|Volume 1: No. 21|
Fraunhofer Gesellschaft will spend $280M to create 19 research institutes and outstations in the former East Germany, combining several existing groups. Institutes will include study of CS, information processing, and cybernetics. [Nature. CACM, 8/91.]
William C. Norris, Chairman Emeritus of CDC, points out that Soviet scientists are now available for joint work with others -- at one fourth the cost of U.S. scientists. The Soviet research establishment is similar to our own, and is just now opening up to commercial R&D. Germany is aggressively seeking such R&D relationships in the Soviet Union. [Computerworld, 8/12.]
The U.K.'s Science and Engineering Research Council (SERC) will limit the amount that grantees can spend in the first years of multiyear grants. A three-year grant might get only 15% in the first year, and as little as 50% in the first two. This will not only spread current funds further; it will force recipients to provide much of the hardware and startup costs. [CACM, 8/91.] (NSF grants tend to be the opposite, buying equipment up front and tapering to just salary and maintenance in later years. It's up to the proposer to set forth a rational plan, though.)
The news at IJCNN was that Japan and Europe are each planning to increase neural-network research to about $40M/year, plus an unspecified amount of military funding. DARPA, on the other hand, is planning to cut its $14M program back to $8M. U.S. government funding will then be about $2.5M for commercial research ($1.5M NSF, $1M NIH) plus $27.5 military ($8M DARPA, $12.5M Navy, $7M Army). Barbara Yoon, the DARPA program director, is fighting the cut, but DARPA Director Victor Reis and Microelectronics Technology Director Arati Prabhakar have other priorities. Barbara claims that all objectives of the initial 28-month study have been met, and it's time to start doing real applications in control, filtering, system diagnosis, forecasting, and data fusion. [R. Colin Johnson and Tom Schwartz, EE Times, 7/29.] (Has DARPA decided to fund CBR instead?)
A report by the Office of Science and Technology Policy (OSTP) cites brain research as an important input to AI/CS advances. Presidential Science Advisor D. Allan Bromley calls this the "decade of the brain," as significant as the first space launch. [CACM, 8/91.]
Robert White, Undersecretary of Commerce for Technology, is telling Congress that copyright protection should be extended to federally produced software. Without protection, industry has little interest in federal software and the developers have little incentive to help transfer their work to industry. Often it is foreign companies that benefit most from U.S.-sponsored work. Patent protection is available (due to the Stevenson-Wydler Technology Act of 1980), but few federal agencies are interested in patenting their software. The pending Technology Transfer Improvements Act would permit copyright protection for purposes of cooperative R&D agreements. Congress still needs to work out the extent of protection, including the distinction between software and databases. [EE Times, 7/29.]
David Hillman (Quality Systems Inc., Fairfax, VA) has reviewed the 8th Intelligence Community AI/Advanced Computing Symposium, March 1991. This is one I'd really like to have attended. Over 400 people came, with the principal customers being the CIA, FBI, Defense Intelligence Agency, National Security Agency, and the military-specific intelligence branches. Rick Steinheiser, CIA Office of R&D, was the chairman. Topics included text processing, document retrieval, language processing, knowledge visualization, decision aids, expert systems, neural networks, multisource fusion, imagery analysis, parallel processing, and computer security. Research included machine learning and very large knowledge-based systems. [AI Expert, 8/91.] Programs are in danger of cutbacks, and classified access is always a problem, but these people are supporting research in the good stuff. Can any Computists comment on career opportunities with these customers?
According to Jim Warren, who got it from The Privacy Journal, last year's Computerworld-Smithsonian award for innovative information technology in the government sector went to the Thailand Ministry of the Interior. The Ministry runs the world's largest citizen database, tracking ancestral history, family makeup, voting patterns, domestic and foreign travel, and social welfare. 12,000 government agents have access to the data on 55 million citizens. [MicroTimes, 8/5.]
John Perry Barlow is less dogmatic than most attendees of the SF Conference on Computers, Privacy, and Freedom -- not the Politically Correct attitude. However, I like his assertion that any new laws concerning information should restrict its use rather than its acquisition. [CACM, 8/91.] There is plenty of precedent (ham radio; trade secret laws), and the acquisition of knowledge hardly seems like a punishable offense.