Cover image for Mind at light speed : a new kind of intelligence
Title:
Mind at light speed : a new kind of intelligence
Author:
Nolte, D. D.
Personal Author:
Publication Information:
New York : Free Press, [2001]

©2001
Physical Description:
xi, 303 pages : illustrations ; 25 cm
Language:
English
ISBN:
9780743205016
Format :
Book

Available:*

Library
Call Number
Material Type
Home Location
Status
Item Holds
Searching...
TA1630 .N65 2001 Adult Non-Fiction Non-Fiction Area
Searching...

On Order

Summary

Summary

"Mind at Light Speed" is the ultimate story of artificial intelligence -- and how it will revolutionize the world in which we live. David Nolte, Professor of Physics at Purdue University, and his research colleagues have devoted their lives to building computers that use light instead of electricity for computation. These machines will be so fast and efficient that they will generate a new kind of intelligence, which for centuries has only been dreamed of by visionaries and mystics. That science fiction is now real.

Since the invention of the laser right up to the recent news that a light particle was halted as if it were a baseball caught in a mitt, we have watched the manipulation of light grow ever more sophisticated and ingenious. That line of research is about to pay off more dramatically than we could have hoped. Nolte's and his colleagues' simple yet revolutionary idea is that, while electric charge may have always done the calculating in our computers -- "and" inside our brains --we can build machines that compute with light, with photons, instead. Such optical computers would operate at light speed and in the process red


Reviews 1

Publisher's Weekly Review

Purdue University physics professor Nolte charts the future of computing in an excellent book designed to appeal to the specialist as well as the general reader. Someday, Nolte writes, "luminous machines of light made from threads of glass and brilliantly colored crystals that glow and shimmer, pulsating to the beat of intelligence" will be commonplace. In other words, clunky electronics that rely on electrons to regulate the flow of information will be replaced by fiber optics that use laser beams to regulate other information-encoded laser beams. But with this generation of machines already at hand, Nolte envisions a further departure: a computer's "consciousness" will be driven by quantum physics. Light computers will use the qubit, the quantum version of the binary bit, to process all answers to a question simultaneously, and could use holographic symbols rather than binary systems as units of information. Nolte supports his case with a broad foundation of argument that includes chapters drawing together the history of quantum physics, the mechanics of human sight and intelligence, linguistics and semiotics. He also gives compelling insights into the nature of human thought and the technology that, he says, could far exceed it. Nolte's optimism poses a striking contrast to Roger Penrose's contentious and superb The Emperor's New Mind and subsequent Shadows of the Mind, which argued that computers cannot rise to the level of human thought. Nolte sounds at times like a seer caught up in rapture at the shape of things to come, but his research is cutting edge and his predictions forceful. (Dec.) (c) Copyright PWxyz, LLC. All rights reserved


