ECS 289 / MAE 298, Lecture 16 May 22, Diffusion, Cascades and Influence, Part II

Size: px
Start display at page:

Download "ECS 289 / MAE 298, Lecture 16 May 22, Diffusion, Cascades and Influence, Part II"

Transcription

1 ECS 289 / MAE 298, Lecture 16 May 22, 2014 Diffusion, Cascades and Influence, Part II

2 Announcements Homeworks HW3 and HW3b now due Tues May 27 HW3a (project progress report) due Fri May 30 NetSci 2014 attendance and volunteering Visiting Scholar, Trivik Verma, will coordinate. Expect an next week with final details. Tasks: Pass out name tags (starting at 8:15am); Pass around microphone for audience questions; etc Mathematics and Industry Seminar today, 5:10pm Project presentations

3 Mathematics and Industry Seminar today, 5:10pm Two data scientists from Yelp are speaking on their work on A/B testing and the multi-armed bandit problem; Scott Clark and Ben Goldenberg The talk is at 5:10 pm in MSB 2112 Optimally Learning for Fun and Profit Running experiments over user traffic to test feature improvements has become common over the last decade. Segmenting your traffic and showing different code paths or parameters to different buckets of users, A/B testing, allows one to iteratively improve features with statistical confidence. This problem can be aided with the use of multi-armed bandits, which optimally trade off exploration (gaining new knowledge about the system) and exploitation (gain received from the current knowledge of the system) to efficiently control these experiments. When used in experiments involving parameter search this allows for quickly determining which values are viable in the parameter space (exploration) and funneling more traffic towards them (exploitation). This method can be extended by applying optimal experimental design, using Bayesian global optimization, to suggest optimal new values to test as old ones are deemed non-viable, using information about the space gained from the running experiment. In this talk I will give a short overview of the multi-armed bandit problem and show how we are using optimal experimental design to extend the traditional multi-armed bandit approach to A/B testing and targeting.

4 13 Class projects 1. Optimal Location For a New Business Harika Sabella, Mohammad Adil, Sugeerth Murugesan 2. Contagion in Social Resistace Networks Haochen Wu, Felipe Aviles Lucero, Jaime Jackson, Aleksander Zujev, Elizabeth Zarrindast 3. Community Detection and Resilience G. Badeau, W. Cuello, R. Starr 4. Open-Source Collaboration and Following Grace Benefield, Casey Casalnuovo 5. The Effect of Signed Network Topology on Binary Neurons Tom Chartrand, Alec Boyd 6. Who Triggers a Weibo Event? Xiaotao Feng, Jiahui Guan 7. Hierarchical Structure of Wikipedia Brian Weston, Jay Gokhale, Ali Emara 8. Characterizing fmri-based functional networks during attention-demanding tasks at rest Ben Kubit, Saeedeh Komijani 9. Maternal/fetal effects of Bis(2-ethylhexyl) phthalate on hormone regulation during gestation in Zinc deficient rats: a network analysis approach Carlos Ruvalcaba, Trevor Ramsay, and Heidi Kucera

5 13 Class projects, cont 10. California Winegrower Social Networks: Environmental Certifications and Practices Michael Levy and Ryan Parker 11. Opinion Dynamics with reluctant agents Hoi-To Wai, Christopher Patton 12. Extraction and Reasoning about Roles in Power Networks Andrew Smith 13. Yelp Data challenge - Communities for A Better Yelp Edmund Yan, Leyuan Wang Volunteered to speak Thurs May mins total for each presentation (including Q&A) 5-6 presentations on Thurs May 29 (90 min slot) 7-8 presentations on Thurs June 12 (120 min slot)

6 Diffusion and cascades in networks (Nodes in one of two states) Viruses (human and computer) contact processes epidemic thresholds Adoption of new technologies Winner take all Benefit of first to market Benefit of second to market Political or social beliefs and societal norms A long history of study, now trying to add impact of underlying network structure.

7 Diffusion, Cascade behaviors, and influential nodes Part I: Ensemble models (last time) Generating functions for giant components

8 Generating functions Start from a simple probability density function, P k Can manipulate P k to build generating functions (G.F.) for more complicated distributions Edge following G.F. Component at end of random edge G.F. Component for randomly selected node G.F. Take derivatives of G.F. to calculate the moments of those distributions s, expected size of component for randomly chosen node (first moment of that G.F.) An algebraic expression. First instance that s is emergence of a giant component.

9 Generating function approach to adoption of new behavior: Watts PNAS (2002) All nodes, except one, start in inactive state, { 1} Fractional threshold model (Φ i ). Node activated once a fraction of it s neighbors Φ i are active. A vulnerable node is one that needs only a single neighbor to be active before it flips (i.e., Φ i 1/k). Use generating functions to calculate the expected size of clusters of vulnerable nodes. A Global cascade corresponds to a giant component Results Heterogeneity in thresholds (Φ i ) enhances global cascades. Heterogeneity of degree (P k ) suppresses global cascades.

10 Diffusion, Cascade behaviors, and influential nodes Part I: Ensemble models (last time) Master equation approach: Pastor-Satorras and Vespignani Contact process, epidemic spreading Probability of becoming activated is proportional to the number of active neighbors. Results Heterogeneity of degree (P k ) enhances global spreading. For PLRG with 2 < γ < 3 the epidemic threshold λ c 0.

11 SIS disease dynamics ρ k (t) is density of infected nodes of degree k at time t. (Hence [1 ρ k (t)] is probability a node of degree k is NOT infected.) λ = β/γ, the effective spreading rate. Set γ = 1. (Recall β is infection rate, γ is recovery.) The time evolution (a master equation ): dρ k (t) dt = ρ k (t) + λk [1 ρ k (t)] Θ(ρ(t)) First term: nodes recover with unit rate (γ = 1) Second term: Infection rate λ, times number of neighbors k, times prob node of degree k is healthy, times prob of being connected to an infected node Θ(ρ(t)).

