Copyright (C) ELECTRONIC ENGINEERING TIMES 1997

        Net surfers run genetic algorithms, artificial ecologies online -- Smart-technology SIG page links to live demos
By R. Colin Johnson

        Manhasset, N.Y. - The Smart Technology Special Interest Group (SIG) on TechWeb now features a page of links to online simulations that allow users to experiment with real systems in this emerging field. The SIG can be found at  this URL's

        Many of the new approaches to computation bridge different disciplines to arrive at a novel system. One example is Face-It, which harnesses evolutionary principles to mimic facial expressions. Created by professors Luigi Pagliarini and Domenico Parisi at the University of Pavia, Italy, the system automates Leonardo da Vinci's face studies using genetic algorithms (GAs) to evolve meaningful facial expressions representing a predetermined emotion.

        Fact-It is finding practical application as a "neutral" stimulus used in psychological studies. The user clicks on a face from a set of computer-generated alternatives to select the one that most closely matches a given emotion. As the user makes selections, the animated face evolves expressions based on the user's responses. Each time the user selects a face, the GA evolves another. As the user accepts new choices, the GA culls mismatches from a population of facial "genotypes." Each selected face is then "cloned" and randomly mutated to create new members of an evolving population of faces that match the user's choices.

        The artificial chromosome that represents each face requires a bit string that codifies 33 analog variables describing up to 400,000 different facial expressions. The implementation of the animated face appears much less complex than a human face, however.

        Another application with a biological aspect-in this case individuals competing for dominance-pits software robots against one another over the Internet. Called Project von Nuemann, the site offers net surfers the opportunity to create a king-of-the-hill software robot that survives grueling tests only by virtue of the machine intelligence instilled in it. The basic idea is to utilize genetic algorithms and related artificial-life technologies to evolve formidable enemies to battle each other in cyberspace. Game universes, evolved enemies and the artificial-life engine can all be downloaded from the  FTP site  ftp.krl.caltech.edu.

        Each user maintains a personal game arena inside a local computer. When a user evolves particularly interesting enemies, they can be uploaded to the FTP site for free distribution to others. What-ifs can be played out by downloading software robots and running a game session locally. Effective robots evolve by learning and remembering counter measures that worked in the past against tactics that resemble those being used in the current threat. The algorithm used by the artificial-life engine is a straight-forward genetic programming tree. Neural networks and virtual machine code are two areas of current exploration that may be included in upcoming versions of the artificial-life engine that drives Project von Neumann.

        A more cooperative take on robot interaction can be found at Boids, a site maintained by Craig Reynolds of Silicon Graphics Inc. The site uses Java to simulate coordinated animal motions, such as bird flocks and fish schools.

        The individuals in a flock are termed boids. They are able to observe their entire environment, but individual boids only react to other boids in their vicinity. Derivatives of Reynolds's work have been widely used in research labs. Reynolds originally wrote Boids in Common Lisp. A version in C++ has been written by Christopher Kline. A commercial version can be found at reality.sgi.com/employees/craig/boids.html.

        The Artificial Life Garden site, programmed by Tesh Nakamura., features a common ecosystem composed of up to 500 software creatures. Each entity is described by 50 kbytes of artificial DNA, which defines its eating, moving and sexual preferences. Users can assemble a personal "tank" of individual creatures plucked from the common ecosystem.

        The original Game of Life, where ecological simulation got its start, can be found at a site maintained by inventor John Conway. The Web implementation of the Game of Life uses HTML "checkboxes" to represent cells, which grow in number as the population reproduces. Evolve or Die (www.fusebox.com/cb/alife.html) offers a modernized version created by Robert Silverman.
 
        Other sites explore more practical uses for smart technologies by offering computational solutions for a diverse number of disciplines. The Adaptive Filter Algorithm uses neural-like learning to track and compensate for dynamically changing noise. The Java implementation, by Moritz Harteneck, reveals adaptive filtering algorithms at work.

        BrainMap extends a user's understanding of the functional anatomy of the human brain by virtue of rapid, exhaustive access to image-derived data on behavioral functions, according to Peter Fox, director of the Research Imaging Center (RIC). BrainMap offers detailed readouts of actual recordings of the brain and relates them to the brain regions known or believed to control various functions.
 
 

        Since its inception in 1994, BrainMap has grown to include over 6,000 entries from over 500 experiments reported in over 150 scholarly papers from 100 institutional research participants. BrainMap is sponsored by RIC at the Health Science Center of the University of Texas in San Antonio.

        The Visual Models of Morphogenesis: A Guided Tour depicts biological structures as acquiring complex properties through the interaction in space and time of various component modules. The complex forms and patterns of living organisms result from development, and their emergence is traditionally referred to as morphogenesis. By studying emergent processes in a biological context, engineers can gain a better understanding of the processes they must model in the real world.

        TranslateNow can automatically translate a message from English into French, German or Spanish, or vise versa, at no charge. TranslateNow demonstrates Globalink's Barcelona translation technology. By pasting up to 500 words in French, German or Spanish into a Web form page, Barcelona performs the translation automatically and e-mails it back within minutes.

        CarbNet, the Carbohydrate Neural Network (CarbNet), uses neural networks to perform online pattern recognition at the University of Georgia's Complex Carbohydrate Research Center. CCRC is using a neural network to automate the recognition of spectral data from complex chemical formulas. Any researchers with unidentified compounds can submit their spectral data to the university's neural network over the Web.

            The BITMed Web site offers a wide selection of smart technology demonstrations. BITMed online simulators use the Cellular Device Machine Development System (CDM-DS) made by InSilico Technologies Inc.

        Online simulators using the CDM-DS include a genetic algorithm, the classic cellular automata called the game of life, a neural dendritic growth simulator, a robotic simulator, wire world, the wave rule, the Hodge rule, a lymph node simulator and a hypothalamic-pituitary-adrenal simulator.

 
Copyright © 1997 CMP Media Inc.