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

Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Data Mining Practical Machine Learning Tools and Techniques

How has Japan become a major economic power, a world leader in the automotive and electronics industries? What is the secret of their success? The consensus has been that, though the Japanese are not particularly innovative, they are exceptionally skilful at imitation, at improving products that already exist. But now two leading Japanese business experts, Ikujiro Nonaka and Hiro Takeuchi, turn this conventional wisdom on its head: Japanese firms are successful, they contend, precisely because they are innovative, because they create new knowledge and use it to produce successful products and technologies. Examining case studies drawn from such firms as Honda, Canon, Matsushita, NEC, 3M, GE, and the U.S. Marines, this book reveals how Japanese companies translate tacit to explicit knowledge and use it to produce new processes, products, and services.

The knowledge-creating company : how Japanese companies create the dynamics of innovation

How to Do Things With Words

Temporal difference (TD) learning is a model-free reinforcement learning technique, which adopts an infinite horizon discount model and uses an incremental learning technique for dynamic programming. The state value function is updated in terms of sample episodes. Utilising eligibility traces is a key mechanism in enhancing the rate of convergence. TD(λ) represents the use of eligibility traces by introducing the parameter λ. However, the underlying mechanism of eligibility traces with an approximation function has not been well understood, either from theoretical point of view or from practical point of view. The TD(λ) method has been proved to be convergent with local tabular state representation. Unfortunately, proving convergence of TD(λ) with function approximation is still an important open theoretical question. This paper aims to investigate the convergence and the effects of different eligibility traces. In this paper, we adopt Sarsa(λ) learning control algorithm with a large, stochastic and dynamic simulation environment called SoccerBots. The state value function is represented by a linear approximation function known as tile coding. The performance metrics generated from the simulation system can be used to analyse the mechanism of eligibility traces.

Experimental analysis of eligibility traces strategies in temporal difference learning

A large number of studies on genetic algorithms (GAs) emphasize finding a globally optimal solution. Some other investigations have also been made for detecting multiple solutions. If a global optimal solution is very sensitive to noise or perturbations in the environment then there may be cases where it is not good to use this solution. We have proposed a new scheme, GA/RS/sup 3/, which extends the application of GAs to domains that require the discovery of robust solutions and a mathematical model for this scheme has been developed restricting their search space to one-dimensional. We analyze properties of GAs/RS/sup 3/ in multi-dimension search spaces. The effectiveness of the scheme is demonstrated by solving two-dimensional functions having broad and sharp peaks.

http://www2.hannan-u.ac.jp/~tsutsui/ps/kes98-rs3.pdf

Properties of robust solution searching in multi-dimensional space with genetic algorithms

Fuzzy Sets & their Application to Clustering & Training

There is a worldwide interest in the development of handwritten character and word recognition systems. These systems are used in many situations such as recognition of postcodes, interpretation of amount and verification of signature on bank checks, and law enforcement on public roads. This chapter introduces principles of handwriting recognition for on-line and off-line systems. It discusses some considerations about performance evaluation of a handwriting recognition system. Handwriting recognition systems acquire data by means of either offline devices or on-line devices. Most current approaches to off-line handwriting recognition consist of three main phases, namely, preprocessing, segmentation, and recognition. Recognition of handwritten characters is undoubtedly the basic task of most automatic reading systems. Most approaches to on-line handwriting recognition fall into two main categories: analysis by synthesis, which is based on the modeling of handwriting generation, and application of off-line recognition methods to on-line recognition.

An introduction to handwritten character and word recognition

Knowledge-based techniques have been used to automatically detect surface land mines present in thermal and multispectral images. Polarisation sensitive infrared sensing is used to highlight the polarisation signature of man-made targets such as landmines over natural features in the image. Processing the thermal polarisation images using a background discrimination algorithm we were able to successfully identify eight of the nine man-made targets, three of which were mines with only three false targets. A digital camera was used to collect a number of multispectral bands of the test mine area containing three surface landmines with natural and man-made clutter. Using a supervised and unsupervised neural network technique on the textural and spectral characteristics of selected multispectral bands we successfully identified the three surface mines but obtained numerous false targets with varying degrees of accuracy. Finally to further improve our detection of land mines we use a fuzzy rule based fusion technique on the processed polarisation resolved image together with the output results of the two best classifies. Fuzzy rule based fusion identified the locations of all three landmines and reduced the false alarm rate from seven (as obtained by the polarisation resolved image) to two.

Lakhmi C. Jain

Papers

Experimental analysis of eligibility traces strategies in temporal difference learning

Properties of robust solution searching in multi-dimensional space with genetic algorithms

Fuzzy Sets & their Application to Clustering & Training

An introduction to handwritten character and word recognition

Computational Intelligence Techniques in Landmine Detection