12 Steady state of master eqn, dρ k dt = 0 implies: ρ k = λ k Θ 1 + λ Θ Inserting into expression for Θ: Θ = 1 k k k p k λ k Θ 1 + λ Θ (Note Θ = 0 always satisfies, but is quite dull!... ρ k = 0)

13 Searching for more solutions to last equation, in interval 0 < Θ 1 A 1 B 1 y 2 (Θ) y 2 (Θ) Slope < 1 1 Θ Θ 1 Θ If the slope of Θ > 1 at the origin, there will be a non-trivial solution in the interval 0 < Θ 1.

14 Searching for more solutions to last equation, in interval 0 < Θ 1 Taking derivative w.r.t. Θ of both sides of last equation: d dθ [ 1 k p k k solving this: 1 k k ] λ k Θ 1 + λ Θ Θ=0 k p k λ c k = k λ c = k k 2 k 2 = 1, at λ = λ c k λ c = 1 If k 2 but k finite, then λ c 0. Unlike a threshold model, they find node heterogeniety greatly enhances onset of global cascade

15 Diffusion, Cascade behaviors, and influential nodes Part II: Contact processes with individual node preferences Long history of empirical / qualitative study in the social sciences (Peyton Young, Granovetter, Martin Nowak...; diffusion of innovation; societal norms) Recent theorems: network coordination games (bigger payout if connected nodes in the same state) (Kleinberg, Kempe, Tardos, Dodds, Watts, Domingos) Finding the influential set of nodes, or the k most influential Often NP-hard and not amenable to approximation algorithms Key distinction: Thresholds of activation (leads to unpredictable behaviors) Diminishing returns (submodular functions nicer)

16 Part II. Network Coordination Games The most basic model: Reviewed in Kleinberg Cascading Behavior in Networks: Algorithmic and Economic Issues, Chap 24 of Algorithmic Game Theory, (Cambridge University Press, 2007). Again each node in one of two states, say { 1, +1}. Play a game with each connected neighbor independently. Total payout is sum over all games. Assume neighbor(s) of j in fixed state while j updates. Positive payout if connected nodes i and j adopt the same state. No payout if they differ. And -1 can have different payout that +1 coordinated behavior. Payout matrix: q 0 0 (1-q)

17 How each node operates Again assume all other nodes fixed while node j updates. It has k A j nodes in state 1, and kb j nodes in state +1. If it chooses state 1, payout of qk A j. If it chooses state +1, payout of (1 q)k B j. Chooses 1 if qk A j > (1 q)kb j. Substitute in k j = k A j + kb j and rearrange: Criteria: choose 1 if k B j < qk j and +1 if k B j > qk j. A threshold model! Adopt +1 if a fraction q of your neighbors have state +1.

18 Contagion threshold and cascades Start all nodes in 1. And all nodes update synchronously at discrete time steps. Key question: When is there a small set of nodes S, that when set to +1 convert all (or almost all) of the population? A set S is contagious if every other node is converted by S. Easier for S to be contagious if the threshold q is small. Define the contagion threshold of a graph G to be the maximum q for which there exists a finite contagous set. (Like with generating functions, here no notion of how long it takes for the full network to be activated. Just a final steadystate answer.)

19 Progressive vs. non-progressive processes The model thus far is non-progressive: nodes can flip from 1 to +1 and back to 1. This makes the situation less stable. Consider a line of all 1 at the start with a single +1 in the center, and q = 1/2. At next time steps neighbors of the +1 flip, but the +1 switches back to 1! And the whole system ends up blinking. Progressive: Once you flip, always stay in that state. The line above now all flips to +1 in a wavefront moving right and left-wards.

20 Theorem: The Contagion Threshold for any Graph is at most 1/2. (Recall the contagion threshold is the maximum value of q for which a finite contagious set exists.) Independent of progressive vs non-progressive. A behavior cant spread very far if it requires a strict majority of your friends to convince you to adopt it. This means if q > 1/2 on any graph, it cannot support a cascade and the full graph will not be activated. This is for any graph: uniform degree, power law, etc.

21 Extending this simple model So far all nodes have same fractional threshold q, and all neighbors contribute equally in calculation of fraction. The General Linear Threshold Model Directed graphs (not reciprocal influence necessarily). Each node has a threshold chosen uniformly at random between [0, 1]. Each neighbor exerts a non-negative weight. The only constraint is that sum over all the weights is less than or equal to 1. Note we now have diversity of influence (e.g., spouse/relative can exert stronger weight than coworker/friend).

22 Finding the influential nodes Motivation Viral marketing use word-of-mouth effects to sell product with minimal advertising cost. Design of search tools to track news, blogs, and other forms of on-line discussion about current events Finding the influential nodes: formally The minimum set S V that will lead to the whole network being activated. The optimal set of a specified size k = S that will lead to largest portion of the network being activated.

23 Due to thresholds/ critical mass In general NP-hard to find optimal set S. NP-hard to even find a approximate optimal set (optimal to within factor η 1 ɛ where n is network size and ɛ > 0.) ( inapproximability ) Due to thresholds (esp if each node can have its own) might have a tiny activated final set of nodes but it jumps abruptly if just a few more nodes or, moreover, the right nodes activated. Kleinberg calls this abrupt response the Knife edge property

24 Diminishing returns (No longer a threshold, but a concave function) Each additional friend who adopts the new behavior enhances your chance of adopting the new behevaior, but with less influence for each additional friend (from Leskovec talk)

25 Diminishing returns (Submodular / concave function) The benefit of adding elements decreases as the set to which they are being added grows. So no longer get to have more influence from family or other special nodes. (Instead its the first nodes exert more influence.) Since no longer have special nodes easy to build up optimal set S of k nodes. Hill climbing add one at the time nodes to the set S that cause maximum impact.

26 Hill climbing (from Leskovec talk)

27 Submodular and hill climbing more formally: (from Leskovec talk)

28 Empirical observations (from Leskovec talk)

29 (from Leskovec talk)

30 Joining Livejournal: on online bulletin board network Probability of joining a community when k friends are already members probability k Diminishing returns only sets in once k > 3. Network effect not illustrated by curve: If the k friends are highly clustered, the new user is more likely to join.