Excerpts

Excerpts

Preface Light is the quintessential messenger. It travels faster than anything else can travel. It weighs nothing, and costs almost as little to make. A million rays of light carrying a thousand colors can travel along with each other or through each other without interacting, carrying data and commands between millions of locations. This capability is called the parallelism of light, and it represents massive communication and computational power. With it, machines of light will do a million things at once. Indeed, visual information is streaming into our eyes and hitting our retinas at a rate over a billion bits per second. The need to feed the information-hungry eyes is one of the principal forces driving the exponential growth of information carried by the optical Internet. The desire for ever more sophisticated visual content puts demands on the Internet that can be solved only by using the parallelism of light moving in transparent glass fibers. The optical revolution that began at the end of the twentieth century was launched by the human eye, but it will move far beyond serving simple human senses. The power of parallelism is the basis of whole new classes of machines of light. These will become ever faster. But faster intelligence is not a revolution -- it is just more of the same. The real revolution will come when all-optical intelligence distributes itself over optical networks with light controlling light. The net will have a multiplicity of interconnections that rivals the complexity contained in human minds. This book is a journey. It begins with the oldest (yet the most sophisticated) machine of light: the human eye. It ends by exploring the quantum optical computers that will be realized late in this new century. To reach that end it will take three generations of machines of light, which I introduce in chapter 2. The first is the Optoelectronic Generation that we are using now, supporting the optical Internet. The second is the All-Optical Generation, when light will control light and images become the units of information. The third and last is the Quantum Optical Generation, when quantum effects that defy classical logic will be used to transport (even teleport) quantum information and perform "uncomputable" computations in the wink of an eye. What will these machines of light look like? How will they manipulate information? Will they have intelligence? These are some of the questions that I ask when exploring the structure of visual intelligence in chapter 3. The neural networks of the human eye and brain are the most sophisticated image-processing machines that we know. They provide the starting point for artificial networks in optical machines. Detecting spatial features in a crowded scene is one of the simplest things our eyes and mind can do, yet it is one of the most challenging problems to artificial intelligence. Why? Our neurons are so slow. Our rate of reading is millions of times slower than the processing rates of our simple PCs. How can such slow machines as our minds perform so well? These questions are explored through chapters 4 and 5. Which raises a tantalizing question: What if machines of light could tap into the parallelism of light without being hampered by human limitations? This is the challenge for the three generations of the machines of light. The Optoelectronic Generation, supporting the bandwidth explosion on the Internet, is described in chapter 6, followed by the migration to optical intelligence, described in chapter 7, when information as light controls light and intelligence on the Internet becomes distributed over more intelligent nodes than there are neurons in the human brain. What kind of intelligence will that represent? To tap fully the parallelism of light requires that images become the units of information. What if the bit, a simple yes-no, is replaced by an entire image as the "unit" of information? In such machines, one image will tell another image what to do. Chapter 8 describes holographic machines that store information optically inside brilliantly clear crystals and that dream visually. At the apex of optical evolution driven by parallelism will be the quantum optical computer. Nothing we have ever experienced can prepare us for the astronomical shift that quantum technology represents. What will become possible when quantum neural networks connect together through quantum teleportation across the Quantum Internet? The entire network will become a macroscopic quantum wavefunction. Will it be conscious? These are questions raised on our journey from hieroglyphics, one of the first optical languages invented at the dawn of civilization, to the holographic quantum computers of the new century. Plug in your eyes. Chapter One The Glass Bead Game Visual Knowledge THIS SAME ETERNAL IDEA, WHICH FOR US HAS BEEN EMBODIED IN THE GLASS BEAD GAME, HAS UNDERLAIN EVERY MOVEMENT OF MIND...WITH THE DREAM OF PAIRING THE LIVING BEAUTY OF THOUGHT AND ART WITH THE MAGICAL EXPRESSIVENESS OF THE EXACT SCIENCES. Hermann Hesse, The Glass Bead Game, 1942 Our lives are filled with images. Every day we see signals, read signs, and learn symbols. We find our way with maps, look for news and bargains in newspapers, calculate our bills and taxes. We turn printed music into wonderful sounds, often without conscious effort. Icons fill our churches, synagogues, and mosques, dot our computer screens, and are sprawled on billboards, on clothing and advertisement pages. Architecture and art conspire to