Category: Nanodot

from the good-news-bad-news dept.
Robert Trombatore writes "ZDNet's UK edition has an interesting article about AI. It's fairly well balanced and makes mention of Foresight Institute and its role in raising public awareness of emerging technologies."

CP: Here's an excerpt: "Minsky agrees that extinction at the mechanical hands of a robot race may be just around the corner, but says that developments in the field of artificial intelligence call for considered debate. He says he is encouraging artificial intelligence experts to participate in the work of the Institute." You can do that here on nanodot, and at the upcoming Senior Associate Gathering.

from the My-they-move-fast,-don't-they? dept.

The full text of a number of papers on evolutionary robotics design, artificial life (Alife), and swarm programming are available either online or as Acrobat PDF files as the "Feature of the Week" for the 31 August 2000 issue, on the website of the journal Nature.

Items of recent interest include:

• the paper by Pollack & Lipson of Brandeis University on evolutionary robotics design (see item posted here on 30 August)
• another paper titled, "Ant-like task allocation and recruitment in cooperative robots," by Krieger, et al. (no, not that Krieger)
• a commentary article on these two above
• related past items, including a paper on the complexity and evolution of artificial genomes, and another by the Swarm group at the Santa Fe Institute on the use of social insects as a model for programming robotic swarms.

Note: It is unclear how long Nature provides full access to articles in its "Feature of the Week" series; some past features several weeks old are still available, but others are not. Visit this site soon if you'd like to access or download these papers.

from the machines-with-a-past? dept.
Adam Brown writes "It has been theorized that in order for a computer to evolve its personality, it should have past memories to draw upon. While browsing the web I found the site randomaccessmemory.org which encourages members to post their life experiences and any memories that stand out in their mind. Could such a database be used as a starting seed for the implanted past memories of an AI persona? Does anybody else know of any similar resources?" CP: there's the separate question of whether seeding with human memories is a good idea…but if so, one could use biographies.

from the early-"engineering-AI" dept.
EddieWehri writes "Researchers developed a computer program that came up with designs for simple robots after being given simple parameters. Found on slashdot. http://www.canoe.ca/CNEWSScience 0008/30_robot.html" RalphMerkle explains that this refers to "recent work at Brandeis by Jordan Pollack. Evidently, this work involves a design system that uses evolutionary algorithms to design a robotic system."

Update: A spate of press coverage of this research was set off by the publication of a paper by Pollack and co-worker Hod Lipson, also of Brandeis, in the 31 August 2000 issue of Nature. Some interesting press accounts include articles in the Los Angeles Times and the New York Times.

from the Do-you-know-me? dept.

Three years after its widely-publicized chess match with Grand Master Garry Kasparov, IBM's Deep Blue computer system outranks many human celebrities in a poll of the public's familiarity with both human and machine personalities. A short article from the Associated Press on the Washington Post web site describes the study, which was conducted by Marketing Evaluation/TvQ, Inc.

According to the article, even though Deep Blue's big public splash was in 1997, about 50 percent of those polled recognized Deep Blue's name — putting it on par with actress and "Baywatch babe" Carmen Electra, and slightly above CNN talk show host Larry King.