From: Ben Goertzel (email@example.com)
Date: Tue Aug 23 2005 - 02:33:30 MDT
Yeah, a link to this DARPA program was sent to this list a long time ago...
Novamente LLC's application for funding under this program was rejected, as I learned a few weeks ago. That particular rejection came with zero explanation.
At the same time as that, I also received a rejection for an application for NIST funding for Novamente, which was more ironic and more informative because it came along with near-maximum reviews for my application in every category.
This second rejection drove home the well-known fact that such funding is allotted on political grounds as much as scientific/technical grounds. (And my own "political" connections are not null, but nor are they the best in the business....)
I note that I have successfully achieved government grants and contracts on many occasions, including recently -- but for narrow-AI work (such as bioinformatics or language processing), not AGI work.
The problem is that the pool of money allocated for AGI-ish work is so so small, that whenever any is advertised, there is a huge body of applicants, and the selection problem for the grant application review panel must be very difficult. Assessing the quality of the Novamente design, or any of its competitors, based on a brief grant application can't be easy. My application referred to various prior publications and also material available online, but I doubt the reviewers took the time to dig through all this stuff enough to fully appreciate the underlying ideas....
Unfortunately, the average quality of AGI-ish work is not very high. So when the gov't offers a relatively small amount of money and couples it with a disbursement process that involves overtaxed reviewers and a strong bias toward applications coming from individuals with appropriate political connections, the outcome is not that likely to be funding decisions that advance AGI significantly...
Of course, I will continue to try to improve the quality of my relevant political connections, and keep on writing and submitting proposals -- which of course gets to be a drag, because I'd rather be creating AGI than writing proposals (but with our current level of funding, creating AGI based on our existing and I believe workable design occurs at an unpleasantly slow, though still greater than zero, pace).
-- Ben G
Jef Allbright <firstname.lastname@example.org> wrote:
I was just discussing DARPA and AI with Peter Voss at dinner Friday
night, and we both agreed there was very little AGI in sight.
Today I saw these interesting developments at
*1 Program Objective *
The goal of the Biologically-Inspired Cognitive Architectures Program
via this BAA is to develop, implement and evaluate psychologically-based
and neurobiologically-based theories, design principles, and
architectures of human cognition. In a subsequent phase, the program has
the ultimate goal of implementing computational models of human
cognition that could eventually be used to simulate human behavior and
approach human cognitive performance in a wide range of situations.
*2 Background and Motivation*
The traditional approach to machine intelligence pursued by the
Artificial Intelligence (AI) community has provided many achievements,
but has fallen short of the grand vision of integrated, versatile,
intelligent systems. Revolutionary advances may be possible by building
upon new approaches inspired by cognitive psychology and neuroscience.
Such approaches have the potential to help us understand and model
significant aspects of intelligence thus far not attained by classic
formal knowledge modeling technology. This program seeks to advance the
design and development of computational models of human cognition based
on architectures using both approaches.
During the past 15 years, intensive research in neuroscience, enabled by
advances in experimental techniques and instrumentation, has led to
substantial progress in our understanding of the brain's physical
structure and function. It is now feasible to develop cognitive
architectures, rooted in neuroscience, which may offer important insight
into the understanding and modeling of human cognition.
ACT-R-users mailing list
This archive was generated by hypermail 2.1.5 : Tue May 21 2013 - 04:00:48 MDT