#### DMCA

## On new characterizations of social influence in social networks

Venue: | in Proceedings of the 2013 American Control Conference, 2013 |

Citations: | 4 - 3 self |

### Citations

7413 | Convex Optimization - Boyd, Vandenberghe - 2004 |

2582 | The structure and function of complex networks
- Newman
- 2010
(Show Context)
Citation Context ...tive search to find the globally optimal selection of FAs for nφ ≤ 7. It can be seen in Table II that the coordinate descent method finds the global solution for nφ = 6, 7. The betweenness centrality =-=[2]-=- of node i is defined as β(i) = ∑ j 6=i 6=l τjl(i) τjl , where τjl is the total number of shortest paths from node j to node l, and τjl(i) is the number of those paths that pass through node i. Betwee... |

1518 |
Introduction to Linear and Nonlinear Programming
- Luenberger
- 1973
(Show Context)
Citation Context ...e coordinate descent method has been widely used in sparse signal processing and machine learning [15]–[18]. Due to its simple implementation, this method is often used for large-scale problems [19], =-=[20]-=-. Furthermore, this method has the desirable property of being applicable to, and convergent for, nonsmooth problems [21], [22]. Instead of minimizing the objective function in (5) with respect to the... |

1175 |
Nonlinear programming
- Bertsekas
- 1995
(Show Context)
Citation Context ...hod The coordinate descent method has been widely used in sparse signal processing and machine learning [15]–[18]. Due to its simple implementation, this method is often used for large-scale problems =-=[19]-=-, [20]. Furthermore, this method has the desirable property of being applicable to, and convergent for, nonsmooth problems [21], [22]. Instead of minimizing the objective function in (5) with respect ... |

985 | Maximizing the spread of influence through a social network
- Kempe, Kleinberg, et al.
- 2003
(Show Context)
Citation Context ... influence. The present work is motivated by the papers [1]–[4], which consider the problem of quantifying the influence of agents in a social network. Similar problems have also been investigated in =-=[5]-=-, [6]. In this work we use techniques from systems theory to develop alternative characterizations of social influence, and employ them to formulate pertinent social network design problems. To this e... |

719 | Regularization paths for generalized linear models via coordinate descent
- Friedman
- 2010
(Show Context)
Citation Context ...imple but efficient algorithm which we describe in Section III-A. A. Coordinate Descent Method The coordinate descent method has been widely used in sparse signal processing and machine learning [15]–=-=[18]-=-. Due to its simple implementation, this method is often used for large-scale problems [19], [20]. Furthermore, this method has the desirable property of being applicable to, and convergent for, nonsm... |

396 |
Reaching a Consensus
- DeGroot
- 1974
(Show Context)
Citation Context ... agent updates its belief by taking a weighted average of the beliefs of those agents it socially interacts with. This is a widely-used belief evolution scheme, often referred to as the DeGroot model =-=[7]-=-. We point out that although the DeGroot Financial support from the National Science Foundation under awards CMMI-0927509 and CMMI-0927720 and under CAREER Award CMMI0644793 is gratefully acknowledged... |

368 | Sparse reconstruction by separable approximation - Wright, Nowak, et al. - 2009 |

323 | Pathwise coordinate optimization
- Friedman, Hastie, et al.
(Show Context)
Citation Context ..., a simple but efficient algorithm which we describe in Section III-A. A. Coordinate Descent Method The coordinate descent method has been widely used in sparse signal processing and machine learning =-=[15]-=-–[18]. Due to its simple implementation, this method is often used for large-scale problems [19], [20]. Furthermore, this method has the desirable property of being applicable to, and convergent for, ... |

296 | Convergence of a block coordinate descent method for nondifferentiable minimization
- Tseng
(Show Context)
Citation Context ...imple implementation, this method is often used for large-scale problems [19], [20]. Furthermore, this method has the desirable property of being applicable to, and convergent for, nonsmooth problems =-=[21]-=-, [22]. Instead of minimizing the objective function in (5) with respect to the entire vector φ, the coordinate descent method holds fixed n − 1 variables φj for j 6= i and minimizes the objective fun... |

209 | Introduction to Probability - Grinstead, Snell - 1997 |

205 | A dual coordinate descent method for large-scale linear SVM - Hsieh, Chang, et al. - 2008 |

159 |
A coordinate gradient descent method for nonsmooth separable minimization
- Tseng, Yun
- 2009
(Show Context)
Citation Context ...implementation, this method is often used for large-scale problems [19], [20]. Furthermore, this method has the desirable property of being applicable to, and convergent for, nonsmooth problems [21], =-=[22]-=-. Instead of minimizing the objective function in (5) with respect to the entire vector φ, the coordinate descent method holds fixed n − 1 variables φj for j 6= i and minimizes the objective function ... |

