Understanding Consciousness with Model Abstractions
2007-7-9 到 2007-7-10
报告人 : Firmo FREIRE (PUC-Rio, Brésil)The quest for the understanding of consciousness pre dates Aristotle (384 A.C.) and has yet to produce a definition of what consciousness really is. One of the reasons for this difficulty lies in the fact that most of what goes on in the human mind is nonconscious, i.e., we are only aware of a very small part of the results or manifestations of the complex and massively parallel processes that go on in our minds. Since we cannot observe these processes, introspection fails as a means for understanding the workings of the mind. What is needed is an experimental framework for consciousness studies.
Based on the assumptions that unconscious processes are related to consciousness and that there are degrees of consciousness, we propose a strategy that has the potential to further the understanding of human consciousness and elicit knowledge for the development of Artificial Consciousness. This strategy is based on a simulation of an open ended environment of medium complexity and the expectation is that such an approach will allow for easy experimentation, hypothesis making and testing, and lead to insights to the conscious and unconscious processes of the mind.
The above approach takes into account computation and engineering theories and practices. It also relies on expertise in other fields such as psychology, neurology and philosophy. Although these latter fields are concerned mainly with the workings of the human mind, their theories, findings, and speculations on this subject are an invaluable as a source of knowledge.
The fundamental concept used in this work is that of Internal Models, i.e., the mind not only creates inner models to represent its inner workings, but also to represent the environment with which it has to interact. The above assumption is the basis for the architecture of the proposed simulator.
yasmine.charif (at) nulllip6.fr