?>

Go to content

Home
ETSIINF en Twitter ETSIINF en Facebook
Inicio > Programas en extinción > Official postgraduate programme (to be phased out) > Advanced courses

Advanced courses

These seminars are the course offerings in the computer sciences module of the postgraduate programme master's degrees. Each master's degree requires or recognizes a number of different credits from the computer sciences module, depending on the curriculum of the master in question.

Students must enrol for the computer science course unit and earn the required number of credits by attending the seminars. For this purpose, students must give the Associate Dean for Research and Posgraduate Programmes advance notice of each seminar they intend to attend.

Seminars taught by guest lecturers

The seminars on offer for the 2006/07 academic year are as follows. They are all worth 2 ECTS. Unless otherwise stated, they will all be held in Seminar Room H1002 (Building 1) from 4.00 pm to 7.00 pm.

For undergraduate degree students, attendance of any of these seminars will be recognized as 1 free choice credit.

 

Seminars for 2006/07 Academic Year
From 26/02/07 to 02/03/07 Privacy, security, and data mining
Yucel Saigin, Sabancı Üniversitesi
From 06/03/07 to 09/03/07 Designing expert systems for disease diagnosis
Luis Laita, Professor Emeritus (UPM)
H1004 (Building I)
From 19/03/07 to 23/03/07 Data mining processes
Pedro Sousa, Universidade Nova de Lisboa
From 26/03/07 to 30/03/07 Molecular computing
Yakoov Benenson, Harvard University
From 23/04/07 to 27/04/07 Learning from data streams
Joao Manuel Portela da Gama, Universidade de Porto
07/05/07 a 11/05/07 Robustness programming patterns
Christof Fetzer, Technische Universität Dresden
From 16/05/07 to 18/05/07 Distributed computing environments
Bettina Kemme, McGill University
(from 4 pm to 8 pm)
From 21/05/07 to 25/05/07 Static analysis, program verification, and automated theorem proving
Deepak Kapur, University of New Mexico
From 28/05/07 to 01/06/07 Fault-tolerance in distributed systems
Jean Charles Fabre, Université  Toulouse/CNRS
From 18/06/07 to 22/06/07 Web usage mining
Bettina Berendt, Humboldt-Universität zu Berlin
From 25/06/07 to 29/06/07 Scalable storage systems: efficient remote block-level I/O & reliability
Angelos Bilas, University of Crete/FORTH

The above seminars are taught by guest lecturers. This gives students the chance to get acquainted with the research going on at other institutions, particulary abroad. Another set of seminars puts students into touch with research undertaken at this university, as they are taught by professors and researchers working at our university, attached to official research groups.

Seminars taught by in-house researchers

These seminars are held throughout the academic year. Each seminar has a duration of five days. The seminars are held on the same day of the week for five consecutive weeks (save minor exceptions). They are all worth 1 ECTS. Unless otherwise stated, all seminars are held in Lecture Theatre A6201 (Building 6) from 7.00 pm to 9.00 pm.

SEMINARS WITHOUT A SYLLABUS ARE TO BE CONFIRMED

(red=holiday; yellow=no seminar)

2006/07 academic year
MONDAY J 8 15 22 29 F   5    
TUESDAY A 9 16 23 30 E   6   Process Security and Improvement
WEDNESDAY N 10 17 24   B
R
  7   Application of ICT to Teaching/Learning Processes
THURSDAY U 11 18 25   U 1 8   18 - ICT... 25 - Natural Computing
FRIDAY A 12 19 26   A 2 9   Natural Computing
SATURDAY R 13 20 27   R 3 10    
SUNDAY Y
14 21 28   Y 4 11    
                     
MONDAY F
E
26     5 12 19 26   Development and Implementation of Languages with Declarative Technology
TUESDAY B 27 M   6 13 20 27   Mathematical Foundations for Soft Computing
WEDNESDAY R 28 A   7 14 21 28   Decision support systems
THURSDAY U   R 1 8 15 22 29   Computing applied to signal and image processing
FRIDAY A   C 2 9 16 23 30   Biomedical Informatics
SATURDAY R   H 3 10 17 24 31    
SUNDAY Y     4 11 18 25      
                     