114 | Persuasion bias, social influence, and unidimensional opinions.
- DeMarzo, Vayanos, et al.
- 2003
(Show Context)
Citation Context ...fluential group of agents that can most effectively sway public opinion. To answer this question, one first needs to develop a measure of social influence. The present work is motivated by the papers =-=[1]-=-–[4], which consider the problem of quantifying the influence of agents in a social network. Similar problems have also been investigated in [5], [6]. In this work we use techniques from systems theor... |

106 |
A graph-theoretic perspective on centrality
- Borgatti, Everett
(Show Context)
Citation Context ...ths that pass through node i. Betweenness centrality can be used to measure the importance of a node, as it quantifies the number of shortest paths that will become longer when a node is removed [2], =-=[3]-=-. To compare the selection of FAs based on betweenness centrality versus that obtained from solving (5), we first rank the nodes in Fig. 1 in descending order based on their betweenness centrality, an... |

98 | Naive learning in social networks and the wisdom of crowds.
- Golub, Jackson
- 2010
(Show Context)
Citation Context ...ntial group of agents that can most effectively sway public opinion. To answer this question, one first needs to develop a measure of social influence. The present work is motivated by the papers [1]–=-=[4]-=-, which consider the problem of quantifying the influence of agents in a social network. Similar problems have also been investigated in [5], [6]. In this work we use techniques from systems theory to... |

42 | Spread of (mis)information in social networks
- Acemoglu, Ozdaglar, et al.
- 2010
(Show Context)
Citation Context ...ing, University of Minnesota, Minneapolis, MN 55455. E-mails: makan@syr.edu, fu@umn.edu, xzhang49@syr.edu, mihailo@umn.edu. model is simpler than those employed in some recent work on social networks =-=[8]-=-–[10], it allows for more constructive results by permitting the definition of performance measures and the formulation of optimization problems for network design. We exploit the property that social... |

40 | Opinion dynamics and learning in social networks - Acemoglu, Ozdaglar - 2011 |

26 | Opinion fluctuations and disagreement in social networks
- Acemoglu, Como, et al.
(Show Context)
Citation Context ... University of Minnesota, Minneapolis, MN 55455. E-mails: makan@syr.edu, fu@umn.edu, xzhang49@syr.edu, mihailo@umn.edu. model is simpler than those employed in some recent work on social networks [8]–=-=[10]-=-, it allows for more constructive results by permitting the definition of performance measures and the formulation of optimization problems for network design. We exploit the property that social netw... |

22 |
Distributed control of positive systems,” in
- Rantzer
- 2011
(Show Context)
Citation Context ...e results by permitting the definition of performance measures and the formulation of optimization problems for network design. We exploit the property that social networks are positive systems [11], =-=[12]-=- to introduce multiple new characterizations of social influence that capture the efficiency of a given set of nodes in propagating its opinion through society. We then use these characterizations to ... |

21 |
Leader selection for optimal network coherence
- Patterson, Bamieh
- 2010
(Show Context)
Citation Context ...Conference (ACC) Washington, DC, USA, June 17-19, 2013 978-1-4799-0176-0/$31.00 ©2013 AACC 4784 1s2s4s10s5s6s3s7s8s9s12s11s13s14s15s16s17 18s22s25s23s24s21s19s20sFig. 1: The social network example of =-=[13]-=-. Consider the graph of a social network given in Fig. 1. We assume that the matrix T is such that at every time instant every agent updates its belief by taking the average of its own belief with tha... |

12 | Marketing in a random network
- Amini, Draief, et al.
- 2009
(Show Context)
Citation Context ...uence. The present work is motivated by the papers [1]–[4], which consider the problem of quantifying the influence of agents in a social network. Similar problems have also been investigated in [5], =-=[6]-=-. In this work we use techniques from systems theory to develop alternative characterizations of social influence, and employ them to formulate pertinent social network design problems. To this end, w... |

12 | Algorithms for leader selection in stochastically forced consensus networks
- Lin, Fardad, et al.
- 2014
(Show Context)
Citation Context ... some existing ones in the literature. Using these measures, we address the problems of optimal leader selection in social networks and the optimal creation of new social interactions; also see [23], =-=[24]-=- for related efforts. We show that although the formulated optimization problems are nonconvex, they still possess the property of being convex with respect to individual entries of the optimization v... |

9 | On the optimal dissemination of information in social networks
- Fardad, Zhang, et al.
(Show Context)
Citation Context ...hem to some existing ones in the literature. Using these measures, we address the problems of optimal leader selection in social networks and the optimal creation of new social interactions; also see =-=[23]-=-, [24] for related efforts. We show that although the formulated optimization problems are nonconvex, they still possess the property of being convex with respect to individual entries of the optimiza... |

5 |
KYP lemma for internally positive systems and a tractable class of distributed H∞ control problems
- Tanaka, Langbort
- 2010
(Show Context)
Citation Context ...ructive results by permitting the definition of performance measures and the formulation of optimization problems for network design. We exploit the property that social networks are positive systems =-=[11]-=-, [12] to introduce multiple new characterizations of social influence that capture the efficiency of a given set of nodes in propagating its opinion through society. We then use these characterizatio... |