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

On robust estimation of the location parameter

It is shown that, particularly for terrestrial cellular telephony, the interference-suppression feature of CDMA (code division multiple access) can result in a many-fold increase in capacity over analog and even over competing digital techniques. A single-cell system, such as a hubbed satellite network, is addressed, and the basic expression for capacity is developed. The corresponding expressions for a multiple-cell system are derived. and the distribution on the number of users supportable per cell is determined. It is concluded that properly augmented and power-controlled multiple-cell CDMA promises a quantum increase in current cellular capacity. >

On the capacity of a cellular CDMA system

Imaging spectrometers measure electromagnetic energy scattered in their instantaneous field view in hundreds or thousands of spectral channels with higher spectral resolution than multispectral cameras. Imaging spectrometers are therefore often referred to as hyperspectral cameras (HSCs). Higher spectral resolution enables material identification via spectroscopic analysis, which facilitates countless applications that require identifying materials in scenarios unsuitable for classical spectroscopic analysis. Due to low spatial resolution of HSCs, microscopic material mixing, and multiple scattering, spectra measured by HSCs are mixtures of spectra of materials in a scene. Thus, accurate estimation requires unmixing. Pixels are assumed to be mixtures of a few materials, called endmembers. Unmixing involves estimating all or some of: the number of endmembers, their spectral signatures, and their abundances at each pixel. Unmixing is a challenging, ill-posed inverse problem because of model inaccuracies, observation noise, environmental conditions, endmember variability, and data set size. Researchers have devised and investigated many models searching for robust, stable, tractable, and accurate unmixing algorithms. This paper presents an overview of unmixing methods from the time of Keshava and Mustard's unmixing tutorial to the present. Mixing models are first discussed. Signal-subspace, geometrical, statistical, sparsity-based, and spatial-contextual unmixing algorithms are described. Mathematical problems and potential solutions are described. Algorithm characteristics are illustrated experimentally.

/pdf/hyperspectral-unmixing-overview-geometrical-statistical-and-2lb11fr9l4.pdf

Hyperspectral Unmixing Overview: Geometrical, Statistical, and Sparse Regression-Based Approaches

Stable non-Gaussian random processes , by G. Samorodnitsky and M. S. Taqqu. Pp. 632. £49.50. 1994. ISBN 0-412-05171-0 (Chapman and Hall).

The increased deployment of gigabit Ethernet (GbE) optical LANs will be followed by the use of 10 gigabit Ethernet (10GbE) of server farm, campus backbone, inter campus links and storage networks. While 10GbE standard includes use of single-mode fiber (SMF), numerous fiber links in such networks will demand the cost-effectiveness of multimode fiber (MMF). However, just as with SMF, dispersion in MMF can severely limit the link distance. Specifically, the different propagation velocities of the distinct optical modes in MMF lead to large differential mode delay (DMD) that typically overwhelms the chromatic dispersion. For speeds approaching 10-Gbps, MMF is limited to distances ranging up to tens of meters. Moreover, in contrast to typical dispersion in singlemode fiber, DMD can vary dramatically from fiber to fiber, and even launch condition. Consequently, to compensate for dispersion in MMF, a dynamic compensation or a technique that is independent of channel response must be used. We present results of work that implement a combination of both approaches that compensates for DMD in multimode fiber links, in an efficient and robust manner.

Spatially resolved equalization and decision feedback equalization for multimode fiber links

This article presents a comprehensive overview of the research conducted within Mobile VCE's Core Wireless Access Research Programme, a key focus of which has naturally been on MIMO transceivers. The article, the second of four, provides a follow on from the first article, which dealt with MIMO channels and comprises a compilation of material that has been presented in detail elsewhere (see references within the article). In Part I MIMO channel measurements, analysis, and modeling were presented, which were then utilized to develop compact and distributed antenna arrays. This article reports parallel activities into low-complexity MIMO single-user space-time coding techniques, as well as SISO and MIMO multi-user CDMA-based transceivers for B3G systems. Also, overall B3G system designs have been suggested for beamformed multilayer CDMA-based networks

Implementable wireless access for B3G networks. II. MIMO receiver architectures [Topics in Radio Communications]

A technique of modelling the indoor radio channel that provides an insight into the propagation mechanisms involved is presented. The model by its nature provides a detailed account of the channel, with access to system parameters that are otherwise inaccessible. To demonstrate the validity of this model, results are presented which show the close match between the real environment and the simulated environment. Although gross approximations are made about the environment, the model still gives good results, highlighting the important parameters that need to be considered in channel models.

A ray tracing approach to channel modelling for the indoor environment

In this paper, the source-filter model of speech production is adapted to represent the speech signal as the superposition and convolution of a dynamic source and resonant modes. The aim is to increase the resolution of the time-instantaneous-frequency representation of each of the individual contributions of different sections of the human phonatory system. We present a framework based on dynamic mode predictors and filters, which are adapted, using gradient-based techniques, to track the modal dynamics of speech yielding a representation which is free from quasi-stationary assumptions thus allowing flexible manipulation of the speech signal. Several examples are offered including intonation modifications to illustrate the potential of the proposed approach.

/pdf/speech-analysis-and-synthesis-based-on-dynamic-modes-n96tf4y4s4.pdf

Speech Analysis and Synthesis Based on Dynamic Modes

A set of ordinary differential equations (ODEs) has been constructed to represent a continuous sigma delta modulator (C/spl Sigma//spl Delta/M) [1], i.e. a continuous equivalent of a conventional sigma delta modulator (/spl Sigma//spl Delta/M). This paper is concerned with a stability analysis of these equations with particular reference being made to values representing the gains of the feedback loop and the integrator leakage in the circuit. Conditions for the stability of the equations representing second and third order modulation are derived. The results from this process are used on a conventional simulation of sigma-delta modulation, and the applicability of this method is discussed. The criteria for stable operation are extended to n/sup th/ order sigma delta modulation and to a generalised architecture for sigma delta modulation. The results are compared to the conditions for stability derived from the work of various authors. >

Stephen McLaughlin

Papers

Spatially resolved equalization and decision feedback equalization for multimode fiber links

Implementable wireless access for B3G networks. II. MIMO receiver architectures [Topics in Radio Communications]

A ray tracing approach to channel modelling for the indoor environment

Speech Analysis and Synthesis Based on Dynamic Modes

On the stability and configuration of sigma delta modulators