Bart Kosko

From Wikipedia, the free encyclopedia

Bart Kosko (born February 7, 1960) is a writer and professor of electrical engineering at the University of Southern California (USC). He is a leading researcher and popularizer of fuzzy logic, neural networks, and noise, and author of several trade books and textbooks on these and related subjects of machine intelligence.

Contents

[edit] Writing

Kosko’s most popular book to date was the international best-seller Fuzzy Thinking, about man and machines thinking in shades of gray, and his most recent book was Noise. He has also published short fiction and the cyber-thriller novel Nanotime, about a possible World War III that takes place in two days of the year 2030. The novel’s title coins the term “nanotime” to describe the time speed-up that occurs when fast computer chips, rather than slow brains, house minds.

Kosko has a distinctive minimalist prose style, not even using commas in his last several books. [1]. A typical example is the opening of Noise: “I hate noise.” His USC homepage [2] lists several of his newspaper op-ed and magazine articles [3] on topics that include digital privacy, abortion, cryonic suspension, digital uploading minds to computer chips (the subject of his book Heaven in a Chip), and the war on terror.

[edit] Personal background

Kosko holds degrees in philosophy, economics, mathematics, electrical engineering, and law. He is an attorney licensed in California and federal court, and worked part-time as a law clerk for the Los Angeles District Attorney's Office. He started his career in music composition, and is an award-winning composer. He left his native Kansas when, at age 18, he won a scholarship in music composition to USC based on his orchestral “Overture to the Count of Monte Cristo,” which he wrote the year before [4]. Apparently the overture, like his other symphonic works, has never been performed, although several of his smaller chamber works have been, with one of his string quartets winning a young composer’s contest while he was in high school [5]. Kosko worked as an engineer in aerospace in San Diego while working on his Ph.D. in engineering and organizing conferences in neural networks and fuzzy systems. He returned to USC as a professor in 1988. He is an elected governor of the International Neural Networks Society, and sits on the editorial board of several technical journals.

Kosko is a political and religious skeptic. He is a contributing editor of the libertarian periodical Liberty, where he has published essays on “Palestinian vouchers” [6] and the experience of taking the infamous California bar examination. The essay details what it was like to take and pass[7] the same California Bar Exam in 2006. He expanded his essay “In Defense of God” [8] in the last section of Fuzzy Thinking, just as in Heaven in a Chip he expanded the essay of the same name [9] and the essay “The Future of God” [10]. Kosko stated in these books that he has made arrangements to be cryonically suspended in liquid nitrogen when he dies. Along with artificial intelligence pioneer Marvin Minsky, he sits on the science advisory board of the nonprofit Alcor [11] cryonics foundation [12] .

[edit] Research

Kosko’s technical contributions have been in three main areas: fuzzy logic, neural networks, and noise.

In fuzzy logic, he introduced fuzzy cognitive maps [13] , [14] , fuzzy subsethood [15], additive fuzzy systems [16] , fuzzy approximation theorems [17], optimal fuzzy rules [18] , fuzzy associative memories, various neural-based adaptive fuzzy systems [19], ratio measures of fuzziness [20], the shape of fuzzy sets [21], the conditional variance of fuzzy systems [22] , and the geometric view of (finite) fuzzy sets as points in hypercubes and its relationship to the on-going debate of fuzziness versus probability [23] , [24] . More recently, he proposed a probabilistic conditioning interval [25] that has standard conditional probability as a lower bound and a new “Q” superset-based conditional measure as the upper bound. Kosko’s 1993 book Fuzzy Thinking helped popularize the fuzzy field in the early 1990s, as did his article “Fuzzy Logic” in "Scientific American" [26] in the same year.

In neural networks, Kosko introduced the unsupervised technique of differential Hebbian learning [27], sometimes called the “differential synapse,” and most famously the BAM or bidirectional associative memory [28] family of feedback neural architectures, with corresponding global stability theorems [29]. He also introduced a hybrid learning law called differential competitive learning, and showed that competitive learning converges by implicitly minimizing the mean squared error of vector quantization [30]. He showed that many spiking neural and retinal models benefit from the addition of many types of noise [31]. Kosko is also credited with organizing the first IEEE international conference on neural networks in 1987 [32], helping reestablish the second wave of neural network research in the 1980s after earlier neural-net research, based in large part on the perceptron model, fell out of favor in artificial intelligence. He discusses this conference, and his early neural research, in a chapter profile in the book Talking Nets: An Oral History of Neural Research. He is the youngest “neural pioneer” profiled in the book.

In noise, Kosko introduced the concept of adaptive stochastic resonance [33] , using neural-like learning algorithms to find the optimal level of noise to add to many nonlinear systems to improve their performance. He proved many versions of the so-called “forbidden interval theorem,” which guarantees that noise will benefit a system if the average level of noise does not fall in an interval of values [34] . The result still holds for “robust” stochastic resonance in the sense of allowing one to add impulsive noise, such as Cauchy noise, that has infinite variance [35] , [36] . He proposed and proved many versions of the so-called “forbidden interval theorem,” which guarantees that by adding electrical noise to a carbon nanotube antenna the noise helped the nanotube detect faint signals [37]. Kosko appeared on National Public Radio’s “Science Friday” on September 15, 2006 to discuss the pros and cons of noise during the promotion of Noise. The audio recording of the interview is available at the NPR site [38].

[edit] Books

[edit] External links

  • Official site
  • Profile from IEEE Spectrum, February 1996 (PDF file)
  • Q & A from Wired, February 1995 {PDF file}
  • Radio interview: “In Defense of Noise” on National Public Radio [39]
  • External link: "In Defense of God" from IEEE Expert, February 1990 {PDF file} [40]
  • U.S. Patent 5,539,769  "Adaptive Fuzzy Frequency Hopping System", filed March 1994, issued July 1996
Languages