- Shopping Bag ( 0 items )
"IEEE Press is proud to present the first selected reprint volume devoted to the new field of intelligent signal processing (ISP). ISP differs fundamentally from the classical approach to statistical signal processing in that the input-output behavior of a complex system is modeled by using "intelligent" or "model-free" techniques, rather than relying on the shortcomings of a mathematical model. Information is extracted from incoming signal and noise data, making few assumptions about the statistical structure of signals and their environment.
Intelligent Signal Processing explores how ISP tools address the problems of practical neural systems, new signal data, and blind fuzzy approximators. The editors have compiled 20 articles written by prominent researchers covering 15 diverse, practical applications of this nascent topic, exposing the reader to the signal processing power of learning and adaptive systems.
This essential reference is intended for researchers, professional engineers, and scientists working in statistical signal processing and its applications in various fields such as humanistic intelligence, stochastic resonance, financial markets, optimization, pattern recognition, signal detection, speech processing, and sensor fusion. Intelligent Signal Processing is also invaluable for graduate students and academics with a background in computer science, computer engineering, or electrical engineering.
About the Editors
Simon Haykin is the founding director of the Communications Research Laboratory at McMaster University, Hamilton, Ontario, Canada, where he serves as university professor. His research interests include nonlinear dynamics, neural networks and adaptive filters and their applications in radar and communications systems. Dr. Haykin is the editor for a series of books on "Adaptive and Learning Systems for Signal Processing, Communications and Control" (Publisher) and is both an IEEE Fellow and Fellow of the Royal Society of Canada.
Bart Kosko is a past director of the University of Southern California's (USC) Signal and Image Processing Institute. He has authored several books, including Neural Networks and Fuzzy Systems, Neural Networks for Signal Processing (Publisher, copyright date) and Fuzzy Thinking (Publisher, copyright date), as well as the novel Nanotime (Publisher, copyright date). Dr. Kosko is an elected governor of the International Neural Network Society and has chaired many neural and fuzzy system conferences. Currently, he is associate professor of electrical engineering at USC."
ISP differs fundamentally from the classical approach to statistical signal processing in that the input-output behavior of a complex system is modeled by using "intelligent" or "model-free" techniques, rather than relying on the shortcomings of a mathematical model.
Humanistic Intelligence (HI) is proposed as a new signal processing framework in which the processing apparatus is inextricably intertwined with the natural capabilities of our human body and mind. Rather than trying to emulate human intelligence, HI recognizes that the human brain is perhaps the best neural network of its kind, and that there are many new signal processing applications, within the domain of personal cybernetics, that can make use of this excellent but often overlooked processor. The emphasis of this chapter is on personal imaging applications of HI, to take a first step toward an intelligent wearable camera system that can allow us to effortlessly capture our day-to-day experiences, help us remember and see better, provide us with personal safety through crime reduction, and facilitate new forms of communication through collective connected HI. The wearable signal processing hardware, which began as a cumbersome backpack-based photographic apparatus of the 1970s, and evolved into a clothing-based apparatus in the early 1980s, currently provides the computational power of a UNIX workstation concealed within ordinary-looking eyeglasses and clothing. Thus it may be worn continuously during all facets of ordinary day-to-day living; so that, through long-term adaptation, it begins to function as a true extension of the mind and body.
Signals, Image processing, Human factors, Mobile communication, Machine vision, Photoquantigraphic imaging. Cybernetic sciences, Humanistic property protection, Consumer electronics
What is now proposed, is a new form of "intelligence" whose goal is to not only work in extremely close synergy with the human user, rather than as a separate entity, but more importantly, to arise, in part, because of the very existence of the human user. This close synergy is achieved through a user-interface to signal processing hardware that is both in close physical proximity to the user, and is constant.
The constancy of user-interface (interactional constancy) is what separates this signal processing architecture from other related devices such as pocket calculators and Personal Digital Assistants (PDAs).