fill our views with meaningful shapes and form. Pictures capture an instant in time, while movies and video entertain us with visual motion. We live in a visual world, full of information transmitted by light. Writing is the verbal made visual, put into physical form as combinations of letters incised in clay or stone, or on a printed page, or on a computer screen. We understand the words as we see them because the visual impressions on our retinas ultimately connect with the language centers of our human brain. Similar mental processes occur for mathematics and music. Mathematical symbols represent something specific, some thought or quantity, or a relationship between abstract concepts. Notes on a score represent pitch and duration. We see the symbols, visually, and we know what they represent. But how do we know? More neurons are used to transmit visual sensory information to the brain than for any of the other senses. The retina, the light-sensitive layer at the back of the eye, has a special status above all other sensory organs as a direct extension and outgrowth of the brain itself. In the early fetus, portions of the nascent forebrain extend forward to develop eventually into the eyes and retinas. The mature retina is composed of multiple interconnected layers of neurons that take the images coming into the eye and begin to analyze them for spatial relationships. After the retina performs considerable neural computation, the visual information is coded into electrochemical impulses that are the language of the brain and of intelligence. This all happens before the signals are even transmitted to the visual cortex of the brain. Thus, we already have intelligence in our eyes. Natural selection has driven the evolution of organisms that have sophisticated image acquisition and analysis capabilities because the visual image is an information format with significant advantages for survival. None of the other senses can give the type of explicit spatial information that eye and vision can, especially the ability to provide information about distant predators or prey. What is it about the image that makes it so informative? A visual image such as a picture is a parallel data structure. That is, all points in the picture or scene either emit, transmit, or reflect light all at the same time -- in parallel. A single square centimeter of a picture has well over a million points of light, all emitting together. When the image falls on the retina, a million micron-size receptive fields in the retina process and send information simultaneously to the brain. The parallel data rate on the optic nerve is over 1 megabyte per second -- comparable to the data transfer rate of a computer hard drive. By considerable contrast, during oral communication the ear receives words one at a time -- that is, serially -- at the rate of only a few bytes per second. The parallel processing capability of the eyes, and their highly advanced structure and function, far exceeds the information speed that the serial mode of speech and ears can offer. Is a picture worth a thousand words? What information is conveyed when a picture is seen compared to when a thousand words are read? Images carry texture and form, and above all provide spatial relationships "at a glance." They present a whole world to which language can only allude. Inevitably, we must ask: How can we better use the advantages of light and image? THE GLASS BEAD GAME The search for a universal language of visual symbols that can express the essence and subtleties of all knowledge has had a long and energetic history since the English philosopher Sir Francis Bacon (1561-1626) first suggested such a project. One of the early proponents of universal visual languages was the brilliant and influential German philosopher Gottfried Wilhelm Leibniz (1646-1716), who envisioned a universal "character" that could express all knowledge and act as an instrument of discovery to uncover new concepts and truths. At the time, hints that a universal language might be possible came from the growing awareness of Chinese character writing, as well as the rediscovery of Egyptian hieroglyphic writing. The opening up of the Far East and the growing infatuation with Egyptian artifacts presented European scholars with a treasure of mind-expanding possibilities. There was an impression (albeit false) that the hieroglyphs represented things directly, and were divorced from the peculiarities of the spoken language. The existence of these forms of writing was cited as proof that a universal language was possible, which could impart ideas and concepts directly (visually) through written characters. The difficulty was in finding an efficient means to do this. Leibniz outlined the goals of the project in his Dissertatio de arte combinatoria of 1666. Many of his activities related to the project, even his development of the calculus. He corresponded extensively with Johann Bernoulli (1667-1748), a co-inventor of the calculus, to discuss fine points of notation, striving to find the most consistent and efficient set of visual symbols to express the calculus. The standardized notation we use today for calculus was contributed almost exclusively by Leibniz, superseding the English physicist Isaac Newton's (1643-1727) clumsy notation developed at the same time. But Leibniz was unable to find the time in a busy life to tackle the problem of a more general universal language. Others took up the call. In the Twentieth Century, the psychologist Carl Jung (1875-1961) strove