A Graph Theoretic Approach Leo Grady Cognitive and Neural Systems Boston University p.2/65 Outline of talk Space-variant vision - Why and how of graph theory Anisotropic

More information

Laplacian Eigenmaps for Dimensionality Reduction and Data Representation

Laplacian Eigenmaps for Dimensionality Reduction and Data Representation Introduction and Data Representation Mikhail Belkin & Partha Niyogi Department of Electrical Engieering University of Minnesota Mar 21, 2017 1/22 Outline Introduction 1 Introduction 2 3 4 Connections to

More information

Scalable Subspace Clustering

Scalable Subspace Clustering Scalable Subspace Clustering René Vidal Center for Imaging Science, Laboratory for Computational Sensing and Robotics, Institute for Computational Medicine, Department of Biomedical Engineering, Johns

More information

A convex relaxation for weakly supervised classifiers

A convex relaxation for weakly supervised classifiers A convex relaxation for weakly supervised classifiers Armand Joulin and Francis Bach SIERRA group INRIA -Ecole Normale Supérieure ICML 2012 Weakly supervised classification We adress the problem of weakly

More information

Graph-Laplacian PCA: Closed-form Solution and Robustness

Graph-Laplacian PCA: Closed-form Solution and Robustness 2013 IEEE Conference on Computer Vision and Pattern Recognition Graph-Laplacian PCA: Closed-form Solution and Robustness Bo Jiang a, Chris Ding b,a, Bin Luo a, Jin Tang a a School of Computer Science and

More information

Communities, Spectral Clustering, and Random Walks

Communities, Spectral Clustering, and Random Walks Communities, Spectral Clustering, and Random Walks David Bindel Department of Computer Science Cornell University 26 Sep 2011 20 21 19 16 22 28 17 18 29 26 27 30 23 1 25 5 8 24 2 4 14 3 9 13 15 11 10 12

More information

Quadratic reformulation techniques for 0-1 quadratic programs

Quadratic reformulation techniques for 0-1 quadratic programs OSE SEMINAR 2014 Quadratic reformulation techniques for 0-1 quadratic programs Ray Pörn CENTER OF EXCELLENCE IN OPTIMIZATION AND SYSTEMS ENGINEERING ÅBO AKADEMI UNIVERSITY ÅBO NOVEMBER 14th 2014 2 Structure

More information

25 : Graphical induced structured input/output models

25 : Graphical induced structured input/output models 10-708: Probabilistic Graphical Models 10-708, Spring 2016 25 : Graphical induced structured input/output models Lecturer: Eric P. Xing Scribes: Raied Aljadaany, Shi Zong, Chenchen Zhu Disclaimer: A large

More information

Local Lanczos Spectral Approximation for Community Detection

Local Lanczos Spectral Approximation for Community Detection Local Lanczos Spectral Approximation for Community Detection Pan Shi 1, Kun He 1( ), David Bindel 2, John E. Hopcroft 2 1 Huazhong University of Science and Technology, Wuhan, China {panshi,brooklet60}@hust.edu.cn

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

Graphs in Machine Learning

Graphs in Machine Learning Graphs in Machine Learning Michal Valko Inria Lille - Nord Europe, France TA: Pierre Perrault Partially based on material by: Mikhail Belkin, Jerry Zhu, Olivier Chapelle, Branislav Kveton October 30, 2017

More information

18.S096: Spectral Clustering and Cheeger s Inequality

18.S096: Spectral Clustering and Cheeger s Inequality 18.S096: Spectral Clustering and Cheeger s Inequality Topics in Mathematics of Data Science (Fall 015) Afonso S. Bandeira bandeira@mit.edu http://math.mit.edu/~bandeira October 6, 015 These are lecture

More information

Spectral Clustering on Handwritten Digits Database Mid-Year Pr

Spectral Clustering on Handwritten Digits Database Mid-Year Pr Spectral Clustering on Handwritten Digits Database Mid-Year Presentation Danielle dmiddle1@math.umd.edu Advisor: Kasso Okoudjou kasso@umd.edu Department of Mathematics University of Maryland- College Park

More information

Convex Optimization and Modeling

Convex Optimization and Modeling Convex Optimization and Modeling Introduction and a quick repetition of analysis/linear algebra First lecture, 12.04.2010 Jun.-Prof. Matthias Hein Organization of the lecture Advanced course, 2+2 hours,

More information

Spectral Graph Theory Lecture 2. The Laplacian. Daniel A. Spielman September 4, x T M x. ψ i = arg min

Spectral Graph Theory Lecture 2. The Laplacian. Daniel A. Spielman September 4, x T M x. ψ i = arg min Spectral Graph Theory Lecture 2 The Laplacian Daniel A. Spielman September 4, 2015 Disclaimer These notes are not necessarily an accurate representation of what happened in class. The notes written before

More information

Diffusion and random walks on graphs

Diffusion and random walks on graphs Diffusion and random walks on graphs Leonid E. Zhukov School of Data Analysis and Artificial Intelligence Department of Computer Science National Research University Higher School of Economics Structural

