|Current Workshop | Overview||Back | Home | Search ||
Rich Cognitive Models for Policy Design and Simulation
Author: Jeffrey M. Bradshaw
Title: We Regulate to Coordinate: Toward a Theory of Joint Activity in Human-Machine Collaboration
One of the most important prerequisites for joint activity among people—and indeed for the entire functioning of human cultures—is the presence of regulatory systems by which such activity can be coordinated. These kinds of joint activity run the gamut of life, including processes as diverse as a conversation, a couple dancing, driving on a busy highway, and military operations. In this talk, I will describe research explorations undertaken with Paul Feltovich to understand the role of regulation in joint activity among people, and how a better understanding of these processes can enhance the efforts of researchers seeking to develop effective means to coordinate the performance of consequential work within mixed teams of humans, agents, and robots. Such mixed teams are likely to become more common in the future. We are developing an ontology of regulatory systems and a set of policy-based services (KAoS) that can be applied to human-agent-robotic teamwork to help with mutual understanding and complex coordination. I will also describe current applications of this work in adversarial modeling, organizational change, and human-robotic systems.
Author: Rosaria Conte
Institute for Cognitive Science and Technology
Title: For a Knowledge Based Governance: Institution Reputation
Abstract: Social evaluation is one fundamental aspect of social cognition, the pillar on which knowledge-based society is based upon. In this paper, one aspect of social evaluation will be focused, the formation of a supra-individual entity’s reputation (SER). SER will be argued to be one of the main factors on which to resort and rely in order to favour virtuous circuits of selection and control of members on the part of a complex structure. Not only: based on the assumption that, in order to defend own reputation, it is necessary to strengthen the reputation of groups and collective entities one belongs to, one can expect good SER to provide incentive to members.But how to obtain such a result? How avoid courteous evaluation in general, and concerning supra-individual entities in particular? In this work, a social cognitive theory of reputation and its properties (Conte e Paolucci, 2002) will be extended to higher-level agencies, namely to the level of organizations and institutions. SER will be shown to include at least two distinct components, the evaluations circulating among its members (internal reputation and image), and the evaluations and rumors running throughout the external audience (external reputation). The necessity for collecting and aggregating both these components in a dynamically updating SER coefficient will be finally discussed.
Author: Frank Dignum
norms influence behavior and behavior influences norms
Author: Nigel Gilbert
Centre for Research in Social Simulation
(CRESS), University of
Title: Do we really need cognitive models for policy design and simulation?
Social simulation has advanced greatly over the last nearly twenty years and is now being considered as ripe for practical application in policy making. However, there remains a gap between the omewhat abstract and theoretically oriented models that constitute the bulk of academic work, and the rather specific, data driven models that would be required for useful policy support. One way in which it has been proposed that abstract models should be made 'richer' is by using more complicated and supposedly more adequate cognitive models of agents.
Author: Gert Jan Hofstede
Title: What makes us tick?
The presentation argues that when one models negotiations, the following factors are crucial: perceived group membership of the partner, and culture of both negotiators as a perception filter for a host of other factors. The argument runs as follows. All humans share a social psychology in which self-centered and group-centered motives compete. We also share moral systems that promote pro-social behavior and penalize egoistic behavior. And we share strongly developed powers of empathy and communication. All of this is the result of over a million years of competition between groups in which groups that were plagued by egoism lost to, were abandoned by, or were assimilated by, more cohesive groups.As a result, the distinction between a stranger and a known person is crucial to us. We are conditioned to limit moral rights to our group members. But we are quick to use fission or fusion logic to change the delineation of our ‘moral circle’.The small print of the rules of a ‘moral circle’, a community, or a society, are unique to each group. They specify who is admitted, how status and rights are allocated, what constitutes good and bad behavior, and who is cast out. We call these unwritten rules ‘culture’. The culture that we acquire as infants stays with us for the rest of our lives. It primes feelings of closeness, respect for authority, fight or collaborate logic, fear or tolerance of the unknown, indulgence or restraint. The consequence of our human nature as group animals, and of our cultures, is predictable negotiation behavior – though of course allowances must be made for personality and context. The lecture will introduce the main ways in which cultural pre-programming affects negotiation behaviour.
Author: Barry G. Silverman
Collaboratory for Advancement of the Systems
Approach, University of
Title: Can richly detailed, socio-cognitive agent models help with social dilemmas?
When we look at the social sciences we perceive fields that are silo-ed into academic sub-divisions, and scientists within these specialties applying the scientific or hypothetico-reductive method of inquiry and falsification. Yet, social dilemmas are rarely so narrowly organized. They defy reductive inquiry. A simple test of this is that one can rarely find the explanation for and/or policies to solve social dilemmas solely by piecing together individual social science theories. Social dilemmas often are systemic, generally evidence of larger underlying “diseases”, and invariably are the emergent macro-behavior of numerous teleologic actors often unwittingly interacting as they pursue their own myopic agendas and purposes. The global economic meltdown is but one example. For this reason, the post-modern view is that social science and indeed social theory are dead. This talk, however, posits an alternative conception – that of systems synthesis of scientific knowledge together with elicitation and profiling of stakeholder values and perspectives. The point is to construct agent based socio-cognitive models that can recreate social dilemmas and expose them to alternative methods of bottom up as well as top down inquiry and policy investigation. In the end, this “systems social science” approach might provide insights into social dilemmas at the same time as offering a new paradigm for the social sciences.