May 3, 2005
On the discussion of data
regularity à
On the Stephenson Cyber Attack
Taxonomy à
An important discussion is occurring about what type of technology might solve the current data synthesis problem that is seen in ANY ONE of the 27 100M plus per year federal IT deployments. à [54]
Your comments are appreciated. ( mailto:portal@ontologystream.com )
The BCNGroup Roadmap is our attempt to help in this critical issue. All we ask is permission to work on this problem with a small reasonable support level. Without any support, we have been able to bring forward this $750,000 deployment proposal. The program we are trying to help (one supporting US Customs) is spending 1.3 M per day, and the results are questionable.
The deployment would depend on the good faith contribution of evaluation copies of 10 commercial software systems, and would have the advise of a board of science advisors.
Our problem is that peer review is controlled by individuals like Jim Hendler and other computer science professors, and individuals whose "need" is to continue funding for a school of thought that no longer has a grounding in either the foundations of mathematics or natural science.
At least this is the claim that my group is making. Is this a reasonable claim? Yes. Does this claim ignore the hard work and dedication of those who have worked in AI? Yes again, and we do not mean to suggest that getting a PhD in AI is trivial.
What we observe is that a phone survey of 20 new PhDs in Artificial Intelligence reveals that in no case (in this random set of 20 new PhDs) does the individual know any of the cognitive science literature about the nature of human intelligence.
These individuals then serve with those who served on their PhD committee as peer reviewers at DARPA or NIST.
Is this a problem for our society? Yes
Can this be fixed?
Yes
http://www.bcngroup.org/area2/KSF/openLetter.htm