More information

Convex Methods for Transduction

Convex Methods for Transduction Convex Methods for Transduction Tijl De Bie ESAT-SCD/SISTA, K.U.Leuven Kasteelpark Arenberg 10 3001 Leuven, Belgium tijl.debie@esat.kuleuven.ac.be Nello Cristianini Department of Statistics, U.C.Davis

More information

Statistical Learning. Dong Liu. Dept. EEIS, USTC

Statistical Learning. Dong Liu. Dept. EEIS, USTC Statistical Learning Dong Liu Dept. EEIS, USTC Chapter 6. Unsupervised and Semi-Supervised Learning 1. Unsupervised learning 2. k-means 3. Gaussian mixture model 4. Other approaches to clustering 5. Principle

More information

Lecture Introduction. 2 Brief Recap of Lecture 10. CS-621 Theory Gems October 24, 2012

Lecture Introduction. 2 Brief Recap of Lecture 10. CS-621 Theory Gems October 24, 2012 CS-62 Theory Gems October 24, 202 Lecture Lecturer: Aleksander Mądry Scribes: Carsten Moldenhauer and Robin Scheibler Introduction In Lecture 0, we introduced a fundamental object of spectral graph theory:

More information

STAT 200C: High-dimensional Statistics

STAT 200C: High-dimensional Statistics STAT 200C: High-dimensional Statistics Arash A. Amini May 30, 2018 1 / 57 Table of Contents 1 Sparse linear models Basis Pursuit and restricted null space property Sufficient conditions for RNS 2 / 57

More information

Data Analysis and Manifold Learning Lecture 3: Graphs, Graph Matrices, and Graph Embeddings

Data Analysis and Manifold Learning Lecture 3: Graphs, Graph Matrices, and Graph Embeddings Data Analysis and Manifold Learning Lecture 3: Graphs, Graph Matrices, and Graph Embeddings Radu Horaud INRIA Grenoble Rhone-Alpes, France Radu.Horaud@inrialpes.fr http://perception.inrialpes.fr/ Outline

More information

Communities, Spectral Clustering, and Random Walks

Communities, Spectral Clustering, and Random Walks Communities, Spectral Clustering, and Random Walks David Bindel Department of Computer Science Cornell University 3 Jul 202 Spectral clustering recipe Ingredients:. A subspace basis with useful information

More information

Learning Spectral Clustering

Learning Spectral Clustering Learning Spectral Clustering Francis R. Bach Computer Science University of California Berkeley, CA 94720 fbach@cs.berkeley.edu Michael I. Jordan Computer Science and Statistics University of California

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

Semi-Supervised Learning by Multi-Manifold Separation

Semi-Supervised Learning by Multi-Manifold Separation Semi-Supervised Learning by Multi-Manifold Separation Xiaojin (Jerry) Zhu Department of Computer Sciences University of Wisconsin Madison Joint work with Andrew Goldberg, Zhiting Xu, Aarti Singh, and Rob

More information

An indicator for the number of clusters using a linear map to simplex structure

An indicator for the number of clusters using a linear map to simplex structure An indicator for the number of clusters using a linear map to simplex structure Marcus Weber, Wasinee Rungsarityotin, and Alexander Schliep Zuse Institute Berlin ZIB Takustraße 7, D-495 Berlin, Germany

More information

Statistical and Computational Analysis of Locality Preserving Projection

Statistical and Computational Analysis of Locality Preserving Projection Statistical and Computational Analysis of Locality Preserving Projection Xiaofei He xiaofei@cs.uchicago.edu Department of Computer Science, University of Chicago, 00 East 58th Street, Chicago, IL 60637

More information

Semi-Markov/Graph Cuts

Semi-Markov/Graph Cuts Semi-Markov/Graph Cuts Alireza Shafaei University of British Columbia August, 2015 1 / 30 A Quick Review For a general chain-structured UGM we have: n n p(x 1, x 2,..., x n ) φ i (x i ) φ i,i 1 (x i, x

More information

Mining of Massive Datasets Jure Leskovec, AnandRajaraman, Jeff Ullman Stanford University

Mining of Massive Datasets Jure Leskovec, AnandRajaraman, Jeff Ullman Stanford University Note to other teachers and users of these slides: We would be delighted if you found this our material useful in giving your own lectures. Feel free to use these slides verbatim, or to modify them to fit

More information

1 Matrix notation and preliminaries from spectral graph theory

1 Matrix notation and preliminaries from spectral graph theory Graph clustering (or community detection or graph partitioning) is one of the most studied problems in network analysis. One reason for this is that there are a variety of ways to define a cluster or community.

More information

Lecture 13: Spectral Graph Theory

Lecture 13: Spectral Graph Theory CSE 521: Design and Analysis of Algorithms I Winter 2017 Lecture 13: Spectral Graph Theory Lecturer: Shayan Oveis Gharan 11/14/18 Disclaimer: These notes have not been subjected to the usual scrutiny reserved

More information