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 aim of this paper is to discuss the use of adaptive subcarrier-user allocation in a multicellular environment. The paper also proposes a priority swapping based allocation algorithm that can utilise the diversity of the channel selectivity. The channel information required by the base-station for a successful detection of the received signals from the different users can be used by the proposed algorithm to enable the base-station to adaptively and speedily allocate the different subcarriers to its users such that the total throughput is maximised. This type of downlink transmission can best be likened to a frequency hopping (FH) system with the slight difference that the later uses a pseudorandom hopping pattern for each user while the former uses a dynamic-time-varying and channel-dependent orthogonal hopping pattern for each user. It will be shown that the algorithm proposed here has a very high convergence rate and achieves a diversity gain equivalent to that obtained using an optimum allocation algorithm based on the maximum likelihood criterion.

Priority Swapping Subcarrier-User Allocation Technique for Adaptive Multicarrier Based Systems

This paper presents the Time-Correlated Single-Photon Counting (TCSPC) technique applied to underwater environments in order to reconstruct three-dimensional scenes. Two different transceiver systems approaches are described. The first transceiver comprised a single-pixel monostatic scanning unit, which used a fiber-coupled silicon single-photon avalanche diode (SPAD) detector, and a fiber-coupled supercontinuum laser source used in conjunction with an acousto-optic tunable filter (AOTF) for wavelength selection. The experiments were performed using the supercontinuum pulsed laser source operating at a repetition rate of 19.5 MHz, fiber coupled to the AOTF in order to select one operational wavelength, tuned for best performance for the level of scattering of the particular underwater environment. Laboratory-based experiments were performed using average optical powers of less than 1 mW and depth profiles were acquired at up to 8 attenuation lengths between the transceiver and target. The second transceiver system was based on a complementary metal-oxide semiconductor (CMOS) SPAD detector array in a bistatic configuration. It comprised an array of 192 × 128 SPAD detectors, with each detector having an integrated time to digital converter, and a laser diode operating at a wavelength of 670 nm, a repetition rate of 40 MHz, and average optical power up to 9 mW. The experiments demonstrated the recovery of intensity and depth profiles associated with moving targets at up to 4 attenuation lengths. Using data from both systems, various image processing techniques were investigated to reconstruct target depth and intensity profiles in highly scattering underwater environments.

Underwater depth imaging using time-correlated single-photon counting at video frame rates

We propose a novel, ultra-fast, iterative equalizer that has lower complexity than conventional equalizers but has performance that approaches the optimum receiver. Using this technique we demonstrate significantly increased reach for 10 Gb/s over differential modal delay limited multi-mode fiber.

http://ieeexplore.ieee.org/iel5/9410/29870/01363459.pdf

Equalization via iterative detection in multi-mode fiber optic links

Empirical mode decomposition (EMD) has lately received much attention due to the many interesting features exhibits However it lacks a strong theoretical basis which would allow a performance analysis and, hence, the enhancement and optimization of the method in a systematic way In this paper, the investigation of EMD is realized by means of a genetic algorithm (GA) based optimization The above study results in novel directions for both the performance enhancement and also the theoretical investigation of the method

On the performance enhancement of the Empirical Mode Decomposition

A method of assessing a communication route comprising a plurality of links between nodes in a mobile ad-hoc network comprises calculating the two-hop residual bandwidth of each node I of the route as B I ( t ) = B − ∑ J ∈ N ( I ) B ( J ) φ 
where B is the raw channel bandwidth, the summation is the overall consumed bandwidth from node I 's two-hop neighborhood nodes J ∈ N(I), and ϕ is a factor to account for protocol overhead, which may include handshaking, packet collision, re-transmission and/ or back-off scheme traffic. An estimated transmission time for each of a plurality of links between said nodes may be calculated taking said two-hop residual bandwidth into account. For each possible route, a route efficiency function is determined at least by summing the estimated transmission times for all the links in the route, and the route in which the value of the route efficiency function is smallest is selected.

Stephen McLaughlin

Papers

Priority Swapping Subcarrier-User Allocation Technique for Adaptive Multicarrier Based Systems

Underwater depth imaging using time-correlated single-photon counting at video frame rates

Equalization via iterative detection in multi-mode fiber optic links

On the performance enhancement of the Empirical Mode Decomposition

Route selection in a mobile ad-hoc network