PhD Thesis Proposal Evolution in Social Networks Candidate Giulio Rossetti Supervisor Dino Pedreschi Supervisor Fosca Giannotti Pisa, Computer Science.

Slides:



Advertisements
Similar presentations
Classes will begin shortly. Networks, Complexity and Economic Development Class 5: Network Dynamics.
Advertisements

Peer-to-Peer and Social Networks Power law graphs Small world graphs.
Complex Networks: Complex Networks: Structures and Dynamics Changsong Zhou AGNLD, Institute für Physik Universität Potsdam.
Complex Networks Advanced Computer Networks: Part1.
1 Dynamics of Real-world Networks Jure Leskovec Machine Learning Department Carnegie Mellon University
Emergence of Scaling in Random Networks Albert-Laszlo Barabsi & Reka Albert.
Analysis and Modeling of Social Networks Foudalis Ilias.
Lecture 21 Network evolution Slides are modified from Jurij Leskovec, Jon Kleinberg and Christos Faloutsos.
VL Netzwerke, WS 2007/08 Edda Klipp 1 Max Planck Institute Molecular Genetics Humboldt University Berlin Theoretical Biophysics Networks in Metabolism.
Information Networks Small World Networks Lecture 5.
1 Evolution of Networks Notes from Lectures of J.Mendes CNR, Pisa, Italy, December 2007 Eva Jaho Advanced Networking Research Group National and Kapodistrian.
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.
Masters Thesis Defense Amit Karandikar Advisor: Dr. Anupam Joshi Committee: Dr. Finin, Dr. Yesha, Dr. Oates Date: 1 st May 2007 Time: 9:30 am Place: ITE.
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.
Networks FIAS Summer School 6th August 2008 Complex Networks 1.
Network Design IS250 Spring 2010 John Chuang. 2 Questions  What does the Internet look like? -Why do we care?  Are there any structural invariants?
Web as Graph – Empirical Studies The Structure and Dynamics of Networks.
CS Lecture 6 Generative Graph Models Part II.
Sampling from Large Graphs. Motivation Our purpose is to analyze and model social networks –An online social network graph is composed of millions of.
INFERRING NETWORKS OF DIFFUSION AND INFLUENCE Presented by Alicia Frame Paper by Manuel Gomez-Rodriguez, Jure Leskovec, and Andreas Kraus.
Advanced Topics in Data Mining Special focus: Social Networks.
Social Networks: Advertising, Pricing and All That Zvi Topol & Itai Yarom.
User Interactions in OSNs Evangelia Skiani. Do you have a Facebook account? Why? How likely to know ALL your friends? Why confirm requests? Why not remove.
1 Algorithms for Large Data Sets Ziv Bar-Yossef Lecture 7 May 14, 2006
1 IEEE Intelligent Systems, Special Issue on Social Learning, 2010.
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 –
Peer-to-Peer and Social Networks Random Graphs. Random graphs E RDÖS -R ENYI MODEL One of several models … Presents a theory of how social webs are formed.
Models of Influence in Online Social Networks
Large-scale organization of metabolic networks Jeong et al. CS 466 Saurabh Sinha.
(Social) Networks Analysis III Prof. Dr. Daning Hu Department of Informatics University of Zurich Oct 16th, 2012.
Topic 13 Network Models Credits: C. Faloutsos and J. Leskovec Tutorial
Graph Theory in 50 minutes. This Graph has 6 nodes (also called vertices) and 7 edges (also called links)
Survey on Evolving Graphs Research Speaker: Chenghui Ren Supervisors: Prof. Ben Kao, Prof. David Cheung 1.
DEMON A Local-first Discovery Method For Overlapping Communities Giulio Rossetti 2,1,Michele Coscia 3, Fosca Giannotti 2, Dino Pedreschi 2,1 1 Computer.
Jure Leskovec PhD: Machine Learning Department, CMU Now: Computer Science Department, Stanford University.
Clustering of protein networks: Graph theory and terminology Scale-free architecture Modularity Robustness Reading: Barabasi and Oltvai 2004, Milo et al.
COM1721: Freshman Honors Seminar A Random Walk Through Computing Lecture 2: Structure of the Web October 1, 2002.
Complex Networks First Lecture TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: AA TexPoint fonts used in EMF. Read the.
A Graph-based Friend Recommendation System Using Genetic Algorithm
Gennaro Cordasco - How Much Independent Should Individual Contacts be to Form a Small-World? - 19/12/2006 How Much Independent Should Individual Contacts.
Social Network Analysis Prof. Dr. Daning Hu Department of Informatics University of Zurich Mar 5th, 2013.
A Local Seed Selection Algorithm for Overlapping Community Detection 1 A Local Seed Selection Algorithm for Overlapping Community Detection Farnaz Moradi,
Complex Contagions Models in Opportunistic Mobile Social Networks Yunsheng Wang Dept. of Computer Science, Kettering University Jie Wu Dept. of Computer.
Class 9: Barabasi-Albert Model-Part I
Lecture 10: Network models CS 765: Complex Networks Slides are modified from Networks: Theory and Application by Lada Adamic.
Social Network Analysis Aerospace Data Mining Center
Most of contents are provided by the website Network Models TJTSD66: Advanced Topics in Social Media (Social.
+ Big Data, Network Analysis Week How is date being used Predict Presidential Election - Nate Silver –
Spontaneous Formation of Dynamical Groups in an Adaptive Networked System Li Menghui, Guan Shuguang, Lai Choy-Heng Temasek Laboratories National University.
What Is A Network? (and why do we care?). An Introduction to Network Theory | Kyle Findlay | SAMRA 2010 | 2 “A collection of objects (nodes) connected.
Du, Faloutsos, Wang, Akoglu Large Human Communication Networks Patterns and a Utility-Driven Generator Nan Du 1,2, Christos Faloutsos 2, Bai Wang 1, Leman.
COMMUNITY DISCOVERY PART 1: A (BRIEF) INTRODUCTION Giulio Rossetti WMA - 4 May 2015.
RTM: Laws and a Recursive Generator for Weighted Time-Evolving Graphs Leman Akoglu, Mary McGlohon, Christos Faloutsos Carnegie Mellon University School.
Social Networking: Large scale Networks
Quantification in Social Networks Letizia Milli, Anna Monreale, Giulio Rossetti, Dino Pedreschi, Fosca Giannotti, Fabrizio Sebastiani Computer Science.
Mining information from social media
1 Friends and Neighbors on the Web Presentation for Web Information Retrieval Bruno Lepri.
March 3, 2009 Network Analysis Valerie Cardenas Nicolson Assistant Adjunct Professor Department of Radiology and Biomedical Imaging.
Netlogo demo. Complexity and Networks Melanie Mitchell Portland State University and Santa Fe Institute.
Multidimensional Network Analysis Foundations of multidimensional Network Analysis, Berlingerio, Coscia, Giannotti, Monreale, Pedreschi. WWW Journal 2012.
Structures of Networks
Wenyu Zhang From Social Network Group
DEMON A Local-first Discovery Method For Overlapping Communities
Social and Information Network Analysis: Review of Key Concepts
The Watts-Strogatz model
Peer-to-Peer and Social Networks Fall 2017
Topology and Dynamics of Complex Networks
Graph and Tensor Mining for fun and profit
Lecture 21 Network evolution
Presentation transcript:

