CASCON 2018 Keynote: “It Only Took 60 Years to Solve Artificial Intelligence – That Wasn’t so Hard, Was it?”
John Tsotsos will give a keynote talk at the 28th Annual International Conference on Computer Science and Software Engineering (CASCON 2018) in Toronto, Canada, Oct. 31, 2018.
More details: https://www-01.ibm.com/ibm/cas/cascon/
Abstract:
It has been 62 years since the founding Dartmouth AI workshop. In the 1960’s, researchers at MIT believed that developing a computer program to simulate a significant part of human vision could be accomplished by employing several undergraduate students over a summer. They quickly found out they were mistaken. Fast forward to the present day. AI is everywhere, solving all manner of hard problems and triggering debates about its ethics and societal impact. In contrast, scientists studying human (and animal) vision and intelligence have made no corresponding claims for breakthroughs in their understanding of human intelligence. Although a great deal has been learned in the past 60 years to be sure, each major discovery seems to emphasize how much remains unknown. Understanding the brain and how intelligent behaviour is produced remains a major challenge. Noting that the current AI successes are often claimed to be due to mimicking brain processes, is there a disconnect here? This presentation will suggest that there has been some tacit, gradual moving of the goalposts taking place and that there is indeed a disconnect. An understanding of the brain and behavioural sciences reveals many directions for future AI research, and importantly, how far there is still to go. Dramatically, this will require the abandonment of some of the longest held computational standards. Examples from my lab will illustrate.