Description:

How can we make progress towards human-level AI or, less ambitiously, perhaps "dog-level" or even "mouse-level" AI?

This workshop will examine the original AI dream of building systems that exhibit human-level intelligence, and try to define new research directions that might lead us towards this goal.

Machine learning has developed a large set of techniques over the last 10 years, to the point that they are practically applicable in circumscribed domains. These techniques include: Bayesian networks, Bayesian learning, kernel methods, supervised and unsupervised learning, and reinforcement learning. Can these techniques be integrated and applied to make progress towards the human-level AI problem? If not, where and why do they come up short? Additionally, are there new research problems that now urgently need our attention, if our long-term goal is human- or even dog-level AI?

The workshop will address these questions, and attempt to define challenge problems and new research directions towards "true" AI. We will also seek to identify major gaps in our understanding and technology, and explore the role (if any) of more traditional methods such as logical AI and connectionist architectures.


Invited Speakers: Tom Dietterich, Yann LeCun, Jitendra Malik, Stuart Russell, Richard Sutton.
See the Workshop Schedule for details.