When can a system S be decomposed into A and B?
( Professor of Information Ecology, Science and Engineering Faculty, Queensland University of Technology )
In order to model a system, we habitually assume the system can be decomposed into component parts, and then the “the whole is the sum of the parts”. So much so, the question of how to view, understand or model a system in a non-compositional way has an oddly unfamiliar ring to it. This seminar will explore the validity of this assumption by means of analytic methods, which define a boundary between compositionality and non-compositionality. Whilst the examples used will shed light on the supposed compositionality of conceptual representation in natural language, the analytic methods are generally applicable. The seminar will close with some reflections about what the implications may hold for modeling systems.
Peter Bruza is leader of the information ecology discipline, which adopts a holistic view in researching the interaction between people, information and context.
He holds a PhD in mathematics and computer science from the University of Nijmegen, The Netherlands (1993). Since graduating, he is primarily known for his development of formal models of information retrieval. In recognition of his research contribution in this field, he was appointed programme co-chair of ACM SIGIR, the premier international conference in information retrieval in 2004, and again in 2014. He serves on the international editorial boards of the "Journal of Applied Logic", (Elsevier) "Information Retrieval", (Springer) "The Logic Journal of IGPL" (Oxford University Press) and the book series "Information Science and Knowledge Management" (Springer). His current research centres around how to develop models of information systems which are more aligned with human cognition in complex information environments.