31 (from Leskovec talk)

32 (from Leskovec talk)

33 (from Leskovec talk)

34 (from Leskovec talk)

35 (from Leskovec talk)

36 (from Leskovec talk)

37 (from Leskovec talk)

38 For a wealth of additional information see Leskovec talk: leskovec dcbn/ The role product category plays (books, dvds, videos, anime dvds) Predicting recommendation success with linear models. How do people actually get recommendations Amazon recommendation of similar purchases Personalized rec based on previous purchases/likes 68% of people consult friends and family before purchasing home electronics [Burke 2003]. (i.e. More influenced by friends than strangers.) 94% of users make recommendations w/o having received one (they are the seed nodes)

39 Another interesting recent piece Challenging the Influentials Hypothesis, Duncan J. Watts, Measuring Word of Mouth, Volume 3, Aug 2007.

40 Summary Important distinctions for cascade processes Contagion (e.g. a virus) versus social behaviors. Threshold models / critical mass (abrupt changes as set S increased) Diminishing returns (submodular / concave) KKT03, KKT05: If individual function for all nodes exhibit diminishing returns, the resulting influence function for the graph will be submodular ( local to global ). Can approximate such sets (hill climbing)

41 Other interesting, related models on networks Voter models Synchronization (related to the spectral properties of A the adjacency matrix). Finally: Part III: Markov chains and mixing times Montanari and Saberi, The Spread of Innovations in Social Networks, PNAS 2010 Unlike any of the above models (which tell us about equilibrium sizes of activated populations), Markov Chain and mixing times tell us about the time it takes for innovations to be adopted! sluggish rapid fire spread

ECS 289 F / MAE 298, Lecture 15 May 20, Diffusion, Cascades and Influence

ECS 289 F / MAE 298, Lecture 15 May 20, Diffusion, Cascades and Influence ECS 289 F / MAE 298, Lecture 15 May 20, 2014 Diffusion, Cascades and Influence Diffusion and cascades in networks (Nodes in one of two states) Viruses (human and computer) contact processes epidemic thresholds

More information

ECS 253 / MAE 253, Lecture 15 May 17, I. Probability generating function recap

ECS 253 / MAE 253, Lecture 15 May 17, I. Probability generating function recap ECS 253 / MAE 253, Lecture 15 May 17, 2016 I. Probability generating function recap Part I. Ensemble approaches A. Master equations (Random graph evolution, cluster aggregation) B. Network configuration

More information

ECS 253 / MAE 253, Lecture 13 May 15, Diffusion, Cascades and Influence Mathematical models & generating functions

ECS 253 / MAE 253, Lecture 13 May 15, Diffusion, Cascades and Influence Mathematical models & generating functions ECS 253 / MAE 253, Lecture 13 May 15, 2018 Diffusion, Cascades and Influence Mathematical models & generating functions Last week: spatial flows and game theory on networks Optimal location of facilities

More information

Diffusion of Innovation and Influence Maximization

Diffusion of Innovation and Influence Maximization Diffusion of Innovation and Influence Maximization Leonid E. Zhukov School of Data Analysis and Artificial Intelligence Department of Computer Science National Research University Higher School of Economics

More information

ECS 289 / MAE 298, Lecture 7 April 22, Percolation and Epidemiology on Networks, Part 2 Searching on networks

ECS 289 / MAE 298, Lecture 7 April 22, Percolation and Epidemiology on Networks, Part 2 Searching on networks ECS 289 / MAE 298, Lecture 7 April 22, 2014 Percolation and Epidemiology on Networks, Part 2 Searching on networks 28 project pitches turned in Announcements We are compiling them into one file to share

More information

CSCI 3210: Computational Game Theory. Cascading Behavior in Networks Ref: [AGT] Ch 24

CSCI 3210: Computational Game Theory. Cascading Behavior in Networks Ref: [AGT] Ch 24 CSCI 3210: Computational Game Theory Cascading Behavior in Networks Ref: [AGT] Ch 24 Mohammad T. Irfan Email: mirfan@bowdoin.edu Web: www.bowdoin.edu/~mirfan Course Website: www.bowdoin.edu/~mirfan/csci-3210.html

More information

Diffusion of Innovation

Diffusion of Innovation Diffusion of Innovation Leonid E. Zhukov School of Data Analysis and Artificial Intelligence Department of Computer Science National Research University Higher School of Economics Social Network Analysis

More information

Cascading Behavior in Networks: Algorithmic and Economic Issues

Cascading Behavior in Networks: Algorithmic and Economic Issues CHAPTER 24 Cascading Behavior in Networks: Algorithmic and Economic Issues Jon Kleinberg Abstract The flow of information or influence through a large social network can be thought of as unfolding with

More information

Maximizing the Spread of Influence through a Social Network. David Kempe, Jon Kleinberg, Éva Tardos SIGKDD 03

Maximizing the Spread of Influence through a Social Network. David Kempe, Jon Kleinberg, Éva Tardos SIGKDD 03 Maximizing the Spread of Influence through a Social Network David Kempe, Jon Kleinberg, Éva Tardos SIGKDD 03 Influence and Social Networks Economics, sociology, political science, etc. all have studied

More information

Probability Models of Information Exchange on Networks Lecture 6

Probability Models of Information Exchange on Networks Lecture 6 Probability Models of Information Exchange on Networks Lecture 6 UC Berkeley Many Other Models There are many models of information exchange on networks. Q: Which model to chose? My answer good features

More information

Modeling, Analysis, and Control of Information Propagation in Multi-layer and Multiplex Networks. Osman Yağan

Modeling, Analysis, and Control of Information Propagation in Multi-layer and Multiplex Networks. Osman Yağan Modeling, Analysis, and Control of Information Propagation in Multi-layer and Multiplex Networks Osman Yağan Department of ECE Carnegie Mellon University Joint work with Y. Zhuang and V. Gligor (CMU) Alex

More information

CS224W: Analysis of Networks Jure Leskovec, Stanford University

CS224W: Analysis of Networks Jure Leskovec, Stanford University Announcements: Please fill HW Survey Weekend Office Hours starting this weekend (Hangout only) Proposal: Can use 1 late period CS224W: Analysis of Networks Jure Leskovec, Stanford University http://cs224w.stanford.edu

More information

Epidemics and information spreading

Epidemics and information spreading Epidemics and information spreading Leonid E. Zhukov School of Data Analysis and Artificial Intelligence Department of Computer Science National Research University Higher School of Economics Social Network

More information

CS 322: (Social and Information) Network Analysis Jure Leskovec Stanford University

CS 322: (Social and Information) Network Analysis Jure Leskovec Stanford University CS 322: (Social and Inormation) Network Analysis Jure Leskovec Stanord University Initially some nodes S are active Each edge (a,b) has probability (weight) p ab b 0.4 0.4 0.2 a 0.4 0.2 0.4 g 0.2 Node

More information

A Note on Maximizing the Spread of Influence in Social Networks

A Note on Maximizing the Spread of Influence in Social Networks A Note on Maximizing the Spread of Influence in Social Networks Eyal Even-Dar 1 and Asaf Shapira 2 1 Google Research, Email: evendar@google.com 2 Microsoft Research, Email: asafico@microsoft.com Abstract.

