Advanced Topics in Computer Science 2 Networks, Crowds and Markets Fall, 2012-2013 Prof. Klara Kedem klara@cs.bgu.ac.il Office 110/37 Office hours by email appointment
The textbook Chapters from the book of Jon Kleinberg and David Easley : Networks, Crowds and Markets: Reasoning about a Highly Connected World The chapters will be 1-3,5,6,8-10, 12-15
Course requirements Attendance is obligatory Each student will give a 1 hour lecture, or 2 students will work on a chapter. The power point presentation will be presented to me at least 5 days before presented in class There will be an in-class quiz at the end of the semester The grade will consist of: 10 points the quiz, 75 the talk and 15 attendance (5 points will be taken off for each unjustified class missing).
Chapter 1 - Overview The connectedness of modern society Networks social, information, economics, finances Links, web pages, blogs Behavior and dynamics, interconnectivity Working with massive network datasets Theory involved: graph theory, game theory, Strategic interaction on networks.
Chapter 2 Graphs (notice that the book is for wide audience, for CS no need to go into graph theory only in connection with ideas we ll be talking on later, e.g., networks, giant components etc.) Some graph terms and algorithms Small world phenomenon Giant components Large network datasets Leave time to ask and discuss in class the exercises with answers
Chapter 3 - Strong and weak ties How information flows through a social network, how different nodes can play structurally distinct roles in this process Triadic closure if 2 people on the net have a common friend then there is high likelihood that they will become friends in the future The strength of weak ties why will get more leads to job interviews from weak acquaintances than from your very close friends? Tie strength and network structure in large-scale data - weak ties serve to link together different tightly-knit communities Tie strength, social media, and passive engagement (on Facebook) closure, structural holes, and social capital Betweenness measures Exercises
Chapter 5 Positive and negative relationships The last bullet here is a long and somewhat hairier sub section, so either I give it in 1.5 meetings and 2-3 students or skip it altogether. Without the exercises. The structural balance for networks friends and enemies balance (as in like and dislike) The balance theorem and its proof Applications of structural balance Weaker Structural balance (with the proof) Extending the balance idea to (a) non cliques (general networks), (b) approximately balanced networks
Chapter 6 Games Connectedness of the network and behavior of individuals. Definitions and examples. What is a game, underlying assumptions, common knowledge of the game analyzing a game for 2 Strictly dominant strategy and rational game playing Definitions : best response, strict best response, dominant strategy and strcitly dominant strategy Nash equilibrium-we should expect players to use strategies that are best responses to each other (equilibrium of beliefs). Emphasize this definition in all the examples Multiple equilibria
Chapter 6 Games (continued) Randomization, equilibrium with mixed strategies (lots of examples) Finding all Nash equilibria of a two-player two-strategy game Pareto optimality and social optimality definitions Exercises
Chapter 8 Modeling network traffic using game theory Traffic at equilibrium Braess s paradox (adding network capacity may slow traffic) The social cost of traffic at equilibrium (up to and without A and afterward)
Chapter 9 Auctions Types of auctions, and what in this chapter Second price auctions - bidding your true value used in ebay and in keyword based ads - formulated as a game. Truthful bidding in second-price auctions. Truthfulness is a dominant strategy First-price auctions and all-pay auctions Common values and the winner s curse Bidding strategies in first-price and all-pay auctions: models of bidder behaviour that derive equilibrium bidding strategies. Equilibrium with two bidders
Chapter 9 Auctions (continued) Equilibrium with many bidders (stop before General Distribution) Seller revenue (stop before Revenue Equivalence) Present in class 2 exercises from 9.8, from questions 1-10 (a) and (b), discuss and solve
Chapter 10 Matching markets Markets are a prime example of network-structured interaction between many agents, buyers and sellers. Matching markets is one of several models of interactions among market participants. Bipartite graphs and perfect matchings (binary preference) Optimal assignments when value is assigned to preference (centralized assignment) Prices and the market-clearing property (market assignment) Constructing a set of market-clearing prices How does this relate to single-item auctions? 2 exercises with discussion and solution after discussion
Chapter 12 Bargaining and power in networks Power in social networks power as a property of the relation between two individuals, network exchange. Examples The Nash bargaining solution The ultimatum game Stable and balanced outcomes (do not do 12.9) 2 exercises with discussion and solution after discussion
Chapter 13 The structure of the web The Web, information networks, hypertext, associative memory. Other information networks. The notion of associative memory (V. Bush) and the Memex. History and evolution of the Web. Navigational and transactional links. The Web as a directed graph (only on navigational links), SCC (strongly connected components). The bow-tie structure of the web, giant SCC The important ideas behind the Web 2.0. Designers of web sites today need to think not just about organizing information but also about the social feedback effects inherent in maintaining an audience of millions of users who might interact with the site and/or among themselves
Chapter 14 Link analysis and Web search Find a needle in a hay stack vs. Link analysis and answering searches by using hubs (send links out to multuple pages) and authorities (pages linked to by others). PageRank, a problem of rank slow leak and its solution scaled PageRank update rule Random walks on the Web graph and the probability of ending up in page X (it is equal to its PageRank!!) How link analysis is applied in modern search engines Combining also text and usage data Improving one s PageRank and the game theoretics principle Omit 14.5-14.6, discuss some exercises and solve with class
Chapter 15 Sponsored search markets (if there is time for this chapter then subsections 10.1-10.4 should be given before 15) Advertising tied to search behaviour Paying per click, setting prices through auctions, second-price auction Advertising as a matching market - assigning ad slots (buyers) to advertisers (buyers) Encouraging truthful bidding : the VCG principle Truth telling as a dominant strategy (!!!!) The generalized second-price auction Equilibria of the above and constructing the bids (price per click) Ad quality (this subsection 15.7 finishes the talk) Few exercises