"How do we know what we know about what we know?"
AI has traditionally thought of mental competence as deriving from internally represented knowledge, and has devoted great effort towards providing notations, methodologies and semantic theories for such representations. This representationalit assumption has often been attached, but seems to be surviving. This talk will revisit some of the basic assumptions of representationalist AI and critique them from an empirical perspective, focusing on the question: what kind of empirical test could we use to tell if our representations are right or wrong? Along the way we will touch on a number of topics, including the notorious frame problem, some rival ontologies for time and change, the evolution of consciousness, the fact that it requires a graduate education to be worried by paradoxes, and the semantic web. The talk has no technical conclusion, but makes some tentative suggestions for future research.
for more detailed talk description.
Please email me ([email protected]), if you are interested in coming, as I'll need to make arrangements for parking/entrance into ISI.
Added by mote on March 8, 2005