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

An Introduction to MultiAgent Systems.

Digital processing of speech signals

This paper proposes an approach for pedestrian tracking using dead reckoning enhanced with a mode detection using a standard smartphone. The mode represents a specific state of carrying device, and it is automatically detected while a person is walking. This paper presents a new approach, which extends and enhances previous methods by identifying in real-time three typical modes of carrying the device and using the identified mode to enhance tracking accuracy. The way of carrying the device in all modes is unconstrained to offer reliable person-independent tracking. Based on the identification of modes, a lightweight step-based tracking algorithm is developed with a novel step length estimation model. The tracking system is implemented on a commercial off-the-shelf smartphone equipped with a built-in inertial measurement unit with 3-D accelerometer and gyroscope. It achieves real-time tracking and localization performance with an average position accuracy of 98.91%.

A Multi-Mode Dead Reckoning System for Pedestrian Tracking Using Smartphones

Adaptive sliding window segmentation for physical activity recognition using a single tri-axial accelerometer

IEC 61499 has been endorsed as the standard for modeling and implementing distributed industrial process measurement and control systems. The standard prescribes the use of function blocks for designing systems in a component-oriented approach. The execution model of a basic function block and the manner for event/data connections between blocks are described therein. Unfortunately, the standard does not provide exhaustive specifications for function block execution. Consequently, multiple standard-compliant implementations exhibiting different behaviors are possible. This not only defeats the purpose of having a standard but also makes verification of function block systems difficult. To overcome this, we propose synchronous semantics for function blocks and show its feasibility by translating function blocks into a subset of Esterel, a well-known synchronous language. The proposed semantics avoids causal cycles common in Esterel and is proved to be reactive and deterministic under any composition. Moreover, verification techniques developed for synchronous systems can now be applied to function blocks.

A Synchronous Approach for IEC 61499 Function Block Implementation

SystemJ: A GALS language for system level design

After the cancellation of selective availability in May 2000, the accuracy of plain uncorrected global positioning systems (GPSs) is significantly increased. Currently, the main cause of error in plain GPS receivers is imperfect compensation of the ionospheric error. Tests performed in New Zealand show that even a simple code range differential GPS (DGPS) provides significant improvement in accuracy over a plain GPS and can be implemented using low-cost off-the-shelf equipment. A simple statistical model of positioning error estimation is developed and applied on data obtained by measurements in order to verify the findings. The method, which needs only position information without satellite constellation, provides sufficient accuracy when simultaneously gathered plain GPS and DGPS data are compared

Zoran Salcic

Papers

A Multi-Mode Dead Reckoning System for Pedestrian Tracking Using Smartphones

Adaptive sliding window segmentation for physical activity recognition using a single tri-axial accelerometer

A Synchronous Approach for IEC 61499 Function Block Implementation

SystemJ: A GALS language for system level design

A Comparison of Accuracy Using a GPS and a Low-Cost DGPS