1 “Expansion” in Power Law and Scale Free Graphs Milena Mihail Georgia Tech with Christos Gkantsidis, Christos Papadimitriou and Amin Saberi.

Slides:



Advertisements
Similar presentations
Routing Complexity of Faulty Networks Omer Angel Itai Benjamini Eran Ofek Udi Wieder The Weizmann Institute of Science.
Advertisements

Approximate Max-integral-flow/min-cut Theorems Kenji Obata UC Berkeley June 15, 2004.
Routing in Undirected Graphs with Constant Congestion Julia Chuzhoy Toyota Technological Institute at Chicago.
1 Complex Networks: Connectivity and Functionality Milena Mihail Georgia Tech.
Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.
Scale Free Networks.
Jennifer Tour Chayes Joint work with N. Berger, C. Borgs, A. Ganesh, A. Saberi, D. B. Wilson Controlling the Spread of Viruses on Power-Law Networks.
Poly-Logarithmic Approximation for EDP with Congestion 2
1 On the Eigenvalue Power Law Milena Mihail Georgia Tech Christos Papadimitriou U.C. Berkeley &
1 Algorithmic Performance in Power Law Graphs Milena Mihail Christos Gkantsidis Christos Papadimitriou Amin Saberi.
CSE 522 – Algorithmic and Economic Aspects of the Internet Instructors: Nicole Immorlica Mohammad Mahdian.
Topology Generation Suat Mercan. 2 Outline Motivation Topology Characterization Levels of Topology Modeling Techniques Types of Topology Generators.
Complex Networks Third Lecture TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: AA TexPoint fonts used in EMF. Read the.
On the Spread of Viruses on the Internet Noam Berger Joint work with C. Borgs, J.T. Chayes and A. Saberi.
Scale-free networks Péter Kómár Statistical physics seminar 07/10/2008.
Analysis of Network Diffusion and Distributed Network Algorithms Rajmohan Rajaraman Northeastern University, Boston May 2012 Chennai Network Optimization.
Small Worlds Presented by Geetha Akula For the Faculty of Department of Computer Science, CALSTATE LA. On 8 th June 07.
The Barabási-Albert [BA] model (1999) ER Model Look at the distribution of degrees ER ModelWS Model actorspower grid www The probability of finding a highly.
The structure of the Internet. How are routers connected? Why should we care? –While communication protocols will work correctly on ANY topology –….they.
Network Design IS250 Spring 2010 John Chuang. 2 Questions  What does the Internet look like? -Why do we care?  Are there any structural invariants?
Network Statistics Gesine Reinert. Yeast protein interactions.
Small World Networks Somsubhra Sharangi Computing Science, Simon Fraser University.
Advanced Topics in Data Mining Special focus: Social Networks.
SDSC, skitter (July 1998) A random graph model for massive graphs William Aiello Fan Chung Graham Lincoln Lu.
The structure of the Internet. How are routers connected? Why should we care? –While communication protocols will work correctly on ANY topology –….they.
Expanders Eliyahu Kiperwasser. What is it? Expanders are graphs with no small cuts. The later gives several unique traits to such graph, such as: – High.
CSE 522 – Algorithmic and Economic Aspects of the Internet Instructors: Nicole Immorlica Mohammad Mahdian.
1 Algorithms for Large Data Sets Ziv Bar-Yossef Lecture 7 May 14, 2006
Small Subgraphs in Random Graphs and the Power of Multiple Choices The Online Case Torsten Mütze, ETH Zürich Joint work with Reto Spöhel and Henning Thomas.
Summary from Previous Lecture Real networks: –AS-level N= 12709, M=27384 (Jan 02 data) route-views.oregon-ix.net, hhtp://abroude.ripe.net/ris/rawdata –
Computer Science 1 Web as a graph Anna Karpovsky.
Algorithmic Problems in the Internet Christos H. Papadimitriou
Large-scale organization of metabolic networks Jeong et al. CS 466 Saurabh Sinha.
The Erdös-Rényi models
Optimization Based Modeling of Social Network Yong-Yeol Ahn, Hawoong Jeong.
1 Milena Mihail Web Science Tea Feb 29, 08 Discussion Topic:
1 Milena Mihail Georgia Tech. with Stephen Young, Giorgos Amanatidis, Bradley Green Flexible Models for Complex Networks.
1 Algorithmic Performance in Complex Networks Milena Mihail Georgia Tech.
1 Spectral Analysis of Power-Law Graphs and its Application to Internet Topologies Milena Mihail Georgia Tech.
Flexible Graph Models for Complex Networks Complex Networks: Internet and its applications: WWW, content sharing, social online Other: further social,
Clustering of protein networks: Graph theory and terminology Scale-free architecture Modularity Robustness Reading: Barabasi and Oltvai 2004, Milo et al.
1 Burning a graph as a model of social contagion Anthony Bonato Ryerson University Institute of Software Chinese Academy of Sciences.
1 “Erdos and the Internet” Milena Mihail Georgia Tech. The Internet is a remarkable phenomenon that involves graph theory in a natural way and gives rise.
Jure Leskovec Computer Science Department Cornell University / Stanford University Joint work with: Jon Kleinberg (Cornell), Christos.
Emergence of Scaling and Assortative Mixing by Altruism Li Ping The Hong Kong PolyU
1 Milena Mihail Georgia Tech. “with network elements maintaining characteristic profiles” Models and Algorithms for Complex Networks “with categorical.
Percolation Processes Rajmohan Rajaraman Northeastern University, Boston May 2012 Chennai Network Optimization WorkshopPercolation Processes1.
On-line Social Networks - Anthony Bonato 1 Dynamic Models of On-Line Social Networks Anthony Bonato Ryerson University WAW’2009 February 13, 2009 nt.
Random Dot Product Graphs Ed Scheinerman Applied Mathematics & Statistics Johns Hopkins University IPAM Intelligent Extraction of Information from Graphs.
Lecture 10: Network models CS 765: Complex Networks Slides are modified from Networks: Theory and Application by Lada Adamic.
KPS 2007 (April 19, 2007) On spectral density of scale-free networks Doochul Kim (Department of Physics and Astronomy, Seoul National University) Collaborators:
Miniconference on the Mathematics of Computation
1 How to burn a graph Anthony Bonato Ryerson University GRASCan 2015.
Scaling Properties of the Internet Graph Aditya Akella, CMU With Shuchi Chawla, Arvind Kannan and Srinivasan Seshan PODC 2003.
Graph Partitioning using Single Commodity Flows
Performance Evaluation Lecture 1: Complex Networks Giovanni Neglia INRIA – EPI Maestro 10 December 2012.
Anonymous communication over social networks Shishir Nagaraja and Ross Anderson Security Group Computer Laboratory.
1 Milena Mihail Georgia Tech. Algorithmic Performance in Complex Networks.
Scaling Properties of the Internet Graph Aditya Akella With Shuchi Chawla, Arvind Kannan and Srinivasan Seshan PODC 2003.
Algorithms and Computational Biology Lab, Department of Computer Science and & Information Engineering, National Taiwan University, Taiwan Network Biology.
1 “Hybrid Search Schemes for Unstructured Peer- to-Peer Networks” “Random Walks in Peer-to-Peer Networks” Christos Gkantsidis, Milena Mihail, Amin Saberi.
Algorithmic Game Theory and Internet Computing
Random Walk for Similarity Testing in Complex Networks
Sequential Algorithms for Generating Random Graphs
Christos H. Papadimitriou UC Berkeley christos
Complex Networks: Connectivity and Functionality
Random Graph Models of large networks
On the effect of randomness on planted 3-coloring models
Problem Solving 4.
Modelling and Searching Networks Lecture 2 – Complex Networks
Presentation transcript:

