There is, I think, something ethereal about i —the square root of minus one. I remember first hearing about it at school. It seemed an odd beast at that time—an intruder hovering on the edge of reality.

Usually familiarity dulls this sense of the bizarre, but in the case of i it was the reverse: over the years the sense of its surreal nature intensified. It seemed that it was impossible to write mathematics that described the real world in …

I and i

Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Deposits of clastic carbonate-dominated (calciclastic) sedimentary slope systems in the rock record have been identified mostly as linearly-consistent carbonate apron deposits, even though most ancient clastic carbonate slope deposits fit the submarine fan systems better. Calciclastic submarine fans are consequently rarely described and are poorly understood. Subsequently, very little is known especially in mud-dominated calciclastic submarine fan systems. Presented in this study are a sedimentological core and petrographic characterisation of samples from eleven boreholes from the Lower Carboniferous of Bowland Basin (Northwest England) that reveals a >250 m thick calciturbidite complex deposited in a calciclastic submarine fan setting. Seven facies are recognised from core and thin section characterisation and are grouped into three carbonate turbidite sequences. They include: 1) Calciturbidites, comprising mostly of highto low-density, wavy-laminated bioclast-rich facies; 2) low-density densite mudstones which are characterised by planar laminated and unlaminated muddominated facies; and 3) Calcidebrites which are muddy or hyper-concentrated debrisflow deposits occurring as poorly-sorted, chaotic, mud-supported floatstones. These

Phd by thesis

I have developed "tennis elbow" from lugging this book around the past four weeks, but it is worth the pain, the effort, and the aspirin. It is also worth the (relatively speaking) bargain price. Including appendixes, this book contains 894 pages of text. The entire panorama of the neural sciences is surveyed and examined, and it is comprehensive in its scope, from genomes to social behaviors. The editors explicitly state that the book is designed as "an introductory text for students of biology, behavior, and medicine," but it is hard to imagine any audience, interested in any fragment of neuroscience at any level of sophistication, that would not enjoy this book. The editors have done a masterful job of weaving together the biologic, the behavioral, and the clinical sciences into a single tapestry in which everyone from the molecular biologist to the practicing psychiatrist can find and appreciate his or

Principles of Neural Science

1. Introduction. Control System Design Steps. Adaptive Control. A Brief History. 2. Models for Dynamic Systems. Introduction. State-Space Models. Input/Output Models. Plant Parametric Models. Problems. 3. Stability. Introduction. Preliminaries. Input/Output Stability. Lyapunov Stability. Positive Real Functions and Stability. Stability of LTI Feedback System. Problems. 4. On-Line Parameter Estimation. Introduction. Simple Examples. Adaptive Laws with Normalization. Adaptive Laws with Projection. Bilinear Parametric Model. Hybrid Adaptive Laws. Summary of Adaptive Laws. Parameter Convergence Proofs. Problems. 5. Parameter Identifiers and Adaptive Observers. Introduction. Parameter Identifiers. Adaptive Observers. Adaptive Observer with Auxiliary Input. Adaptive Observers for Nonminimal Plant Models. Parameter Convergence Proofs. Problems. 6. Model Reference Adaptive Control. Introduction. Simple Direct MRAC Schemes. MRC for SISO Plants. Direct MRAC with Unnormalized Adaptive Laws. Direct MRAC with Normalized Adaptive Laws. Indirect MRAC. Relaxation of Assumptions in MRAC. Stability Proofs in MRAC Schemes. Problems. 7. Adaptive Pole Placement Control. Introduction. Simple APPC Schemes. PPC: Known Plant Parameters. Indirect APPC Schemes. Hybrid APPC Schemes. Stabilizability Issues and Modified APPC. Stability Proofs. Problems. 8. Robust Adaptive Laws. Introduction. Plant Uncertainties and Robust Control. Instability Phenomena in Adaptive Systems. Modifications for Robustness: Simple Examples. Robust Adaptive Laws. Summary of Robust Adaptive Laws. Problems. 9. Robust Adaptive Control Schemes. Introduction. Robust Identifiers and Adaptive Observers. Robust MRAC. Performance Improvement of MRAC. Robust APPC Schemes. Adaptive Control of LTV Plants. Adaptive Control for Multivariable Plants. Stability Proofs of Robust MRAC Schemes. Stability Proofs of Robust APPC Schemes. Problems. Appendices. Swapping Lemmas. Optimization Techniques. Bibliography. Index. License Agreement and Limited Warranty.

