| Seminar on "Social Commerce and Trust"
Date: August 6, 2012 Speaker: Dr. Bruce Spencer Abstract: Currently social networks are used by buyers to become more aware of sellers and their goods. Increasingly, they are used with three other market mechanisms: Buyers use social networks to form buyer groups that leverage purchasing power. Sellers incentivize buyers to post recommendations on social media. Sellers or buyers compose compatible items into bundles with increased value and reduced price. We discuss the potential effects of social networks on these mechanisms with a simulation that allows combinations of these mechanisms, based published metrics for authenticity. We also consider the real life implications of these market mechanisms for businesses and (mobile) consumers. Bio: Bruce Spencer is a senior researcher at the National Research Council, Canada, and the founder of its Internet Logic group, which applies the principles of logic to the usage of Internet resources. It maintains strong external ties to industrial partners, academic colleagues and standards bodies. Bruce has lead six multi-year projects that deliver software to industrial and academic partners, including Bell Sympatico/MSN, Agfa Healthcare, Kibboko.com and Canarie. He was formerly a full professor, and is now an adjunct at the University of New Brunswick, where he co-founded the automated reasoning group, and continues to supervise graduate students. 4:30pm - 4:45pm: break 4:45pm - 5:15pm: Coalition Formation Game Based Reputation System Speaker: Yuan Liu Abstract: In e-marketplaces, reputation systems are helpful for modeling the trustworthiness of sellers, especially when buyers do not have much personal experience with the sellers. However, reputation systems bear a challenging problem: the subjectivity problem, where buyers have different subjectivity in providing ratings about the same seller. Such subjectivity difference impedes the validity of the reputation systems. In this work, we propose a theoretical model based on coalition formation game to avoid the subjectivity problem. We also construct a specific coalition formation game based reputation system with a proportional allocation algorithm. In the proposed system, buyers contribute to building reputation systems within their coalitions and benefit from the allocated utility which is created by the reputation systems. Our theoretical analysis shows that buyers with the same subjectivity have incentives to form one coalition to avoid the subjectivity problem, if specific conditions are satisfied. 5:15pm - 5:45pm: A General Trust Framework Based on Diffusion Theory and its Instantiation on Modeling Trustworthiness of Advisors Speaker: Hui Fang Abstract: Trust mechanisms have been considered as an important part for the successful design of online environments. In this paper, we first propose a general trust framework called DiffTrust derived from diffusion theory in social science to address a user (agent)'s trust building process in online environments. In DiffTrust, a user's trust building in a system is considered as a trust diffusion process, and both the spatial and temporal information are highlighted. DiffTrust is both evolutionary and robust, and emphasizes trust's dynamic characteristic (intrinsic nature of trust). It can also cope with different kinds of trust in different scenarios (contexts). We then apply DiffTrust framework to address the trustworthiness of advisors (providing feedbacks) for each buyer in e-marketplaces. Specifically, we propose a computational model to build context-aware trust social network for each buyer, which models the trustworthiness of advisors based on shared interactions between the buyer and each advisor, susceptibility of the buyer, and the diffusive influence of other buyers already adopting the corresponding advisor. And, the temporal effect is taken into account. We conduct experiments on real datasets to verify its effectiveness by comparing it with other competing approaches. 5:45pm - 6:30pm: Discussions and meeting with Dr. Bruce Spencer |