1 “Expansion” in Power Law and Scale Free Graphs Milena Mihail Georgia Tech with Christos Gkantsidis, Christos Papadimitriou and Amin Saberi

2 Graphs with Skewed Degree Sequences Communication Networks This Talk: Algorithmic Issues, “Expansion”, spectral gap determine performace of key algorithms. Metabolic Networks

3 How does Congestion Scale? Sprint AT&T Demand: n 2, uniform. What is load of max congested link, in optimal routing ? ISPs: K Routers: K WWW: 500K-3B P2P: hundred Ks

4 CAIDA Degrees not Concentrated around mean E[degree]~3 Real Internet Topologies Not Erdos-Renyi

5 Degree-Frequency Power Law frequenc y E[d] = const., but No sharp concentration degree E[d] = const., but No sharp concentration Erdos-Renyi sharp concentration Models by Kumar et al 00, x Bollobas et al 01, x Fabrikant et al 02

6 Power Laws Degree-Frequency Rank-Degree Eigenvalues (Adjacency Matrix) [WWW: Kumar et al 99, Barabasi-Albert 99] [Interdomain Routing: Faloutsos et al 99]

7 Models for Power Law Graphs EVOLUTIONARY Macroscopic : Growth & Preferential Attachment Simon 55, Barabasi-Albert 99, Kumar et al 00, Bollobas-Riordan 01. Microscopic : Growth & Multiobjective Optimization, QoS vs Cost Fabrikant-Koutsoupias-Papadimitriou 02. STRUCTURAL (aka CONFIGURATIONAL) “Random” graph with “power law” degree sequence.

