This year eight PhD theses were nominated for the Victor Lesser Distinguished Dissertation Award. All nominees are of outstanding quality and made the choosing of a single winner extremely difficult. After many discussions and even several rounds of voting it was decided that this year the Victor Lesser Distinguished Dissertation Award will have a winner and an explicit runner up.
The winner of this year's award is:
Daniel Villatoro with the thesis titled: "Social Norms for Self-Policing Multi-agent Systems and Virtual Societies", supervised by Jordi Sabater-Mir at the Autonomous University of Barcelona, Spain.
and the runner up is:
Albert Jiang with the thesis titled: "Representing and Reasoning with Large Games", supervised by Kevin Leyton-Brown at the University of British Columbia, Canada.
The award winner will give a presentation about his research during the AAMAS conference (Thursday, June 7, at 17-18).
Frank Dignum (on behalf of the Victor Lesser Distinguished Dissertation Award committee: Andrea Omicini, Toru Ishida, Ana Paiva, Jaime Sichman and Katia Sycara).
Daniel Villatoro
"Social Norms for Self-policing Multi-agent Systems and Virtual Societies".
Abstract:
Social norms help people self-organizing in many situations where having an authority representative is not feasible. On the contrary to institutional rules, the responsibility to enforce social norms is not the task of a central authority but a task of each member of the society. In recent years, the use of social norms has been considered also as a mechanism to regulate virtual societies and specifically heterogeneous societies formed by humans and artificial agents.
Firstly we sketch a game-theoretical categorization of norms that will organize the rest of the talk. This dissertation generally tackles how norms (assuming their existence) become established inside a virtual society, such as those formed entirely by virtual agents or a combination of them with human subjects.
We initially tackle how conventions emerge when dealing with different topological structures of interactions. In this part we discovered how in social networks (with the theoretical characteristics of a scale-free) conventions cannot always emerge (even in the self-interest of the whole society), because of the emergence of subconventions that are facilitated by the inherent structure of the network. The identification of the Self-Reinforcing Substructures have allowed us to develop the necessary mechanisms to reach full convergence, which was never previously reached by any other researcher in the community.
After that we explore other mechanisms that allow the imposition of social norms, such as incentives mechanisms like punishment. We present an empirical study of how different punishment technologies affect differently human subjects and we develop an agent architecture (EMIL-I-A) which behaves similarly. This architecture is not only affected by the costs associated to punishment but also by the normative message it conveys, allowing the transmission of normative messages, establishing therefore the differentiation between punishment and sanction. This hypothesis is tested using a cross-methodological approach performing human experimentation and agent based simulation.
Finally, we explore another cognitive mechanism that would allow us to explain the voluntary non self-interested compliance, Internalization, by which agents comply with norms because so doing is an end in itself, and not merely because of external sanctions, such as material rewards or punishment."
Short Bio:
Daniel Villatoro completed his PhD at the IIIA-CSIC under the supervision of Dr. Jordi Sabater-Mir. His main research interests focus on self-policing mechanisms for the adaptation of virtual environments, paying special attention to the interaction of virtual entities and human subjects.
He has collaborated with known researchers in the area such as Sandip Sen, Rosaria Conte, Giulia Andrighetto or Michael Luck, and visited important institutions such as the Santa Fe Institute. Daniel has over 20 publications in top tier conferences and specialized journals.
Moreover he has been an active member of the community acting as general chair of the EASSS09 and EASSS11, and the MABS11 Workshop, and reviewer of the most important journals (such as JAAMAS, EAAI, or ACM TAAS) and conferences (such as AAAI, IJCAI, AAMAS or ECAI).
|