More information

Epidemics in Complex Networks and Phase Transitions

Epidemics in Complex Networks and Phase Transitions Master M2 Sciences de la Matière ENS de Lyon 2015-2016 Phase Transitions and Critical Phenomena Epidemics in Complex Networks and Phase Transitions Jordan Cambe January 13, 2016 Abstract Spreading phenomena

More information

KINETICS OF SOCIAL CONTAGION. János Kertész Central European University. SNU, June

KINETICS OF SOCIAL CONTAGION. János Kertész Central European University. SNU, June KINETICS OF SOCIAL CONTAGION János Kertész Central European University SNU, June 1 2016 Theory: Zhongyuan Ruan, Gerardo Iniguez, Marton Karsai, JK: Kinetics of social contagion Phys. Rev. Lett. 115, 218702

More information

Lecture VI Introduction to complex networks. Santo Fortunato

Lecture VI Introduction to complex networks. Santo Fortunato Lecture VI Introduction to complex networks Santo Fortunato Plan of the course I. Networks: definitions, characteristics, basic concepts in graph theory II. III. IV. Real world networks: basic properties

More information

Spreading and Opinion Dynamics in Social Networks

Spreading and Opinion Dynamics in Social Networks Spreading and Opinion Dynamics in Social Networks Gyorgy Korniss Rensselaer Polytechnic Institute 05/27/2013 1 Simple Models for Epidemiological and Social Contagion Susceptible-Infected-Susceptible (SIS)

More information

Lecture 10. Under Attack!

Lecture 10. Under Attack! Lecture 10 Under Attack! Science of Complex Systems Tuesday Wednesday Thursday 11.15 am 12.15 pm 11.15 am 12.15 pm Feb. 26 Feb. 27 Feb. 28 Mar.4 Mar.5 Mar.6 Mar.11 Mar.12 Mar.13 Mar.18 Mar.19 Mar.20 Mar.25

More information

Analytically tractable processes on networks

Analytically tractable processes on networks University of California San Diego CERTH, 25 May 2011 Outline Motivation 1 Motivation Networks Random walk and Consensus Epidemic models Spreading processes on networks 2 Networks Motivation Networks Random

More information

Kristina Lerman USC Information Sciences Institute

Kristina Lerman USC Information Sciences Institute Rethinking Network Structure Kristina Lerman USC Information Sciences Institute Università della Svizzera Italiana, December 16, 2011 Measuring network structure Central nodes Community structure Strength

More information

On the Submodularity of Influence in Social Networks

On the Submodularity of Influence in Social Networks On the Submodularity of Influence in Social Networks Elchanan Mossel Dept. of Statistics U.C. Berkeley mossel@stat.berkeley.edu Sebastien Roch Dept. of Statistics U.C. Berkeley sroch@stat.berkeley.edu

More information

Lecture 11 October 11, Information Dissemination through Social Networks

Lecture 11 October 11, Information Dissemination through Social Networks CS 284r: Incentives and Information in Networks Fall 2013 Prof. Yaron Singer Lecture 11 October 11, 2013 Scribe: Michael Tingley, K. Nathaniel Tucker 1 Overview In today s lecture we will start the second

More information

Diffusion of information and social contagion

Diffusion of information and social contagion Diffusion of information and social contagion Leonid E. Zhukov School of Data Analysis and Artificial Intelligence Department of Computer Science National Research University Higher School of Economics

More information

Marks. bonus points. } Assignment 1: Should be out this weekend. } Mid-term: Before the last lecture. } Mid-term deferred exam:

Marks. bonus points. } Assignment 1: Should be out this weekend. } Mid-term: Before the last lecture. } Mid-term deferred exam: Marks } Assignment 1: Should be out this weekend } All are marked, I m trying to tally them and perhaps add bonus points } Mid-term: Before the last lecture } Mid-term deferred exam: } This Saturday, 9am-10.30am,

More information

The Spreading of Epidemics in Complex Networks

The Spreading of Epidemics in Complex Networks The Spreading of Epidemics in Complex Networks Xiangyu Song PHY 563 Term Paper, Department of Physics, UIUC May 8, 2017 Abstract The spreading of epidemics in complex networks has been extensively studied

More information

Greedy Maximization Framework for Graph-based Influence Functions

Greedy Maximization Framework for Graph-based Influence Functions Greedy Maximization Framework for Graph-based Influence Functions Edith Cohen Google Research Tel Aviv University HotWeb '16 1 Large Graphs Model relations/interactions (edges) between entities (nodes)

More information

Diffusion of Innovations in Social Networks

Diffusion of Innovations in Social Networks Daron Acemoglu Massachusetts Institute of Technology, Department of Economics, Cambridge, MA, 02139, daron@mit.edu Diffusion of Innovations in Social Networks Asuman Ozdaglar Massachusetts Institute of

More information

Multi-Stage Complex Contagions in Random Multiplex Networks

Multi-Stage Complex Contagions in Random Multiplex Networks Multi-Stage Complex Contagions in Random Multiplex Networks Yong Zhuang and Osman Yağan Department of ECE, Carnegie Mellon University, Pittsburgh, PA 523, USA (Dated: July 4, 28) In this work, we aim to

