Informs Annual Meeting 2017

TE27

INFORMS Houston – 2017

TE28

4 - Deriving Global Criticality Conditions from Local Dependencies Les Servi, The MITRE Corporation, M/S.M230, 202 Burlington Road, Bedford, MA, 01730-1420, United States, lservi@mitre.org, Paul R. Garvey To assure greater resilience in a richly connected cyber network from exploited vulnerabilities, it is necessary to identify critical nodes and links that, without which, could compromise the ability of the system to operate as required. This talk provides a new near linear time way to find the globally critical nodes and links from local dependencies between elements that comprise the system. 350C Social Media Analytics III Invited: Social Media Analytics Invited Session Chair: Sriram Somanchi, University of Notre Dame, Notre Dame, IN, 46556, United States, somanchi.1@nd.edu 1 - Apply Text Mining Technology to Assist Design of Mobile Game Pokémon Go Shih-Hsien Tseng, Chung Yuan Christian University, 200 Chung Pei Road, Chung Li District, Taoyuan City, 32023, Taiwan, tsengsh@mail2000.com.tw, Hsiao-Ying Chen, Chung-Yu Lao With the prevalence of Mobile Games, numbers of Games Developers and the players are growing significantly. Many players like to discuss experiences and suggestions at the forum, and therefore online reviews from users become an important source of information for the game developer. In this research, we apply LDA Topic Model to analyze the opinions from users of Mobile Game Pokémon Go on PTT which is the largest forum in Taiwan to help game developers enhance the design quality. 2 - Opinion Dynamics with Bounded Openness to Ideas Youzong Xu, Xi’an Jiaotong-Liverpool University, 111 Ren’ai Road, Business Building 423B, Suzhou, 215123, China, xu.youzong@wustl.edu, Bo Li We study opinion dynamics with agents who have bounded openness to ideas. We characterize necessary and sufficient conditions under which agents with bounded openness to ideas can reach a consensus, or become segregated groups that never listen to each other. We then apply our findings to show why, when social media is providing a platform that people can easily be connected with each other, social media may lead to segregation and even polarization of people’s opinions. 3 - Exploring the Social Media Prevalence and Consequence of “Sharing without Reading” TE27 Extant research on social contagion assumes that information spreads as sharers deliberately process external information, then decide whether or not to share it; as each sharer processes this information, s/he is “infected” with new knowledge. However, emerging evidence suggests that this is not always the case; for example, 16.2% of shared links on Twitter have more retweets than clicks (Zarrella, 2013). We propose that “sharing without reading” represents a distinct phenomenon in which information “carriers” spread content without being infected by it. Evidence from four studies suggests that sharing without reading leads to increases in subjective, but not objective, knowledge. 4 - Exploring Social Influence from Social Opinion to Social Network Shu-Ping Tsai, Graduate Student, National Sun Yat-Sen University, No. 70, Lienhai Rd., Kaohsiung, 80424, Taiwan, Kaohsiung, Taiwan, m054020017@Student.nsysu.edu.tw With the rapid development of social Web and electronic commerce, the potential consumers seek for product-related information before making a purchase decision from various online resources. Among all kinds of online resources, customer review provides critical product features and discussion posts play a vital role for peer communication. However, there is a lack of integration of these rich online resources. This study particularly investigates customer review and interpersonal information from discussion boards and explores how these two social resources associated with each other to improve the insight into each user’s preference with a social role they act. Frank Zheng, Doctoral Student, UT-Austin, Austin, TX, United States, jianqing.zheng@mccombs.utexas.edu

350D Mechanism Design and Pricing

Invited: Auctions Invited Session

Chair: Gabriel Weintraub, Stanford Graduate School of Business, Stanford, CA, 94304, United States, gweintra@stanford.edu 1 - Evaluating Strategic Forecasters Rahul Deb, University of Toronto, 150 St George Street, Toronto, ON, M5S.3G7, Canada, rahul.deb@utoronto.ca Motivated by the question of how one should evaluate professional election forecasters, we study a novel dynamic mechanism design problem without transfers. A principal who wishes to hire only high quality forecasters is faced with an agent of unknown quality. The agent privately observes signals about a publicly observable future event, and may strategically misrepresent information to inflate the principal’s perception of his quality. We show that the optimal deterministic mechanism is simple and easy to implement in practice: it evaluates a single, optimally timed prediction. We study the generality of this result and its robustness to randomization and noncommitment. 2 - The Scope of Sequential Screening with Ex-post Participation Constraints Gabriel Weintraub, Stanford Graduate School of Business, Knight Management Center, Stanford University, 655 Knight Way, E364, Stanford, CA, 94304, United States, gweintra@stanford.edu, Dirk P. Bergemann, Francisco Javier Castro We study the classic sequential screening problem with the additional constraint that the seller is required to satisfy the buyer’s ex-post participation constraints. This setting arises in several practical applications, such as the market for display advertising. First, we provide a necessary and sufficient condition under which the static contract (buyers are not sequentially screened) is optimal for general distributions. When this condition fails, we fully characterize the optimal dynamic contract, which randomizes the low type buyer and exhibits other novel properties. 3 - Understanding and Pricing on a Network with Latent Agents Alexandre Belloni, Duke University, Fuqua School of Business, 100 Fuqua Drive, Durham, NC, 27708, United States, abn5@duke.edu, Baris Ata, Ozan Candogan A recent literature on social networks has focused on understanding how a seller can use the available social network information to improve decisions. This paper investigates how a seller estimates the impact of the social network on agents’ consumption and makes pricing decisions when there are additional latent agents which are not observed by the seller. These latent agents can purchase the same product from a different channel and can influence the agents in the observable part of the network. We establish structural results, compute optimal pricing rules based on observable quantities, and estimate the effective latent network.. Variants with and without random shocks are considered. 4 - Local Computation Mechanism Design Shai Vardi, California Institute of Technology, Pasadena, CA, United States, svardi@caltech.edu Local computation mechanisms (LCMs) are game theoretic mechanisms that produce small parts of a solution extremely quickly, without any pre-processing. LCMs reply to queries in time sub-linear in the input size, so that the replies to different queries are consistent with a single global solution that satisfies some required game-theoretic properties. LCMs are useful in scenarios when the market is so big that mechanisms that require access to all of the input and produce the entire output are impractical. We design local computation mechanisms for several classical game-theoretical problems: stable matching, job scheduling and unit-demand combinatorial auctions.

420