Not only is the apparatus operationally constant, in the sense that although it may have power saving (sleep) modes, it is never completely shut down (dead as is typically a calculator worn in a shirt pocket but turned off most of the time). More important is the fact that it is also interactionally constant. By interactionally constant, what is meant is that the inputs and outputs of the device are always potentially active. Interactionally constant implies operationally constant, but operationally constant does not necessarily imply interactionally constant. Thus, for example, a pocket calculator, worn in a shirt pocket, and left on all the time is still not interactionally constant, because it cannot be used in this state (e.g. one still has to pull it out of the pocket to see the display or enter numbers). A wrist watch is a borderline case; although it operates constantly in order to continue to keep proper time, and it is conveniently worn on the body, one must make a conscious effort to orient it within one's field of vision in order to interact with it.
A. Why Humanistic Intelligence
It is not, at first, obvious why one might want devices such as pocket calculators to be operationally constant. However, we will later see why it is desirable to have certain personal electronics devices, such as cameras and signal processing hardware, be on constantly, for example, to facilitate new forms of intelligence that assist the user in new ways.
Devices embodying HI are not merely intelligent signal processors that a user might wear or carry in close proximity to the body, but instead, are devices that turn the user into part of an intelligent control system where the user becomes an integral part of the feedback loop.
B. Humanistic Intelligence does not necessarily mean "user-friendly"
Devices embodying HI often require that the user learn a new skill set, and are therefore not necessarily easy to adapt to. Just as it takes a young child many years to become proficient at using his or her hands, some of the devices that implement HI have taken years of use before they began to truly behave as if they were natural extensions of the mind and body. Thus, in terms of Human-Computer Interaction, the goal is not just to construct a device that can model (and learn from) the user, but, more importantly, to construct a device in which the user also must learn from the device. Therefore, in order to facilitate the latter, devices embodying HI should provide a constant user-interface-one that is not so sophisticated and intelligent that it confuses the user. Although the device may implement very sophisticated signal processing algorithms, the cause and effect relationship of this processing to its input (typically from the environment or the user's actions) should be clearly and continuously visible to the user, even when the user is not directly and intentionally interacting with the apparatus. Accordingly, the most successful examples of HI afford the user a very tight feedback loop of system observability (ability to perceive how the signal processing hardware is responding to the environment and the user), even when the controllability of the device is not engaged (e.g. at times when the user is not issuing direct commands to the apparatus). A simple example is the viewfinder of a wearable camera system, which provides framing, a photographic point of view, and facilitates the provision to the user of a general awareness of the visual effects of the camera's own image processing algorithms, even when pictures are not being taken. Thus a camera embodying HI puts the human operator in the feedback loop of the imaging process, even when the operator only wishes to take pictures occasionally. A more sophisticated example of HI is a biofeedback-controlled wearable camera system, in which the biofeedback process happens continuously, whether or not a picture is actually being taken. In this sense, the user becomes one with the machine, over a long period of time, even if the machine is only directly used (e.g. to actually take a picture) occasionally.
Humanistic Intelligence attempts to both build upon, as well as re-contextualize, concepts in intelligent signal processing, and related concepts such as neural networks, fuzzy logic, and artificial intelligence. Humanistic Intelligence also suggests a new goal for signal processing hardware, that is, in a truly personal way, to directly assist, rather than replace or emulate human intelligence. What is needed to facilitate this vision is a simple and truly personal computational signal processing framework that empowers the human intellect. It should be noted that this framework which arose in the 1970s and early 1980s is in many ways similar to Engelbart's vision that arose in the 1940s while he was a radar engineer, but that there are also some important differences. Engelbart, while seeing images on a radar screen, envisioned that the cathode ray screen could also display letters of the alphabet, as well as computer generated pictures and graphical content, and thus envisioned computing as an interactive experience for manipulating words and pictures. Engelbart envisioned the mainframe computer as a tool for augmented intelligence and augmented communication, in which a number of people in a large amphitheatre could interact with one another using a large mainframe computer.