More information

Viral Marketing and the Diffusion of Trends on Social Networks

Viral Marketing and the Diffusion of Trends on Social Networks University of Pennsylvania ScholarlyCommons Technical Reports (CIS) Department of Computer & Information Science May 2008 Viral Marketing and the Diffusion of Trends on Social Networks Jennifer Wortman

More information

Cost and Preference in Recommender Systems Junhua Chen LESS IS MORE

Cost and Preference in Recommender Systems Junhua Chen LESS IS MORE Cost and Preference in Recommender Systems Junhua Chen, Big Data Research Center, UESTC Email:junmshao@uestc.edu.cn http://staff.uestc.edu.cn/shaojunming Abstract In many recommender systems (RS), user

More information

Social Choice and Networks

Social Choice and Networks Social Choice and Networks Elchanan Mossel UC Berkeley All rights reserved Logistics 1 Different numbers for the course: Compsci 294 Section 063 Econ 207A Math C223A Stat 206A Room: Cory 241 Time TuTh

More information

Latent voter model on random regular graphs

Latent voter model on random regular graphs Latent voter model on random regular graphs Shirshendu Chatterjee Cornell University (visiting Duke U.) Work in progress with Rick Durrett April 25, 2011 Outline Definition of voter model and duality with

More information

New Journal of Physics

New Journal of Physics New Journal of Physics The open access journal for physics Dynamics of competing ideas in complex social systems Yubo Wang 1, Gaoxi Xiao 1,3 and Jian Liu 2 1 School of Electrical and Electronic Engineering,

More information

MobiHoc 2014 MINIMUM-SIZED INFLUENTIAL NODE SET SELECTION FOR SOCIAL NETWORKS UNDER THE INDEPENDENT CASCADE MODEL

MobiHoc 2014 MINIMUM-SIZED INFLUENTIAL NODE SET SELECTION FOR SOCIAL NETWORKS UNDER THE INDEPENDENT CASCADE MODEL MobiHoc 2014 MINIMUM-SIZED INFLUENTIAL NODE SET SELECTION FOR SOCIAL NETWORKS UNDER THE INDEPENDENT CASCADE MODEL Jing (Selena) He Department of Computer Science, Kennesaw State University Shouling Ji,

More information

DS504/CS586: Big Data Analytics Graph Mining II

DS504/CS586: Big Data Analytics Graph Mining II Welcome to DS504/CS586: Big Data Analytics Graph Mining II Prof. Yanhua Li Time: 6:00pm 8:50pm Mon. and Wed. Location: SL105 Spring 2016 Reading assignments We will increase the bar a little bit Please

More information

Social Influence in Online Social Networks. Epidemiological Models. Epidemic Process

Social Influence in Online Social Networks. Epidemiological Models. Epidemic Process Social Influence in Online Social Networks Toward Understanding Spatial Dependence on Epidemic Thresholds in Networks Dr. Zesheng Chen Viral marketing ( word-of-mouth ) Blog information cascading Rumor

More information

Complex Networks, Course 303A, Spring, Prof. Peter Dodds

Complex Networks, Course 303A, Spring, Prof. Peter Dodds Complex Networks, Course 303A, Spring, 2009 Prof. Peter Dodds Department of Mathematics & Statistics University of Vermont Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License.

More information

Growing a Network on a Given Substrate

Growing a Network on a Given Substrate Growing a Network on a Given Substrate 1 Babak Fotouhi and Michael G. Rabbat Department of Electrical and Computer Engineering McGill University, Montréal, Québec, Canada Email: babak.fotouhi@mail.mcgill.ca,

More information

Web Structure Mining Nodes, Links and Influence

Web Structure Mining Nodes, Links and Influence Web Structure Mining Nodes, Links and Influence 1 Outline 1. Importance of nodes 1. Centrality 2. Prestige 3. Page Rank 4. Hubs and Authority 5. Metrics comparison 2. Link analysis 3. Influence model 1.

More information

CS224W: Social and Information Network Analysis Jure Leskovec, Stanford University

CS224W: Social and Information Network Analysis Jure Leskovec, Stanford University CS224W: Social and Information Network Analysis Jure Leskovec, Stanford University http://cs224w.stanford.edu Find most influential set S of size k: largest expected cascade size f(s) if set S is activated

More information

CS 277: Data Mining. Mining Web Link Structure. CS 277: Data Mining Lectures Analyzing Web Link Structure Padhraic Smyth, UC Irvine

CS 277: Data Mining. Mining Web Link Structure. CS 277: Data Mining Lectures Analyzing Web Link Structure Padhraic Smyth, UC Irvine CS 277: Data Mining Mining Web Link Structure Class Presentations In-class, Tuesday and Thursday next week 2-person teams: 6 minutes, up to 6 slides, 3 minutes/slides each person 1-person teams 4 minutes,

More information

Influence Maximization in Dynamic Social Networks

Influence Maximization in Dynamic Social Networks Influence Maximization in Dynamic Social Networks Honglei Zhuang, Yihan Sun, Jie Tang, Jialin Zhang and Xiaoming Sun Department of Computer Science and Technology, Tsinghua University Department of Computer

More information

Modeling Strategic Information Sharing in Indian Villages

Modeling Strategic Information Sharing in Indian Villages Modeling Strategic Information Sharing in Indian Villages Jeff Jacobs jjacobs3@stanford.edu Arun Chandrasekhar arungc@stanford.edu Emily Breza Columbia University ebreza@columbia.edu December 3, 203 Matthew

More information

DS504/CS586: Big Data Analytics Graph Mining II

DS504/CS586: Big Data Analytics Graph Mining II Welcome to DS504/CS586: Big Data Analytics Graph Mining II Prof. Yanhua Li Time: 6-8:50PM Thursday Location: AK233 Spring 2018 v Course Project I has been graded. Grading was based on v 1. Project report

More information

Diffusion of Behavior and Equilibrium Properties in Network Games

Diffusion of Behavior and Equilibrium Properties in Network Games Diffusion of Behavior and Equilibrium Properties in Network Games By Matthew O. Jackson and Leeat Yariv* Situations in which agents choices depend on choices of those in close proximity, be it social or

