Verwandte Artikel zu The Knowledge Web: From Electronic Agents to Stonehenge...

The Knowledge Web: From Electronic Agents to Stonehenge and Back -- And Other Journeys Through Knowledge - Softcover

Burke, James

 
9780684859354: The Knowledge Web: From Electronic Agents to Stonehenge and Back -- And Other Journeys Through Knowledge

Inhaltsangabe

In The Knowledge Web, James Burke, the bestselling author and host of television's Connections series, takes us on a fascinating tour through the interlocking threads of knowledge running through Western history. Displaying mesmerizing flights of fancy, he shows how seemingly unrelated ideas and innovations bounce off one another, spinning a vast, interactive web on which everything is connected to everything else: Carmen leads to the theory of relativity, champagne bottling links to wallpaper design, Joan of Arc connects through vaudeville to Buffalo Bill.
Illustrating his open, connective theme in the form of a journey across a web, Burke breaks down complex concepts, offering information in a manner accessible to anybody -- high school graduates and Ph.D. holders alike. The journey touches almost two hundred interlinked points in the history of knowledge, ultimately ending where it begins.
At once amusing and instructing, The Knowledge Web heightens our awareness of our interdependence -- with one another and with the past. Only by understanding the interrelated nature of the modern world can we hope to identify complex patterns of change and direct the process of innovation to the common good.

Die Inhaltsangabe kann sich auf eine andere Ausgabe dieses Titels beziehen.

Über die Autorin bzw. den Autor

James Burke is the author of several bestselling books, including Circles, American Connections, and The Knowledge Web. He is a monthly columnist at Scientific American and also serves as director, writer, and host of the television series Connections 3 on The Learning Channel. He is the founder of the James Burke Institute for Innovation in Education, whose flagship project, the Knowledge Web, an interactive website, was recently launched. He lives in London.

Auszug. © Genehmigter Nachdruck. Alle Rechte vorbehalten.

Chapter 1

Feedback

This book takes a journey across the vast, interconnected web of knowledge to offer a glimpse of what a learning experience might be like in the twenty-first century once we have solved the problem of information overload.

In the past when technology generated information overload the contemporary reaction was much the same as it is today. On the first appearance of paper in the medieval West, the English bishop Samson of St. Alban's complained that because paper would be cheaper than animal-skin parchment people would use paper to write too many words of too little value, and since paper was not as durable as parchment, paper-based knowledge would in the long run decay and be lost. When the printing press was developed in the fifteenth century it was said that printed books would make reading and writing "the infatuation of people who have no business reading and writing." Samuel Morse's development of the telegraph promised to link places as far apart as Maine and Texas, triggering the reaction: "What have Maine and Texas to say to each other?" The twentieth-century proliferation of television channels has led to concerns about "dumbing-down."

The past perception that new information technologies would have a destabilizing social effect led to the imposition of controls on their use. Only a few ancient Egyptian administrators were permitted to learn the skills of penmanship. Medieval European paper manufacture was strictly licensed. The output of sixteenth-century printing presses was subject to official censorship by both church and state. The new seventeenth-century libraries were not open to the public. Nineteenth-century European telegraphs and telephones came under the control of government ministries.

The problem of past information overload has generally been of concern only to a small number of literate administrators and their semiliterate masters. In contrast, twenty-first-century petabyte laptops and virtually free access to the Internet may bring destabilizing effects of information overload that will operate on a scale and at a rate well beyond anything that has happened before. In the next few decades hundreds of millions of new users will have no experience in searching the immense amount of available data and very little training in what to do with it. Information abundance will stress society in ways for which it has not been prepared and damage centralized social systems designed to function in a nineteenth-century world.

Part of the answer to the problem may be an information-filtering system customized to suit the individual. The most promising of the systems now being developed will guide users through the complex and exciting world of information without their getting lost. This book provides an opportunity for the reader to take a practice run on such a journey. The journey (the book) begins and ends with the invention of the guidance system itself -- the semi-intelligent agent.

There are several types of agent in existence acting like personal secretaries in a variety of simple ways: filtering genuine e-mail from spam, running a diary, paying bills and selecting entertainment. In the near future agents will organize and conduct almost every aspect of the individual's life. Above all they will journey across the knowledge webs to retrieve information, then process and present it in ways customized to suit the user. In time they will act on behalf of their user because they will have learned his or her preferences by learning from the user's daily requirements.

In the search to develop semi-intelligent agents, one of the most promising systems (and the one which starts this journey) may be the neural network. Such a network consists of a number of cells each reacting to signals from a number of other cells that in turn fire their signals in reaction to input from yet other cells. If input signals cause one cell to fire more frequently than others, its input to the next cell in the series will be given greater weighting. Since cells are programmed to react preferentially to input from cells that fire frequently rather than from those that fire rarely, the system "learns" from experience. This is thought to be similar to the way learning operates in the human brain, where the repetition of a signal generated in response to a specific experience can cause enlargement in the brain cell's synapses.

The synapse is the part of the cell that releases transmitter chemicals that cross the gap to the next cell. If sufficient chemicals arrive on the other side, they generate an impulse. If enough of these signals are generated in the target cell, they cause its synapses to release chemicals in turn, and "pass the message on." A cell with larger synapses, releasing larger amounts of chemical, is therefore more likely to cause another cell to fire. Networks of such frequently firing cells may constitute the building blocks of memory.

This theory of neuronal interaction was first proposed in 1943 by two American researchers, Walter Pitts and Warren McCulloch, who also suggested that such a feedback process might result in purposive behavior when linking the senses with the brain and muscles if the result of the interaction were to cause the muscles to act to reduce the difference between a condition in the real world as perceived by the senses and the condition as desired by the brain.

Pitts and McCulloch belonged to a small group of researchers calling itself the "Teleological Society," another of whose members was the man who invented the name for this neural feedback process. He was Norbert Wiener, and he was the first to see the way in which feedback might work in a machine, during his research on antiaircraft artillery systems during World War II. Wiener was a rotund, irascible, cigar-chomping MIT professor of math who prowled what he described as the "frontier areas" between the scientific disciplines. Between biology and engineering wiener developed a new discipline to deal with feedback processes. He called the new discipline "cybernetics." Wiener recognized that feedback devices are information-processing systems receiving information and acting upon it. When applied to the brain this new information-oriented view was a fundamental shift away from the entirely biological paradigm that had ruled neurophysiology since Freud, and it was to affect all artificial-intelligence work from then on.

Wiener first applied his feedback theory early in World War II, when he and a young engineer named Julian Bigelow were asked to improve the artillery hit rate. At the beginning of the war the problem facing antiaircraft gunners was that as the speed of targets increased (thanks to advances in engine and airframe technology) it became necessary to be able to fire a shell some distance ahead of a fast-moving target in order to hit it. Automating this process involved a large number of variables: wind, temperature, humidity, gunpowder charge, length of gun barrel, speed and height of target, and many others. Wiener used continuous input from radar tracking systems to establish the recent path of the target and use that path to predict what the target's likely position would be in the immediate future. This information would then be fed to the gun-moving mechanisms so that aiming-off was continually updated.

The system had its most outstanding successes in 1944; when British and American gunners shot down German flying bombs with fewer than one hundred rounds per hit. This was an extraordinary advance over previous performance, estimated at one hit per twenty-five hundred rounds. In 1944, during the last four weeks of German V-1 missile attacks on England, the success rate improved dramatically. In-the first week, 24 percent of targets were destroyed; in the second, 46 percent; in the third, 67 percent; and in the fourth, 79...

„Über diesen Titel“ kann sich auf eine andere Ausgabe dieses Titels beziehen.