PhD Thesis Proposal Evolution in Social Networks Candidate Giulio Rossetti Supervisor Dino Pedreschi Supervisor Fosca Giannotti Pisa, Computer Science Dept. 19th December 2012

Presentation Outline I° State of the Art Network Science Multidimensionality Time and Evolution Open Problems Structural issues Topological issues Evolutive issues Proposal What I've done so far… … and what’s my roadmap

Network Science Networks are everywhere – Social interactions, transportation, technological (WWW), biological, economics… Study their structures is an interesting task for a wide set of fields – Computer Science, Physics, Sociology… Through the analysis of networks is possible the discovery (and understanding) of hidden phenomena otherwise difficult to describe – Graph Theory In this technological era, huge networked datasets are easy to collect – Online Social Networks (OSNs), co-authorship networks, product nteworks… State of the Art

How can we describe real world phenomena? Observation The analysis of real world phenomena unveil peculiar charateristic that can be used to describe various typology of networks. Problem Can we build models that generate synthetic networks accordingly? Proposed models – Random Graphs – Small World – Scale Free – Forest Fire State of the Art

Random Graphs & Small World Rarely natural phenomena can be described by regularly connected graphs (such as lattices) and, at the same time, they do not show complete randomness. – [1967] Stanley Milgram: Six degree of separation – [1998] Watts & Strogatz: Small World State of the Art