for universality with his symbols of transformation, and in an altogether different sphere the English logician and philosopher Bertrand Russell (1872-1970) and the English mathematician and logician Alfred North Whitehead (1861-1947) strove for the same thing with the symbolic logic they developed. Yet the most imaginative picture of the potential of light and image was painted by the twentieth century Nobel Prize winning novelist Hermann Hesse (1877-1962). The novel Die Glasperlenspiel (The Glass Bead Game) was the last novel of the author and led to his receiving the Nobel prize in literature in 1946. Hermann Hesse was born in 1877 in the southern German town of Calw by the edge of the Black Forest. As a young man, he developed a voracious appetite for literature as he worked in bookshops in Tübingen and later in Basel, Switzerland. Always a loner and outsider, he immersed himself in books and began a literary career. His first novel, published in 1904 when he was twenty-seven years old, was Peter Camenzind. This novel brought the unknown writer rapid fame and won for him the Bauernfeld Prize of Vienna. He married Maria Bernoulli (of the famous mathematical Bernoulli family) the same year. The following years brought more literary success as Hesse explored the inner turmoil of his youth in his literature. Hesse became acquainted with the theories of Carl Jung, which had a profound influence on his life and writing. In particular, Hesse was fascinated by Jung's ideas concerning dreams and universal symbols. As more novels followed, including Demian, Siddhartha, Steppenwolf, and The Journey to the East , Hesse's writing progressively looked inward, with increasing emphasis on symbolism and vivid imagery. The culmination of his inward growth appeared in 1942, at the age of sixty-five, with Das Glasperlenspiel (The Glass Bead Game). The novel describes a utopian intellectual community called the Order, which occupies itself with the study and playing of the Glass Bead Game. This monastic community exists in some future time, in a country named Castilia that is dedicated solely to the purposes of the Order and of the Game. The story of the Game, and in particular of Joseph Knecht, the Master of the Game, known as the Magister Ludi, unfolds through the narrative of a fictitious biographer. The Game is an idealized version of the universal language envisioned by Leibniz. The narrator tells how the fictitious originator of the Game "invented for the Glass Bead Game the principles of a new language, a language of symbols and formulas, in which mathematics and music played an equal part, so that it became possible to combine astronomical and musical formulas, to reduce mathematics and music to a common denominator." Within this Game, abstract concepts are represented by a set of glass beads, or icons. The visual and spatial arrangement of these beads by players allows all aspects of human knowledge to be related one to another: mathematics to art, music to astronomy, philosophy to architecture, and infinite combinations of these. The winner of the Game was the player who succeeded in weaving the most striking or surprising connections and themes among seemingly disparate concepts. Though fanciful, the Glass Bead Game is a model for the visual representation of knowledge. A quote from Leibniz in 1678, three centuries before, evokes the spirit of the Game: "The true method should furnish us with an Ariadne's thread, that is to say, with a certain sensible and palpable medium, which will guide the mind as do the lines drawn in geometry and the formulas for operations, which are laid down for the learner in arithmetic." It is easy to imagine Leibniz as the Magister Ludi conducting a sublime Glass Bead Game, the players forming threads of colored glass beads, this one representing a theorem of logic, that one an astronomical observation, and between them a musical theme branching to a mathematical formula -- all interrelated, all sharing common forms that span the breadth of human knowledge condensed into symbols. The importance of the Glass Bead Game is not the physical implementation of a set of rules that defines a game. In fact, Hesse was careful never to describe the actual rules by which the Game was played. Furthermore, it must be admitted that universal language schemes (and there have been many) all have failed by being too cumbersome and naive. However, the profound idea at the heart of the Glass Bead Game is that symbols and rules can be visual and that knowledge can be represented and manipulated visually. The Glass Bead Game is an allegory of a new optical language, the language of light and image needed to run the architecture of the future machines of light. This book explores those machines in which the language of the Glass Bead Game is about to become a reality. THE HUMAN BOTTLENECK The measure of any technology is the degree to which we live better by it. This may be posited as the principal thesis of technological humanism. One way that we live better is by reassigning human tasks to alternative agents. James Bailey, in his book After Thought , writes about successive stages of reassignment of human tasks. In the first stage, we reassigned our muscle tasks to animals. Horses provided transportation and oxen pulled our carts. The reassigned work remained on the scale of human effort -- one man could drive a few horses. The revolution came after the second stage, when we reassigned our muscle tasks to machines such as power engines and