(A) Tutorial on
Machine Learning from Data Streams: Techniques and Applications

Abstract: Machine learning from data streams has emerged as a new field of study in the past decade. It is now an established area with techniques of proven performance, and open source and commercial software tools. In this tutorial, an in-depth overview of the area will be given. Both unsupervised and supervised machine learning techniques will be discussed. Enabling machine learning techniques to perform on small computational devices using various approaches will be presented. Finally, the tutorial will be concluded with stimulating applications in different areas with a particular focus on medical applications.

Contents

1- Machine Learning from Data Streams
1.1 Data Streams Basics.
1.2 Clustering.
1.3 Classification.
1.4 Frequency Analysis.
1.5 Time Series Analysis.

2- Machine Learning from Data Streams on Mobile and Embedded Devices
2.1 Motivation.
2.2 Algorithm Granularity.
2.3 Situation-awareness.
2.4 Clutter-awareness for visualisation.
2.5 Pocket Data Mining.

3- Applications of Machine Learning from Data Streams
3.1 Medical Applications.
3.2 Prediction of Electricity Loading.
3.3 Disaster Management.

Biographies

Dr Mohamed Medhat Gaber
Mohamed Medhat Gaber is a Senior Lecturer at the University of Portsmouth, UK. He received his PhD in Artificial Intelligence from Monash University, Australia in 2006. He then held appointments with the University of Sydney, CSIRO, and Monash University, all in Australia. He has published more than 80 papers and edited/co-edited 4 books on data mining and knowledge discovery. Mohamed has served in the program committees of major conferences related to data mining, including ICDM, PAKDD, ECML/PKDD and ICML. He has also co-chaired over 10 workshops and special sessions on various data mining topics. Dr Gaber is recognised as a fellow of the UK Higher Education Academy (HEA). He is also a member of the International Panel of Expert Advisers for the Australasian Data Mining Conferences. In 2007, he was awarded the CSIRO teamwork award.

Dr Joao Gama
Joao Gama is Associate professor at the University of Porto and researcher at LIAAD-Inesc Tec. He served as PI in several FCT projects in learning adaptive systems. He published more than 110 papers in major International conferences and journals, served as PC chair at ECML05, DS09, ADMA09, and Conference Chair at IDA11.He co-organized a series of workshops on learning from data streams in conjunction with ECML-PKDD, KDD, SAC and ICML. He is member of the editorial board of MLJ, DAMI, NGC, and PAI and he is author of a recent book in Knowledge Discovery from Data Streams.

Dr Pedro Pereira Rodrigues
Dr Pedro Pereira Rodrigues is an Invited Assistant Professor at the Health Information and Decision Sciences department of the Faculty of Medicine of the University of Porto, Portugal. He endures research at the Artificial Intelligence and Decision Support Laboratory, and at the Centre for Research in Health Tecnologies and Information Systems, of the same university, which led to the completion of his Bsc (2003) MSc (2005) and PhD (2010) degrees at that university. His publication record includes more than 25 indexed (ISI or Scopus) publications, and he has coauthored 10 book chapters in publishers such as Springer Verlag and Chapman&Hall/CRC Press. He is/was the adviser or co-adviser of 16 MSc students, and has been involved in 6 national projects, 2 European research networks (funded by EU FP5 and FP6), and 2 bilateral agreements with a Slovenian university. Currently, his main research topic is predictive analysis and machine learning from streaming medical data. However, his research interests also include the reliability and evaluation of predictive and clustering analysis in streaming environments in healthcare. Following his research interests, he has organized 8 special sessions on the specific topic of data streams (including two focusing on medical data streams), served as program committee member of more than 30 research events and has reviewed several book chapters and more than 20 journal articles. He was the Local Organization Chair for the 10th International Symposium on Intelligent Data Analysis held in Porto, in October 2011.

(B) Tutorial Multimodal Biometrics: Aspects and challenges

Dr. Waheedah Al Mayyan

Software Technology Research Laboratory (STRL), UK.

 

Abstract

Biometrics technology has been attracting extensive attention from researchers and engineers of personal authentication due to the ever-growing demand on access control, public security, forensics and e-banking. With the fast development of biometric data acquisition sensors and data processing algorithms, diverse biometric systems have been now widely deployed in various applications. However, there are still many challenging problems in improving the accuracy, robustness, efficiency, and user-friendliness of biometric-based systems, and new problems are also emerging with new applications, e.g. personal authentication on mobile devices and internet.

The main focus of this special tutorial is on the recent advances of multimodal biometrics and the challenges in designing, developing, and deploying biometric-based algorithms and systems for various applications along with the latest results and findings in multimodal- biometrics based authentication and directions for the future development.

Topics:

·        Multimodal Biometrics: An Overview

·        Biometric Systems

·        Biometric Recognition System Modes

·        Multimodal Biometrics

·        Limitations of unimodal biometric systems

·        Motivation behind multimodal biometrics

·        Multimodal Biometric Fusion scenarios

·        Multimodal Biometric Architecture

·        Multimodal Biometric Fusion levels

·        Challenges related to Multimodal Biometric Sys. Design

·        Case study

 

Waheeda Almayyan received her B.Sc. and M.Sc. degrees from KuwaitUniversity, Kuwait. In 2012, she received her PhD degree from De Montfort University, UK. She is an Assistant Professor in the College of Business Studies, The Public Authority for Applied Education and Training, Kuwait. Her research interests include pattern recognition and Biometrics.

 

 

 

(C) Tutorial on Logic-based approach for the development of Dependable Computing Systems

Professor Dr. Hussein Zedan

Software Technology Research Laboratory (STRL), UK.

Temporal Logic (s) has been introduced by Amir Pnueli (around the 1960’s) as sound formalisms for the study of dynamic behaviours of computational systems.  It has gained a remarkable popularity and was applied to systems such as real-time, reactive systems which are used in many safety-critical applications. I will introduce Interval Temporal Logic (ITL) and its use in the specification, verification and development of dependable computing system that is used in critical application and in particular security-critical systems. In doing so, I shall illustrate its relationship with Inductive Logic Programming (ILP) and how ITL is more powerful given it is large executable subset, Tempura, and its runtime verification subsystem, AnaTempura.

Illustrative case studies will be taken from hardware systems, secure-critical application s and bioinformatics.

Tutorial structure includes:

1.      Temporal Logic and Computations

2.      Interval Temporal Logic: axiomatic system

3.      Tempura/Ana-Tempura – Realisation/Execution of  ITL

4.      ITL in Hardware Specification

5.      ITL  Transformation/Refinement  system

6.      Policy-based approach with ITL together with enforcement

7.      Case studies:

o    Hardware - Multi-threaded processor,

o    DNA Specification/Verification

o    Time-Critical application

o    Security Enforcement and deployment

 

Professor Hussein Zedan is the Technical Director of the Software Technology Research Laboratory (STRL) at De Montfort University, UK. His research interests include formal methods, verification, semantics, critical systems, re-engineering, computer security, CBD, and IS development.