8 Structural Random Graph Model Given Choose random perfect matching over Molley&Reed 95-98, Aiello,Chung,Lu 00, Tagmunarunkit et al 02 minivertices

9 Congestion in the “Core” Theorem [Gkantsidis,MM, Saberi 02]: For a random graph arising from degree sequence O(n ½ ) ≥ d 1 ≥d 2 ≥…≥d n ≥3 there is a flow that routes demand d i * d j between all vertices i and j with max link congestion O(n log 2 n) almost surely.

10 Proof : Step 1 : Approximation algorithms for multicommodity flow reduce congestion to conductance (special case of sparsest cut). Step 2 : Bound conductance - MAIN LEMMA.

11 By Maximum multi-commodity flow, [Leighton & Rao 95] Proof, Step 1 : Reduce to Conductance

12 Proof, Step 2 : Main Lemma [Gkantsidis,MM, Saberi 02]:

13 Proof of MAIN LEMMA:

14 Proof of MAIN LEMMA:

15 Proof of MAIN LEMMA: Stirling

16 Proof of MAIN LEMMA: ignore Stirling BIGSMALL

17 In an Evolutionary Model ? Growth with Pref. Attachment One vertex at a time New vertex attaches to d existing vertices

18 Reduction to Random Matching [Bollobas & Riordan 01] t=2t=1t=3 t=4 t=5

19 Reduction to Random Matching [Bollobas & Riordan 01] t=2t=1t=3 t=4 t=5

20 Reduction to Random Matching [Bollobas & Riordan 01]

21 In an Evolutionary Model ? Growth with Pref. Attachment Theorem [MM, Saberi 02]: For a graph grown with preferential attachment with d ≥ 3 there is a flow that routes demand d i * d j between all vertices i and j with max link congestion O(n log n) almost surely. Main Lemma: almost surely. Open Question: Analyze a graph grown one vertex or edge at a time, where with probability a new vertex comes and attaches preferentially and with probability a new edge grows preferentially between existing vertices.

22 Spectral Implication Theorem: Eigenvalue separation for stochastic normalization of adjacency matrix [Alon 85, Jerrum&Sinclair 88]

23 Spectra of “Real” Internet

24 Spectral Implications Theorem: Eigenvalue separation for stochastic normalization of adjacency matrix [Alon 85, Jerrum&Sinclair 88] Using matrix perturbation [Courant-Fisher theorem] in a sparse random graph model. Rank-Degree Eigenvalues (Adjacency Matrix) On the eigenvalue Power Law [M.M. & Papadimitriou 02]

25 Theorem : Ffor large enough Wwith probability at least [M.M. & Papadimitriou 02]

26 Proof : Step 1. Decomposition Vertex Disjoint StarsLR-extra RR LL LR =-

27 Proof: Step 2: Vertex Disjoint Stars Degrees of each Vertex Disjoint Stars Sharply Concentrated around its Mean d_i Hence Principal Eigenvalue Sharply Concentrated around

28 Proof: Step 3: LL, RR, LR-extra LR-extra has max degree LL has edges RR has max degree

29 Proof: Step 3: LL, RR, LR-extra LR-extra has max degree RR has max degree LL has edges

30 Proof: Step 4: Matrix Perturbation Theory Vertex Disjoint Stars have principal eigenvalues All other parts have max eigenvalue QED

31 Implication for Info Retrieval Spectral filtering, without preprocessing, reveals only the large degrees. Term-Norm Distribution Problem :

32 Implication for Info Retrieval Term-Norm Distribution Problem : Spectral filtering, without preprocessing, reveals only the large degrees. Local information. No “latent semantics”.

33 Implication for Information Retrieval Application specific preprocessing (normalization of degrees) reveals clusters: WWW: related to searching, Kleinberg 97 IR, collaborative filtering, … Internet: related to congestion, Gkantsidis et al 02 Open : Formalize “preprocessing”. Term-Norm Distribution Problem :

34 Routing Integral paths? Short paths? Reliability? Cover time? Related to Crawling Hitting time? Related to Searching Planted model? Information Retrieval Further Directions: (Experimental work: Gkantsidis, MM, Zegura 02.) Generalize theory of Regular Expanders Peleg&Upfal’88 … Broder,Frieze&Upfal’01 Kleinberg&Rubinfeld’97

35 Metabolic Networks Statistics of fixed size subgraphs? Related to “motifs” in metabolic networks. Model (explain) heavy tailed statistics in noncoding part of DNA? Related to stages of species evolution.

36 Evaluation of Synthetic Topology Generators Core of the Network Entire Topology