Robust adaptive control

We recently considered feedback stabilization over a signal to noise ratio (SNR) constrained channel. The results were examined for the state feedback and minimum phase cases, with links to bit-rate limited control. In this paper, we extend this analysis to non-minimum phase (NMP) plants, and show that for linear time invariant (LTI) control, NMP zeros further constrain the ability to stabilize over an SNR limited channel. This differs from the situation of bit-rate limited stabilization where NMP zeros do not play a role. We show that by considering linear time varying feedback the effect of NMP zeros in SNR limited stabilization may be eliminated.

http://ieeexplore.ieee.org/iel5/9768/30807/01426924.pdf

Stabilization of non-minimum phase plants over signal-to-noise ratio constrained channels

An adaptive control scheme is proposed for rigid-link robots where the control signal computations are performed continuously and the control coefficient computations are performed discretely. A global boundedness result is established for the resulting hybrid scheme, independent of the sampling rate. It is also shown that the position, velocity and acceleration tracking errors are of the order of the sampling period. This study brings the analysis of adaptive robot control closer to practice, where a discrete control law implementation is invariably used. >

http://ieeexplore.ieee.org/iel2/716/5019/00194596.pdf

Hybrid adaptive control for robot manipulators

Many robotic applications involvc time-varying paramctcrs during thc operation of the robot. As remarkcd in (7), the well known skew-symmetric propcrty conccrniiig tlie robot inodcl docs not ncccssarily hold, whcn sonic of tlic parantcters arc timc varying. As a rcsult the existing methods may not be appli- cable, and a new kind of control strategy is nccdcd. 'I'his prob- Icm is considered in the present paper. Using tlie properties of tlie clement by element (or Hadarniad ) product of matrices, we obtain the robot dynamics in parameter-isolated form, from which a new control scheme is developed. The stability is ad- dressed via verifying an integral inequality motivated by the I3arbalat lemma, rather than the more usual Lyapunovniethod. 'I'hc controller proposed yields zero asyniptotic tracking errors when applied to robotic systems with time varying parameters, by using a switching type control law. It is shown that the con- troller presented does not involve much on-line computation sincc it only depends on bounds on the variation of the system paramctcrs, which can be casily off-line prcdetcrniined.

http://ieeexplore.ieee.org/iel5/164/5238/00203364.pdf

Problem of robot manipulators performing path tracking tasks

For strict-feedback nonlinear systems, this paper shows that it is impossible to reduce to zero the optimal cost in the regulation of more states than the number of control inputs in the system, even using an unrestricted control effort. By constructing a near-optimal cheap control law, we characterize the infimum value of the optimal regulation cost as the optimal value of a reduced-order regulator problem. We illustrate our results with an example on the optimal control of a magnetic suspension system.

Cheap control performance of a class of nonright-invertible nonlinear systems

Identification of transformer faults using matched amplitude and phase response characteristics is presented. In frequency response analysis, fault diagnosis is done by detecting changes in frequency response tests. The vast amount of data up to 2 MHz needs to be quantified and characterized for the classification of faults. By a matched fitting with higher order transfer function up to 40, poles, zeros and their relative damping were determined in the frequency plane. Each set of data corresponding to faults is used to train a backpropagation ANN network. Out of 12 faults, the processing technique is able to identify the changes and classify the type of fault for on-line analysis. The location of fault and the severity of faults in the form of more sensitivity are brought out using this signal processing technique.

Richard H. Middleton

Papers

Stabilization of non-minimum phase plants over signal-to-noise ratio constrained channels

Hybrid adaptive control for robot manipulators

Problem of robot manipulators performing path tracking tasks

Cheap control performance of a class of nonright-invertible nonlinear systems

Diagnosis and identification of transformer faults from frequency response data