Scale Free & Forest Fire Scale Free Model [1999] Barabàsi-Albert: Preferential Attachment – “Richer nodes became richer” – Power Law degree distribution State of the Art Forest Fire model [2005] Leskovetc: Community Structure – Shrinking diameter – Power Law degree distribution

Multidimensionality Social Networks are often multidimensional: – Multiple relations could occour among a couple of nodes at the same time Study the interplay among dimensions provides useful informations – Dependence among dimensions, redondance, exclusivity State of the Art [2011] Berlingerio et. al: Foundations of multidimensional network analysis

New measures are needed State of the Art A new set of Neighbor functions Functions able to analyze dimensions and their relevance

Internetworking Scenario & Hub Analysis In this multidimensional setting Hubs needs to be redefined… – [2011] Berlingerio et al.: The pursuit of hubbines State of the Art Data from Online Social Networks could be merged in order to obtain “Social Internetworking Scenarios” (SISs)

Time and Evolution State of the Art Networks are not static objects – Nodes and edges appear and disappear over time – Interactions among the same two nodes could take place multiple times Even when the structure could be assumed frozen in time – Diffusion processes – Information spreading – …

Link Prediction & Information Propagation Link Prediction Identify the rising of new edges observing the actual topology of a network is a complex task – [2003] Kleinberg et al.: The link prediction problem for social networks – [2011] Wang et al.: Human mobility, social ties, and link prediction. State of the Art Information Propagation How rapidly actions became viral? Every action lead necessary to a cascading effect? – [2008] Ma et al.: Mining social networks using heat diffusion processes for marketing candidates selection – [2009] Berlingerio et al.: Mining the information propagation in a network

Presentation Outline II° State of the Art Network Science Multidimensionality Time and Evolution Open Problems Structural issues Topological issues Evolutive issues Proposal What I've done so far… … and what’s my roadmap

Open Problems StructuralTopologicalEvolutive Node & Link Analysis Ego-networks, Communities, … Link Prediction, Community Evolution, … Open Problems

Structural issues “The structure of a network define how the parts of it (nodes and edges) relate to each other, how it is assembled” Open Problems The first step needed to understand how a network evolve is to study how its basic components (nodes and edges) can be characterized – Link analysis – Node analysis

Multidimensional Tie strength & Node Ranking Tie Strength Not all the social connections have the same degree of intimacy – [1973] Granovetter: The strength of weak ties What is a “strong” tie in a multidimensional networks? – Can we exploit this information? Open Problems Node Ranking Can we rank nodes in a multidimensional social graph? – For a person is better to have higher “skills” or being well connected within the network?

Topological issues “Network Topology express the particular shape the nodes are arranged” Open Problems 1.Nodes and edges in a network are organized in complex topologies (i.e. ego- networks, communities…); 2.Since in real networks entities are (strictly) related to each other we need to take care of topology when we want to model their evolution. – Community Discovery – …

Community Discovery The aim of CD algorithms is to identify communities hidden into complex network structure Why Community Discovery? “Cluster” homogeneous nodes relying on topological information (Clustering networked entities) Issues: – Each algorithm model different properties of real world communities – Found an acceptable compromise between number of communities and their sizes Context Dependent – Multidimensonal communities? Open Problems

Evolutive issues “Evolution is the process of change in all forms of networks over generations, and evolutionary analysis is the study of how evolution occurs.” Open Problems 1.Networks are dynamic objects; 2.Structure and Topology changes as time goes by; 3.Novel problems arise: – Community Evolution – Link Prediction – …

Community Evolution & Link Prediction Community Evolution Social groups change over time, can we identify how? – Community Kernel – Community Life-Cicle – Identify nodes roles – … Open Problems Link Prediction Social connections arise\renovate\disappear over time, can we predict how? – Temporal Link Prediction – Network Archeology – Multidimensional Link Prediction – …

Presentation Outline III° State of the Art Network Science Multidimensionality Time and Evolution Open Problems Structural issues Topological issues Evolutive issues Proposal What I've done so far… … and what’s my roadmap

Thesis Proposal Title: Evolution in Social Networks Aim: Exploit structural and topological patterns in order to model and understand evolutionary behavior of social networks. Proposal Understandig properties of Network’s Structure Topological Analysis Evolution in Social Networks

