UCT CS Research Document Archive

Using Bayesian Agents to Enable Distributed Network Knowledge: A Critique

April, Kurt, Anet Potgieter and Richard Cooke (2005) Using Bayesian Agents to Enable Distributed Network Knowledge: A Critique. In Proceedings 4th International Critical Management Studies Conference, Cambridge University.

Full text available as:
PDF - Requires Adobe Acrobat Reader or other PDF viewer.

Abstract

Resource based theory (RBT) states that there are dynamic relationships between individual-resource interactions, which ultimately determine an organisation’s global behaviour in its environment. When combining in idiosyncratic, functional ways to enable an organisation’s global behaviour, we call them complementary resource combinations (CRCs), and socially complex resource combinations (SRCs) when referring to only the complex web of social interactions of these resources. Casual ambiguity refers to the inherent uncertainty when the global behaviour is both tangibly evident and known, but the way in which the unique local interactions between SRCs amongst themselves and the environment ultimately contribute to the global behaviour is often unclear.
Thus, in order to understand social complexity and causal ambiguity of an organization, the SRCs emergent behaviours and the causal local interactions must be observed over time, and the inter-relationships must be identified and made tangible. In our research, we use simple agents to observe the local and global behaviours, to mine the inter-relationships and to model the SRCs. These agents are organized into two types of agencies: Bayesian agencies and competence agencies. The Bayesian agencies are the observers – they collectively implement specialised, distributed Bayesian networks, which enable the agencies to collectively mine relationships between emergent global behaviours and the local interactions that caused them to occur. The competence agencies are the actors – they use the beliefs of selected Bayesian agencies and perform dynamic network analysis. In dynamic network analysis, temporal data is used to predict changes that will occur in the SRCs. Most importantly, the Bayesian agencies observe and mine temporal patterns in various metrics over time, and the competence agencies evolve the SRCs.
Relationships discovered and maintained by Bayesian agencies and competence agencies are integrated into cutting-edge, resource-based topic maps (ISO 13250:2002), which provide a way of modelling the SRCs.

EPrint Type:Conference Paper
Keywords:Bayesian agencies; competence agencies; socially complex resource combinations; causal ambiguity; time-dependent emergent behaviour; topic maps
Subjects:I Computing Methodologies: I.2 ARTIFICIAL INTELLIGENCE
ID Code:221
Deposited By:Cooke, Richard/J E
Deposited On:06 October 2005