locomotives. This stage spurred the industrial revolution, where the scale of the reassigned work extended far beyond human capability, and the change in society was irreversible. In the third stage, we reassigned our mind tasks to calculators and computers, where the increased scale has been mostly one of speed rather than in ways of thinking. The fourth stage is set to begin when we succeed in reassigning our conscious tasks to our mental machines. The way these mental machines think will be the revolution, going beyond mere speed. Some of these machines will be visual. A goal of early artificial vision systems was the detection of features in an image, such as straight edges in a photograph, or the detection of a unique character in a crowd -- like finding the cartoon character named Waldo hidden amongst visual chaos in cartoon books. The machines that perform these image recognition tasks have drawn heavily on the mechanisms of human visual perception as a model of a working visual recognition system. As we see how these machines work, it is possible to envision where machines have a chance to go beyond human capabilities. Our critical weakness in visual communication is the speed limit on comprehension as we read. The data structure of images allows us to see with speeds comparable to the data rate transferred from a computer hard drive; so why do we read so much more slowly than, say, a scanner can scan a page? For instance, you will spend about two minutes looking at this page of this book, while a laser scanner can scan it in a few seconds, and a digital camera can capture it in one thousandth of a second. Our limitations were created by evolution. The human brain has "co-evolved" side by side with the evolution of verbal communication. Despite the superior processing speed of the eye and vision over ear and hearing for receiving information, there is no exclusive biological optical equivalent to the vocal chords. We cannot send visual information to another person in a way that utilizes fully the data capacity of the eye. Sign language is certainly one way in which language can be sent visually. This is a highly efficient and expressive manner of communication, possessing favorable qualities that have no equivalent in spoken language. But one of the important findings of the past decades is that the speed of sign language transmission, even among its most adroit practitioners, remains comparable to the speed of spoken words. The difficulty lies in a bottleneck -- that of comprehension. Visual language, such as reading, starts out purely visually, as signs and symbols entering the eye and transmitted as parallel electrical impulses to the visual cortex. Once in the visual cortex, the neural impulses connect to the language comprehension centers of the brain -- centers that work primarily serially, a word (or a sentence fragment) at a time. In reading comprehension, the parallel processing by which a visual field of data can be perceived at once, as when we look at a picture, is not available. Visual language (writing, mathematical notation, music scores, sign language, etc.) has always required serial processing in the human mind. That is about to change. BEYOND ANTHROPOCENTRICITY Human limitations need not be machine limitations. There is no reason to believe that the specific manner in which we process language is the only possible way. We are free to try new things, to find new ways of interconnecting neurons and nodes in structures that are different from what nature has produced. With the technologies now becoming available to us, we have an opportunity to explore and test alternative hypotheses as to how intelligence functions. The way a system "thinks" reflects the architecture of the system, which is to say, different structures "think" in different ways. Rather than trying to make computers mimic the way we think, we should find different ways of thinking altogether. Intelligent model building has already progressed through one stage -- the reassignment of mind tasks to machines. This stage started with mechanical calculators, first implemented by the French mathematician Blaise Pascal (1623-1662) and Leibniz in the seventeenth century and by the English mathematician and inventor Charles Babbage (1792-1871) in the nineteenth century. It continued with the greatly improved speed and accuracy of the first electronic calculators in the middle part of the twentieth century. Yet today, our advanced computers remain exceedingly unintelligent, and are still far outstripped in reasoning by the human brain. What is currently demonstrated as artificial intelligence is mostly made possible by the tremendous and ever-increasing computing speed of modern-day computers. The high-speed information-processing abilities of computers make up for lack of insight. They get it right, but primarily by brute force. Thus, this stage is not the revolution that some make it out to be. Mathematical computation is noticeably sped up by machines, but the calculations themselves remain the same as we would do by hand. The speed of solution has increased beyond human capability, but the structure has not. The real revolution is beginning only now as the reassigned mind tasks evolve beyond human design by using adaptive and genetic algorithms that change their own structure in response to changing inputs, without human intervention. Such algorithms have the potential to evolve into intelligent systems with no human analog -- possibly evolving beyond human comprehension. Part