Keynote Talks & Panels

 

Prof. Carles Sierra: "Negotiation and Search"

Most existing negotiation algorithms only work for bilateral negotiations with linear additive utility functions. Most real-world negotiations however are much more complex. I introduce in this talk a new family of negotiation algorithms that is applicable to domains with many agents, an intractably large space of possible agreements, non-linear utility functions and limited time so an exhaustive search for the best solution is not feasible. This family of algorithms is called NB3 and applies Branch & Bound search to find good plans to propose. Search and negotiation happen simultaneously and therefore strongly influence each other. It applies a new time-based negotiation strategy that considers two utility aspiration levels: one for the agent itself and one for its opponents. Also, we assume a negotiation protocol that imposes almost no restrictions and is therefore also applicable to negotiations with humans. To analyze the performance of the algorithm I will present the Negotiating Salesmen Problem (NSP): a new variant of the Traveling Salesman Problem, in which several salesmen need to negotiate with each other in order to minimize the lengths of their trajectories.

Prof. Carles Sierra is Full Professor at the Institute of Research on Artificial Intelligence of the Spanish National Research Council (CSIC) and Adjunct Professor at the University of Technology, Sydney, Australia. Recently, he has been particularly active in the area of multi-agent systems, specially on methodological aspects, negotiation and trust modelling. He has participated in around forty research projects funded by the European Commission and the Spanish Government, and has published around three hundred papers in specialised scientific journals, conferences and workshops. He is member of the program committees of around a dozen of conferences and workshops per year, and is a member of seven journal editorial boards including AIJ, JAIR and JAAMAS. He has been General Chair of the conference Autonomous Agents 2000 in Barcelona, AAMAS 2009 in Budapest, and PC chair of AAMAS 2004. He has been the local chair for IJCAI 2011 in Barcelona. More info: http://www.iiia.csic.es/~sierra

 

Sintelnet panel "Social Intelligence: Principles and Engineering"

Andrew J I Jones, Pablo Noriega, Enric Plaza (Chair: Cristiano Castelfranchi)

An exciting discussion.
It is clear that "in the design of socio-technical systems, the need for computational tractability makes it probable that there will have to be some degree of simplification of the principal social concepts involved; but ....  it is essential to achieve as clear a picture as possible of just what it is that is being simplified. .. Any computationally motivated simplifications should be carried out against the background of, and should be properly informed by, precise models of the social concepts themselves." Have been and are computer scientists sufficiently informed by conceptual models of social phenomena that they are engineering?
Is just this the right path: from theory construction, to formalization, implementation, and application? Or is there a feedback from computational modeling to the theoretical analysis. Didn't AI change some crucial theory like the classical notions of "intelligence" and of "rationality" in philosophy, psychology, and economics? Will this feedback on conceptual theories operate again for social and institutional notions (like organization, trust, norms, roles, power, ...)?
A part from a unidirectional or bidirectional model of engineering social notions (from theory to artificial systems or also from computer modeling back to theories), shouldn't we be happy with a "good-enough" but general "framework" for meta-modeling whatever socio-technical systems, artificial institutions, normative and organizational environment? Aren't "satisfycing" criteria good-enough?

Additional information