Finding a team of Experts in Social Networks

Finding a team of Experts in Social Networks Theodoros Lappas UC Riverside Joint work with: Evimaria Terzi (IBM Almaden), Kun Liu (IBM Almaden) 3/10/2021

Motivation “ How can I find a team of experts that can collaborate effectively in order to complete a given task? ” 2 3/10/2021

Problem § Given a task and a set of experts organized in a network, find a subset of experts that can effectively perform the task § Task: set of required skills § Expert: an individual with a specific skill-set § Network: represents strength of relationships 3 3/10/2021

Expertise networks § Collaboration networks (e. g. , DBLP graph, coauthor networks) § Organizational structure of companies § Linked. In § Geographical (map) of experts 4 3/10/2021

What makes a team effective for a task? § T = {algorithms, java, graphics, python} Alice {algorithms} Coverage: Bob {python} Cynthia {graphics, java} David {graphics} {graphics, java, python} every required skill in T is included in the skill-set of at least one team member 5 Eleanor 3/10/2021

Is coverage enough? Alice {algorithms} Bob Cynthia {python} David {graphics, java} {graphics} A A, B, C form an effective group that can communicate B Communication: T={algorithms, java, graphics, python} D C Eleanor {graphics, java, python} A, E could perform the task if they could communicate E the members of the team must be able to efficiently communicate and work together 6 3/10/2021

Problem definition § Given a task and a social network of individuals G, find the subset (team) of G that can effectively perform the given task. § Thesis: Good teams are teams that have the necessary skills and can also communicate effectively 7 3/10/2021

How to measure effective communication? The longest shortest path between any two nodes in the subgraph § Diameter of the subgraph defined by the group members A B D C E diameter = infty =1 8 3/10/2021

How to measure effective communication? The total weight of the edges of a tree that spans all the team nodes § MST (Minimum spanning tree) of the subgraph defined by the group members A B D C E MST= =infty 2 9 3/10/2021

Problem definition – v. 1. 1 § Given a task and a social network G of individuals, find the subset (team) of individuals that can perform the given task and define a subgraph in G with the minimum diameter. § Problem is NP-hard 10 3/10/2021

The Rarest. First algorithm T={algorithms, java, graphics, python} {graphics, python, java} {algorithms, graphics} A B Skills: algorithms E {algorithms, graphics, java} graphics java C {python, java} D python {python} αrare= algorithms Diameter = 2 Srare={Bob, Eleanor} 11 3/10/2021

The Rarest. First algorithm T={algorithms, java, graphics, python} {graphics, python, java} {algorithms, graphics} A B Skills: algorithms E {algorithms, graphics, java} graphics java C {python, java} D python {python} αrare=time: algorithms Diameter = 1 Running Quadratic to the number of nodes Srare={Bob, Efactor: leanor} 2 x. OPT Approximation 12 3/10/2021

Problem definition – v. 1. 2 § Given a task and a social network G of individuals, find the subset (team) of individuals that can perform the given task and define a subgraph in G with the minimum MST cost. § Problem is NP-hard Best known Approximation factor: O(log 3 n log k) 13 3/10/2021

The Steiner. Tree problem § Graph G(V, E) Required vertices § Set of Required Vertices R § Find G’ subgraph of G such that G’ contains all the required vertices (R) and MST(G’) is minimized 14 3/10/2021

The Enhanced. Steiner algorithm T={algorithms, java, graphics, python} graphics {graphics, python, java} java {algorithms, graphics} A B E python C {python, java} {algorithms, graphics, java} algorithms D {python} MST Cost = 1 15 3/10/2021

Experiments 16 3/10/2021

Dataset DBLP Dataset ( DM, AI, DB, T ) ~6000 authors Skills: keywords appearing in paper titles ~2000 features Social Network: Co-Authorship Graph Tasks: Subsets of keywords with different cardinality 17 3/10/2021

Cardinality of teams 18 3/10/2021

Example teams (I) § S. Brin, L. Page: The anatomy of a large-scale hypertextual Web search engine – Paolo Ferragina, Patrick Valduriez, H. V. Jagadish, Alon Y. Levy, Daniela Florescu, Divesh Srivastava, S. Muthukrishnan – P. Ferragina , J. Han, H. V. Jagadish, Kevin Chen-Chuan Chang, A. Gulli, S. Muthukrishnan, Laks V. S. Lakshmanan 19 3/10/2021

Example teams (II) § J. Han, J. Pei, Y. Yin: Mining frequent patterns without candidate generation – F. Bonchi – A. Gionis, H. Mannila, R. Motwani 20 3/10/2021

Extensions § Other measures of effective communication § Other practical restrictions – Incorporate ability levels 21 3/10/2021

Thanks for your attention! 22 3/10/2021