More information

Maximizing the Spread of Influence through a Social Network

Maximizing the Spread of Influence through a Social Network Maximizing the Spread of Influence through a Social Network David Kempe Dept. of Computer Science Cornell University, Ithaca NY kempe@cs.cornell.edu Jon Kleinberg Dept. of Computer Science Cornell University,

More information

Spring 2016 Network Science. Solution of Quiz I

Spring 2016 Network Science. Solution of Quiz I Spring 2016 Network Science Department of Electrical and Computer Engineering National Chiao Tung University Solution of Quiz I Problem Points Your Score 1 5 2 20 25 20 Total 100 Read all of the following

More information

Algorithmic Game Theory and Applications

Algorithmic Game Theory and Applications Algorithmic Game Theory and Applications Lecture 18: Auctions and Mechanism Design II: a little social choice theory, the VCG Mechanism, and Market Equilibria Kousha Etessami Reminder: Food for Thought:

More information

6.207/14.15: Networks Lecture 12: Generalized Random Graphs

6.207/14.15: Networks Lecture 12: Generalized Random Graphs 6.207/14.15: Networks Lecture 12: Generalized Random Graphs 1 Outline Small-world model Growing random networks Power-law degree distributions: Rich-Get-Richer effects Models: Uniform attachment model

More information

Machine Learning and Modeling for Social Networks

Machine Learning and Modeling for Social Networks Machine Learning and Modeling for Social Networks Olivia Woolley Meza, Izabela Moise, Nino Antulov-Fatulin, Lloyd Sanders 1 Spreading and Influence on social networks Computational Social Science D-GESS

More information

Learning to Predict Opinion Share in Social Networks

Learning to Predict Opinion Share in Social Networks Proceedings of the Twenty-Fourth AAAI Conference on Artificial Intelligence AAAI-10 Learning to Predict Opinion Share in Social Networks Masahiro Kimura Department of Electronics and Informatics Ryukoku

More information

Online Social Networks and Media. Opinion formation on social networks

Online Social Networks and Media. Opinion formation on social networks Online Social Networks and Media Opinion formation on social networks Diffusion of items So far we have assumed that what is being diffused in the network is some discrete item: E.g., a virus, a product,

More information

Decision Making and Social Networks

Decision Making and Social Networks Decision Making and Social Networks Lecture 4: Models of Network Growth Umberto Grandi Summer 2013 Overview In the previous lecture: We got acquainted with graphs and networks We saw lots of definitions:

More information

CSI 445/660 Part 3 (Networks and their Surrounding Contexts)

CSI 445/660 Part 3 (Networks and their Surrounding Contexts) CSI 445/660 Part 3 (Networks and their Surrounding Contexts) Ref: Chapter 4 of [Easley & Kleinberg]. 3 1 / 33 External Factors ffecting Network Evolution Homophily: basic principle: We tend to be similar

More information

Models of Communication Dynamics for Simulation of Information Diffusion

Models of Communication Dynamics for Simulation of Information Diffusion Models of Communication Dynamics for Simulation of Information Diffusion Konstantin Mertsalov, Malik Magdon-Ismail, Mark Goldberg Rensselaer Polytechnic Institute Department of Computer Science 11 8th

More information

Contagion and coordination in random networks

Contagion and coordination in random networks Contagion and coordination in random networks Dunia López-Pintado September 9, 2005 Abstract We study the problem of spreading a particular behavior among agents located in a random social network. In

More information

Learning with Temporal Point Processes

Learning with Temporal Point Processes Learning with Temporal Point Processes t Manuel Gomez Rodriguez MPI for Software Systems Isabel Valera MPI for Intelligent Systems Slides/references: http://learning.mpi-sws.org/tpp-icml18 ICML TUTORIAL,

More information

MAE 298, Lecture 8 Feb 4, Web search and decentralized search on small-worlds

MAE 298, Lecture 8 Feb 4, Web search and decentralized search on small-worlds MAE 298, Lecture 8 Feb 4, 2008 Web search and decentralized search on small-worlds Search for information Assume some resource of interest is stored at the vertices of a network: Web pages Files in a file-sharing

More information

CS224W: Social and Information Network Analysis Jure Leskovec, Stanford University

CS224W: Social and Information Network Analysis Jure Leskovec, Stanford University CS224W: Social and Information Network Analysis Jure Leskovec, Stanford University http://cs224w.stanford.edu 10/24/2012 Jure Leskovec, Stanford CS224W: Social and Information Network Analysis, http://cs224w.stanford.edu

More information

CS 6604: Data Mining Large Networks and Time-series. B. Aditya Prakash Lecture #8: Epidemics: Thresholds

CS 6604: Data Mining Large Networks and Time-series. B. Aditya Prakash Lecture #8: Epidemics: Thresholds CS 6604: Data Mining Large Networks and Time-series B. Aditya Prakash Lecture #8: Epidemics: Thresholds A fundamental ques@on Strong Virus Epidemic? 2 example (sta@c graph) Weak Virus Epidemic? 3 Problem

More information

KINETICS OF COMPLEX SOCIAL CONTAGION. János Kertész Central European University. Pohang, May 27, 2016

KINETICS OF COMPLEX SOCIAL CONTAGION. János Kertész Central European University. Pohang, May 27, 2016 KINETICS OF COMPLEX SOCIAL CONTAGION János Kertész Central European University Pohang, May 27, 2016 Theory: Zhongyuan Ruan, Gerardo Iniguez, Marton Karsai, JK: Kinetics of social contagion Phys. Rev. Lett.