While Engelbart himself did not realize the significance of the personal computer, his ideas are certainly embodied in modern personal computing. What is now described is a means of realizing a similar vision, but with the computing re-situated in a different context, namely the truly personal space of the user. The idea here is to move the tools of augmented intelligence and augmented communication directly onto the body, giving rise to not only a new genre of truly personal computing, but to some new capabilities and affordances arising from direct physical contact between the computational apparatus and the human body. Moreover, a new family of applications arises, such as "personal imaging", in which the body-worn apparatus facilitates an augmenting of the human sensory capabilities, namely vision. Thus the augmenting of human memory translates directly to a visual associative memory in which the apparatus might, for example, play previously recorded video back into the wearer's eyeglass mounted display, in the manner of a so-called visual thesaurus.
II. 'WearComp' as means of realizing Humanistic Intelligence
WearComp is now proposed as an apparatus upon which a practical realization of HI can be built, as well as a research tool for new studies in intelligent signal processing.
A. Basic principles of WearComp
WearComp will now be defined in terms of its three basic modes of operation.
A.1 Operational modes of WearComp
The three operational modes in this new interaction between human and computer, as illustrated in Fig 1 are:
Constancy: The computer runs continuously, and is "always ready" to interact with the user. Unlike a hand-held device, laptop computer, or PDA, it does not need to be opened up and turned on prior to use. The signal flow from human to computer, and computer to human, depicted in Fig 1(a) runs continuously to provide a constant user-interface.
Augmentation: Traditional computing paradigms are based on the notion that computing is the primary task. WearComp, however, is based on the notion that computing is NOT the primary task. The assumption of WearComp is that the user will be doing something else at the same time as doing the computing. Thus the computer should serve to augment the intellect, or augment the senses. The signal flow between human and computer, in the augmentational mode of operation, is depicted in Fig 1(b).
Mediation: Unlike hand held devices, laptop computers, and PDAs, WearComp can encapsulate the user (Fig 1(c)). It doesn't necessarily need to completely enclose us, but the basic concept of mediation allows for whatever degree of encapsulation might be desired, since it affords us the possibility of a greater degree of encapsulation than traditional portable computers. Moreover, there are two aspects to this encapsulation, one or both of which may be implemented in varying degrees, as desired:
- Solitude: The ability of WearComp to mediate our perception can allow it to function as an information filter, and allow us to block out material we might not wish to experience, whether it be offensive advertising, or simply a desire to replace existing media with different media. In less extreme manifestations, it may simply allow us to alter aspects of our perception of reality in a moderate way rather than completely blocking out certain material. Moreover, in addition to providing means for blocking or attenuation of undesired input, there is a facility to amplify or enhance desired inputs. This control over the input space is one of the important contributors to the most fundamental issue in this new framework, namely that of user empowerment.
- Privacy: Mediation allows us to block or modify information leaving our encapsulated space. In the same way that ordinary clothing prevents others from seeing our naked bodies, WearComp may, for example, serve as an intermediary for interacting with untrusted systems, such as third party implementations of digital anonymous cash, or other electronic transactions with untrusted parties. In the same way that martial artists, especially stick fighters, wear a long black robe that comes right down to the ground, in order to hide the placement of their feet from their opponent, WearComp can also be used to clothe our otherwise transparent movements in cyberspace. Although other technologies, like desktop computers, can, to a limited degree, help us protect our privacy with programs like Pretty Good Privacy (PGP), the primary weakness of these systems is the space between them and their user. It is generally far easier for an attacker to compromise the link between the human and the computer (perhaps through a so-called Trojan horse or other planted virus) when they are separate entities. Thus a personal information system owned, operated, and controlled by the wearer, can be used to create a new level of personal privacy because it can be made much more personal, e.g. so that it is always worn, except perhaps during showering, and therefore less likely to fall prey to attacks upon the hardware itself. Moreover, the close synergy between the human and computers makes it harder to attack directly, e.g. as one might look over a person's shoulder while they are typing, or hide a video camera in the ceiling above their keyboard.