What I've done so far… Evolutive Analysis Giulio Rossetti, Michele Berlingerio, Fosca Giannotti, “Scalable Link Prediction on Multidimensional Networks”, ICDM DaMNET 2011 IEEE Giulio Rossetti, Michele Berlingerio and Fosca Giannotti, “Link Prediction su Reti Multidimensionali”, SEBD 2011 Topological Analysis Michele Coscia, Giulio Rossetti, Fosca Giannotti, Dino Pedreschi “DEMON: a Local-First Discovery Method for Overlapping Communities”, KDD 2012 ACM SIGKDD Structural Analysis Giulio Rossetti, Luca Pappalardo, Dino Pedreschi “How well do we know each other?: Detecting tie strength in multidimensional social networks”, ASONAM CSNA 2012 IEEE Proposal

…and what’s my roadmap Evolutive Analysis Temporal Link Prediction Community Life- Cycle Diffusion of informations Trust\Privacy feedbacks … Topological Analysis Kernel Communities Roles of nodes within Communities Multidimensional Community Discovery … Structural Analysis Strength of complex structures (ego-networks, communities…) Multidimensional Multilabel Node Ranking …. Proposal

Conclusion Evolutionary analysis of social networks represent a set of interdisciplinary problems. Given the availability of Big Datasets this kind of studies are today feasible on reasonable samples of individuals Proposal

Essential Bibliography Proposal [1] S. Milgram. The small world problem. Psychol, pages 60–67, 1967 [2] M. S. Granovetter. The strength of weak ties. America Journal of Sociology, [3] D. J. Watts and S. H. Strogatz. Collective dynamics of ’small-world’ networks. Nature, 1998 [4] A. L. Barabasi and R. Albert. Emergence of scaling in random networks. In Science, 1999 [5] D. Liben-Nowell and J. Kleinberg. The link prediction problem for social networks. In CIKM, 2003 [6] J. Leskovec, J. M. Kleinberg, and C. Faloutsos. Graphs over time: densification laws, shrinking diameters and possible explanations. In KDD, 2005 [7] H. Ma, H. Yang, M. R. Lyu, and I. King. Mining social networks using heat diffusion processes for marketing candidates selection. In CIKM, 2008 [8] M. Berlingerio, M. Coscia, and F. Giannotti. Mining the information propagation in a network. In SEBD, 2009 [9] M. Berlingerio, M. Coscia, F. Giannotti, A. Monreale, and D. Pedreschi. Foundations of multidimensional network analysis. In ASONAM, 2011 [10] D. Wang, D. Pedreschi, C. Song, F. Giannotti, and A. L. Barabàsi. Human mobility, social ties, and link prediction. In KDD, 2011 [11] M. Coscia, G. Rossetti, D. Pedreschi and F. Giannotti. DEMON: a Loca-first discovery method for overlapping communities, In KDD, 2012

First year original results Appendix How well do we know each other? Detecting tie strength in multidimensional social networks DEMON a local-first approach to community discovery Scalable Link Predition on Multidimensional Networks

Multidimensional Tie Strength Appendix Problem: Given a SIS scenario, are we able to identify strong ties? What they represent? Proposed Metric: Connection Redundancy: Node Similarity and Interactions: Results:

Community Discovery Appendix Problem: Given a complex network are we able to identify meaningful communities form a social point of view? Proposed Algorithm(s): – DEMON – Hierarchical DEMON

Community Discovery (2) Appendix DEMON algorithm: For each node n: Extract the Ego Network of n Remove n from the Ego Network Perform a Label Propagation 1 Insert n in each community found Update the raw community set C For each raw community c in C Merge with “similar” ones in the set (given a threshold) (i.e. merge iff at most the ε% of the smaller one is not included in the bigger one)

Appendix Community Discovery (3) Hierarchical DEMON algorithm: HDemon(Graph G) Cc = connectedComponent(G) C = ExtractCommunities(G) while (|C|>Cc) For c in C: N <- N ∪ make_node(c) For (n,m) in N: If (n share nodes with m): E <- E ∪ (n,m) C <- ExtractCommunities(new Graph(N,E)) ExtractCommunities(Graph G) Egos <- EgoNetworks(G) for e in Egos: C = C ∪ LabelPropagation(e) return C

Multidimensional Link Prediction Problem: Can we exploit multidimensional and temporal information to increase the precision of well known unsupervised link prediction approaches? Proposed metrics and results: Appendix