More information

Contagion. Complex Networks CSYS/MATH 303, Spring, Prof. Peter Dodds

Contagion. Complex Networks CSYS/MATH 303, Spring, Prof. Peter Dodds Complex Networks CSYS/MATH 33, Spring, 211 Basic Social Prof. Peter Dodds Department of Mathematics & Statistics Center for Complex Systems Vermont Advanced Computing Center University of Vermont Licensed

More information

6.207/14.15: Networks Lecture 16: Cooperation and Trust in Networks

6.207/14.15: Networks Lecture 16: Cooperation and Trust in Networks 6.207/14.15: Networks Lecture 16: Cooperation and Trust in Networks Daron Acemoglu and Asu Ozdaglar MIT November 4, 2009 1 Introduction Outline The role of networks in cooperation A model of social norms

More information

Modeling Social Media Memes as a Contagious Process

Modeling Social Media Memes as a Contagious Process Modeling Social Media Memes as a Contagious Process S.Towers 1,, A.Person 2, C.Castillo-Chavez 1 1 Arizona State University, Tempe, AZ, USA 2 Some University, Nowhereville, NE, USA E-mail: smtowers@asu.edu

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

Epidemics on networks

Epidemics on networks Epidemics on networks Leonid E. Zhukov School of Data Analysis and Artificial Intelligence Department of Computer Science National Research University Higher School of Economics Network Science Leonid

More information

CS 598 Statistical Reinforcement Learning. Nan Jiang

CS 598 Statistical Reinforcement Learning. Nan Jiang CS 598 Statistical Reinforcement Learning Nan Jiang Overview What s this course about? A grad-level seminar course on theory of RL 3 What s this course about? A grad-level seminar course on theory of RL

More information

A note on modeling retweet cascades on Twitter

A note on modeling retweet cascades on Twitter A note on modeling retweet cascades on Twitter Ashish Goel 1, Kamesh Munagala 2, Aneesh Sharma 3, and Hongyang Zhang 4 1 Department of Management Science and Engineering, Stanford University, ashishg@stanford.edu

More information

Administration. CSCI567 Machine Learning (Fall 2018) Outline. Outline. HW5 is available, due on 11/18. Practice final will also be available soon.

Administration. CSCI567 Machine Learning (Fall 2018) Outline. Outline. HW5 is available, due on 11/18. Practice final will also be available soon. Administration CSCI567 Machine Learning Fall 2018 Prof. Haipeng Luo U of Southern California Nov 7, 2018 HW5 is available, due on 11/18. Practice final will also be available soon. Remaining weeks: 11/14,

More information

Finite Markov Information-Exchange processes

Finite Markov Information-Exchange processes Finite Markov Information-Exchange processes David Aldous February 2, 2011 Course web site: Google Aldous STAT 260. Style of course Big Picture thousands of papers from different disciplines (statistical

More information

Assortativity and Mixing. Outline. Definition. General mixing. Definition. Assortativity by degree. Contagion. References. Contagion.

Assortativity and Mixing. Outline. Definition. General mixing. Definition. Assortativity by degree. Contagion. References. Contagion. Outline Complex Networks, Course 303A, Spring, 2009 Prof Peter Dodds Department of Mathematics & Statistics University of Vermont Licensed under the Creative Commons Attribution-NonCommercial-ShareAlike

More information

Preventive behavioural responses and information dissemination in network epidemic models

Preventive behavioural responses and information dissemination in network epidemic models PROCEEDINGS OF THE XXIV CONGRESS ON DIFFERENTIAL EQUATIONS AND APPLICATIONS XIV CONGRESS ON APPLIED MATHEMATICS Cádiz, June 8-12, 215, pp. 111 115 Preventive behavioural responses and information dissemination

More information

Detecting Anti-majority Opinionists Using Value-weighted Mixture Voter Model

Detecting Anti-majority Opinionists Using Value-weighted Mixture Voter Model Detecting Anti-majority Opinionists Using Value-weighted Mixture Voter Model Masahiro Kimura, Kazumi Saito 2, Kouzou Ohara 3, and Hiroshi Motoda 4 Department of Electronics and Informatics, Ryukoku University

More information

VCMC: Variational Consensus Monte Carlo

VCMC: Variational Consensus Monte Carlo VCMC: Variational Consensus Monte Carlo Maxim Rabinovich, Elaine Angelino, Michael I. Jordan Berkeley Vision and Learning Center September 22, 2015 probabilistic models! sky fog bridge water grass object

More information

Time varying networks and the weakness of strong ties

Time varying networks and the weakness of strong ties Supplementary Materials Time varying networks and the weakness of strong ties M. Karsai, N. Perra and A. Vespignani 1 Measures of egocentric network evolutions by directed communications In the main text

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

6.207/14.15: Networks Lecture 4: Erdös-Renyi Graphs and Phase Transitions

6.207/14.15: Networks Lecture 4: Erdös-Renyi Graphs and Phase Transitions 6.207/14.15: Networks Lecture 4: Erdös-Renyi Graphs and Phase Transitions Daron Acemoglu and Asu Ozdaglar MIT September 21, 2009 1 Outline Phase transitions Connectivity threshold Emergence and size of

More information

WITH the recent advancements of information technologies,

WITH the recent advancements of information technologies, i Distributed Rumor Blocking with Multiple Positive Cascades Guangmo (Amo) Tong, Student Member, IEEE, Weili Wu, Member, IEEE, and Ding-Zhu Du, arxiv:1711.07412 [cs.si] 1 Dec 2017 Abstract Misinformation

More information

Spatial Epidemic Modelling in Social Networks

Spatial Epidemic Modelling in Social Networks Spatial Epidemic Modelling in Social Networks Joana Margarida Simoes Centre for Advanced Spatial Analysis, University College of London, UK Abstract. The spread of infectious diseases is highly influenced

More information

6.207/14.15: Networks Lecture 7: Search on Networks: Navigation and Web Search

6.207/14.15: Networks Lecture 7: Search on Networks: Navigation and Web Search 6.207/14.15: Networks Lecture 7: Search on Networks: Navigation and Web Search Daron Acemoglu and Asu Ozdaglar MIT September 30, 2009 1 Networks: Lecture 7 Outline Navigation (or decentralized search)

More information

arxiv: v1 [cs.si] 14 Dec 2018

arxiv: v1 [cs.si] 14 Dec 2018 Information Diffusion in Social Networks: Friendship Paradox based Models and Statistical Inference Vikram Krishnamurthy and Buddhika Nettasinghe arxiv:1812.679v1 [cs.si] 14 Dec 218 Abstract Dynamic models

More information

On Susceptible-Infected-Susceptible Epidemic Spreading: an Overview of Recent Study

On Susceptible-Infected-Susceptible Epidemic Spreading: an Overview of Recent Study The 1 st Net-X (2017) On Susceptible-Infected-Susceptible Epidemic Spreading: an Overview of Recent Study Cong Li Adaptive Networks and Control Lab, Fudan University Collaborators: Xiaojie Li, Jianbo Wang,