of this revolution in intelligent model building is the current interest in artificial neural networks based loosely on the structures of biological systems. Scientists have analyzed how the functions of the brain are distributed over neurons, and are trying to translate those structures into electronic or photonic models. Networks of nodes and their interconnections mimic some of the structure of biological networks of neurons and their synapses. However, it is an open question whether mimicking the brain's structure is sufficient to produce an "intelligent" system. Biological model building is still in an early stage of development, with significant work ahead. Furthermore, basing intelligence on the biological neurological model may not be the best solution. Newer, non-biological technologies (such as optical technology) may have more to offer. Optical technology is primed to change intelligent model building. The advantage of the optical computer is its massive parallelism. For a digital computer, the unit of information is the binary unit, known as the "bit." For every tick of the internal clock, only a handful of bits are processed even in the most advanced electronic computers. The bit does not carry much weight: only a "yes" or "no" answer. In some types of optical computer, on the other hand, the unit of information is an image. For every tick of the internal clock, the entire image, with all the information in it, is processed all at once. The parallelism of the image improves the data rate enormously. If the single advantage of optical computers were in parallel processing, then it would still not be the revolution. Higher data rates may mean more computing power but they do not represent expanded function. Optical computers promise something more. They promise abstract and associative "reasoning," based on images and symbols that exist within a language of spatial and spectral (color) relationships. For an optical computer, a picture may well be worth more than a thousand words. A picture may be the program that tells the computer what functions it must perform and what concepts must be employed. The rudimentary and specialized optical computers built so far in the laboratory are not the flexible, programmable machines that will be able to make conjectures and leaps of imagination. Some of the current limitations have been in materials and in technology. More importantly, a fundamental new architecture must be designed for the next-generation machines of light. The new architecture will need a new language in which to express itself. It must be an optical language, where images are like words and the grammar is made up of visual projections and associations; we will need something akin to the language of the Glass Bead Game. THE ARCHITECTURE OF LIGHT Three basic themes are crucial to understanding our own intelligence and how we can go beyond with the next-generation machines of light. First, all manners of human communication, whether audible through speaking and listening, whether visual through writing and reading or the use of sign language by the deaf, or whether tactile through the use of Braille, share a common rate for comprehension that is limited by biological physiology. I call this the Human Comprehension Bottleneck. All communication channels must pass through the same cognitive centers of the brain to provide the ability to make informed decisions. Second, images and words cannot be equivalent (even when considering the same written and spoken word), because the visual and auditory channels use different media that initially access different parts of the brain. Specifically, the visual channel is a massively parallel data channel which has unique attributes and advantages that far outstrip verbal and serial communication -- if only they can be accessed. I call this the Parallel Advantage of Light and Image. Third, and finally, the biological and physiological limitations underlying the Human Comprehension Bottleneck need not be machine limitations. We can build machines that can perform functions that we cannot. Speed alone is not such an advance. Rather, new machine architectures will utilize information in ways that go beyond human capabilities. This process of searching for new visual architectures based on a visual language of spatial and spectral relationships may allow machines to find new ways of thinking that utilize the Parallel Advantage of Light and Image. That new computational structure will be the Architecture of Light, the guiding principal that shapes the three generations of the machines of light. Excerpted from Mind at Light Speed by David D. Nolte. Copyright © 2001 by David D. Nolte. Excerpted by permission. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.

Table of Contents

Preface
1 The Glass Bead Game Visual Knowledge
2 Three Generations of Machines of Light The Paradigm
3 The Structure of Visual Intelligence When a Rose Is a Rose
4 Mechanisms of Vision From the Retina to the Brain
5 The Speed of Seeing Rates of Human Visual Perception
6 Communicating at the Speed of Light The Optical Internet
7 The All-Optical Generation The Control of Light by Light
8 The Telling Image Holographic Computers and the Architecture of Light
9 The Age of Entanglement Quantum Teleportation and Cryptography
10 Quantum Computing the Uncomputable Spinning Coins and Qubits
Epilogue: The Glass Bead Game in a Stream of Light
Appendix A Table of Rates and Measures
Appendix B Optical R&D Companies
Glossary
Bibliography
Notes
Index