Because of its ability to encapsulate us, e.g. in embodiments of WearComp that are actually articles of clothing in direct contact with our flesh, it may also be able to make measurements of various physiological quantities. Thus the signal flow depicted in Fig 1(a) is also enhanced by the encapsulation as depicted in Fig 1(c). To make this signal flow more explicit, Fig 1(c) has been redrawn, in Fig 1(d), where the computer and human are depicted as two separate entities within an optional protective shell, which may be opened or partially opened if a mixture of augmented and mediated interaction is desired.
Note that these three basic modes of operation are not mutually exclusive in the sense that the first is embodied in both of the other two. These other two are also not necessarily meant to be implemented in isolation. Actual embodiments of WearComp typically incorporate aspects of both augmented and mediated modes of operation. Thus WearComp is a framework for enabling and combining various aspects of each of these three basic modes of operation. Collectively, the space of possible signal flows giving rise to this entire space of possibilities, is depicted in Fig 2. The signal paths typically comprise vector quantities. Thus multiple parallel signal paths are depicted in this figure to remind the reader of this vector nature of the signals.
B. The six basic signal flow paths of WearComp
There are six informational flow paths associated with this new human-machine symbiosis. These signal flow paths each define one of the basic underlying principles of WearComp, and are each described, in what follows, from the human's point of view. Implicit in these six properties is that the computer system is also operationally constant and personal (inextricably intertwined with the user). The six basic properties are:
1. UNMONOPOLIZING of the user's attention: it does not necessarily cut one off from the outside world like a virtual reality game or the like does. One can attend to other matters while using the apparatus. It is built with the assumption that computing will be a secondary activity, rather than a primary focus of attention. In fact, ideally, it will provide enhanced sensory capabilities. It may, however, facilitate mediation (augmenting, altering, or deliberately diminishing) these sensory capabilities.
2. UNRESTRICTIVE to the user: ambulatory, mobile, roving, one can do other things while using it, e.g. one can type while jogging, running down stairs, etc.
3. OBSERVABLE by the user: It can get the user's attention continuously if the user wants it to. The output medium is constantly perceptible by the wearer. It is sufficient that it be almost-always-observable, within reasonable limitations such as the fact that a camera viewfinder or computer screen is not visible during the blinking of the eyes.
4. CONTROLLABLE by the user: Responsive. The user can take control of it at any time the user wishes. Even in automated processes the user should be able to manually override the automation to break open the control loop and become part of the loop at any time the user wants to. Examples of this controllability might include a "Halt" button the user can invoke as an application mindlessly opens all 50 documents that were highlighted when the user accidently pressed "Enter"
5. ATTENTIVE to the environment: Environmentally aware, multimodal, multisensory. (As a result this ultimately gives the user increased situational awareness).
6. COMMUNICATIVE to others: WearComp can be used as a communications medium when the user wishes. Expressive: WearComp allows the wearer to be expressive through the medium, whether as a direct communications medium to others, or as means of assisting the user in the production of expressive or communicative media.
Excerpted from Intelligent Signal Processing Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
List of Contributors.
Humanistic Intelligence: "Wear Comp" As a New Framework and Application for Intelligent Signal Processing.
Adaptive Stochastic Resonance.
Learning in the Presence of Noise.
Incorporating Prior Information in Machine Learning by Creating Virtual Examples.
Deterministic Annealing for Clustering, Compression, Classification, Regression, and Speech recognition.
Local Dynamic Modeling with Self-Organizing Maps and Applications to Nonlinear System Identification and Control.
A Signal Processing Framework Based on Dynamic Neural Networks with Application to Problems in Adaptation, Filtering and Classification.
Semiparametric Support Vector Machines for Nonlinear Model Estimation.
Gradient-Based Learning Applied to Document Recognition.
Pattern Recognition Using A Family of Design Algorithms Based Upon Generalized Probabilistic Descent Method.
An Approach to Adaptive Classification.
Reduced-Rank Intelligent Signal Processing with Application to Radar.
Signal Detection in a Nonstationary Environment Reformulated as an Adaptive Pattern Classification Problem.
Data Representation Using Mixtures of Principal Components.
Image Denoising by Sparse Code Shrinkage.
About the Editors.