More information

Controlling Spreading Processes in Networks

Controlling Spreading Processes in Networks Controlling Spreading Processes in Networks (or how to rapidly spread mutants throughout a geographic region) Lorenzo Zino DISMA Weekly Seminar (Seminari Cadenzati - Progetto Eccellenza) July, 08 Why to

More information

CONNECTING DATA WITH COMPETING OPINION MODELS

CONNECTING DATA WITH COMPETING OPINION MODELS CONNECTING DATA WITH COMPETING OPINION MODELS Keith Burghardt with Prof. Michelle Girvan and Prof. Bill Rand Oct. 31, 2014 OUTLINE Introduction and background Motivation and empirical findings Description

More information

Any live cell with less than 2 live neighbours dies. Any live cell with 2 or 3 live neighbours lives on to the next step.

Any live cell with less than 2 live neighbours dies. Any live cell with 2 or 3 live neighbours lives on to the next step. 2. Cellular automata, and the SIRS model In this Section we consider an important set of models used in computer simulations, which are called cellular automata (these are very similar to the so-called

More information

Diusion of Behavior and Equilibrium Properties in Network Games. Matthew O. Jackson and Leeat Yariv. February 26, 2007

Diusion of Behavior and Equilibrium Properties in Network Games. Matthew O. Jackson and Leeat Yariv. February 26, 2007 Diusion of Behavior and Equilibrium Properties in Network Games Matthew O. Jackson and Leeat Yariv February 26, 2007 Situations in which agents' choices depend on choices of those in close proximity, be

More information

Competitive Contagion in Networks

Competitive Contagion in Networks Competitive Contagion in Networks Sanjeev Goyal, Hoda Heidari, Michael Kearns Abstract We develop a game-theoretic framework for the study of competition between firms who have budgets to seed the initial

More information

In biological terms, memory refers to the ability of neural systems to store activity patterns and later recall them when required.

In biological terms, memory refers to the ability of neural systems to store activity patterns and later recall them when required. In biological terms, memory refers to the ability of neural systems to store activity patterns and later recall them when required. In humans, association is known to be a prominent feature of memory.

More information

Modeling face-to-face social interaction networks

Modeling face-to-face social interaction networks Modeling face-to-face social interaction networks Romualdo Pastor-Satorras Dept. Fisica i Enginyería Nuclear Universitat Politècnica de Catalunya Spain http://www.fen.upc.edu/~romu Work done in collaboration

More information

Information Aggregation in Complex Dynamic Networks

Information Aggregation in Complex Dynamic Networks The Combined 48 th IEEE Conference on Decision and Control and 28 th Chinese Control Conference Information Aggregation in Complex Dynamic Networks Ali Jadbabaie Skirkanich Associate Professor of Innovation

More information

Modeling Dynamic Evolution of Online Friendship Network

Modeling Dynamic Evolution of Online Friendship Network Commun. Theor. Phys. 58 (2012) 599 603 Vol. 58, No. 4, October 15, 2012 Modeling Dynamic Evolution of Online Friendship Network WU Lian-Ren ( ) 1,2, and YAN Qiang ( Ö) 1 1 School of Economics and Management,

More information

The decoupling assumption in large stochastic system analysis Talk at ECLT

The decoupling assumption in large stochastic system analysis Talk at ECLT The decoupling assumption in large stochastic system analysis Talk at ECLT Andrea Marin 1 1 Dipartimento di Scienze Ambientali, Informatica e Statistica Università Ca Foscari Venezia, Italy (University

More information

Mechanism Design. Christoph Schottmüller / 27

Mechanism Design. Christoph Schottmüller / 27 Mechanism Design Christoph Schottmüller 2015-02-25 1 / 27 Outline 1 Bayesian implementation and revelation principle 2 Expected externality mechanism 3 Review questions and exercises 2 / 27 Bayesian implementation

More information

Cellular Automata Models for Diffusion of Innovations

Cellular Automata Models for Diffusion of Innovations arxiv:adap-org/9742v 8 Apr 997 Cellular Automata Models for Diffusion of Innovations Henryk Fukś Nino Boccara,2 February 3, 28 Department of Physics, University of Illinois, Chicago, IL 667-759, USA 2

More information

CS 781 Lecture 9 March 10, 2011 Topics: Local Search and Optimization Metropolis Algorithm Greedy Optimization Hopfield Networks Max Cut Problem Nash

CS 781 Lecture 9 March 10, 2011 Topics: Local Search and Optimization Metropolis Algorithm Greedy Optimization Hopfield Networks Max Cut Problem Nash CS 781 Lecture 9 March 10, 2011 Topics: Local Search and Optimization Metropolis Algorithm Greedy Optimization Hopfield Networks Max Cut Problem Nash Equilibrium Price of Stability Coping With NP-Hardness

More information

Lecture: Local Spectral Methods (2 of 4) 19 Computing spectral ranking with the push procedure

Lecture: Local Spectral Methods (2 of 4) 19 Computing spectral ranking with the push procedure Stat260/CS294: Spectral Graph Methods Lecture 19-04/02/2015 Lecture: Local Spectral Methods (2 of 4) Lecturer: Michael Mahoney Scribe: Michael Mahoney Warning: these notes are still very rough. They provide

More information

Improved Bounds on the Epidemic Threshold of Exact SIS Models on Complex Networks

Improved Bounds on the Epidemic Threshold of Exact SIS Models on Complex Networks Improved Bounds on the Epidemic Threshold of Exact SIS Models on Complex Networks Navid Azizan Ruhi, Christos Thrampoulidis and Babak Hassibi Abstract The SIS (susceptible-infected-susceptible) epidemic

More information

Word of Mouth Advertising, Credibility and Learning in Networks 1

Word of Mouth Advertising, Credibility and Learning in Networks 1 Word of Mouth Advertising, Credibility and Learning in Networks 1 Kalyan Chatterjee Department of Economics, The Pennsylvania State University, University Park, Pa. 16802, USA. Bhaskar Dutta Department

More information