MONDAY     16 23 30   7 14 21 Data Mining Engineering
TUESDAY A   17 24   1 8 15 22 Empirical Software Engineering
WEDNESDAY P   18 25 M 2 9 16 23 Ontological Engineering
THURSDAY R 12 19 26 A 3 10 17   Artificial Intelligence
FRIDAY I 13 20 27 Y 4 11 18   Computational Perception and Robotics
SATURDAY L 14 21 28   5 12 19    
SUNDAY   15 22 29   6 13 20    
                     
MONDAY     28     4 11 18 25 Orthogonal Polynomials and Fractal Geometry
TUESDAY M   29 J   5 12 19 26 Distributed Systems (H1005, Building 1)
WEDNESDAY A   30 U   6 13 20 27 Quantum Information and Computing (H1005, Building 1)
THURSDAY Y 24 31 N   7 14 21 28 Computer and communications technology
FRIDAY   25   E 1 8 15 22 29 Validation and business applications (H1005, Building I)
SATURDAY   26     2 9 16 23 30  
SUNDAY   27     3 10 17 24    

Course enrolment

The following master's degrees offer the computer science module:

Students of the Master of Information Technologies must enrol for subjects 103000248 and 103000249, Computer Science I and II, both worth 10 ECTS. Students of the other master's degrees have to enrol for subject 103000242, worth 6 ECTS. You can enrol either at the start of the course or when you have earned the respective credits.

Seminar syllabuses

Designing expert systems for disease diagnosis

Luis Laita, Professor Emeritus (UPM)

Students are expected to search the Internet for a description of the key symptoms of a disease and write a 10-page description of the disease and its symptoms. Symptoms are divided into five groups: emotional, cognitive, health-related, family-related and work- and employer-related. Each group is used to build a production rules base and a potential facts base. A method of condensing knowledge into bi-valued or three-valued modal logic formulas is applied.

The production rules are built using Karnaugh maps relating symptoms in each group and assigning a severity level ("tractable", undefined" and "intractable") to each piece of knowledge. They are then related to other higher levels. The partial results on disease severity are gathered at the end to give a diagnosis. The result is an expert system for disease diagnosis.

Distributed computing environments

Bettina Kemme, McGill University

Distributed computing environments are now a standard in information technologies environments. Examples are workstation clusters, mobile environments, grid systems, multi-layer middleware systems or peer-to-peer systems. Data management in such environments is a challenge. This course will cover topics like data consistency (consistency models, advanced transaction models, advanced concurrency control and distributed retrieval), data replication and querying distributed data. We will also look at fault tolerance issues. We will examine these topics first from an abstract viewpoint and then for specific computing environments, particularly, clusters and peer-2-peer environments. The course is composed of 15 class hours, and the lecturer will dedicate another 15 hours to other activities with the course students.

(1 free choice credit for undergraduate computing students)

Fault-tolerance in distributed systems

Jean Charles Fabre, Université Toulouse/CNRS

The seminar covers several aspects of dependability. It offers a brief introduction to different aspects of dependability, like security and fault tolerance. The seminar then details burning dependability issues. First it deals with intrusion tolerance (Shamir threshold schemes, fragmentation-redundancy-scattering techniques). Second, it deals with fault tolerance in two different contexts: distributed object-oriented systems and standard components (COTS). In the case of distributed object-oriented systems we will look at architectural aspects like using computational reflection to separate fault tolerance (non-functional aspect) from functional aspects. In the case of standard components we will look at wrapping techniques for adding fault tolerance to COTS components to increase their robustness. The course is composed of 15 class hours, and the lecturer will dedicate another 15 hours to other activities with the course students.

(1 free choice credit for undergraduate computing students)

Quantum Information and Computing

Quantum Information and Computing Research Group (Vicente Martín Ayuso)

Empirical Software Engineering

Empirical Software Engineering (Natalia Juristo Juzgado)

Experimental software engineering is a relatively recent area of software engineering related to the collection and analysis of data and experiences that can be used to characterize, evaluate and identify relations between different artefacts used to build software. In other words, the key goal is to come to understand, from an empirical point of view, the strengths and weaknesses of the technologies used in software development.

Experimental software engineering research aims to check the strengths of the technologies used in development. It provides data showing how to improve software production and determines the best conditions for application. As in other disciplines, experiments should also be replicated, either to corroborate the results under the same application conditions or to provide additional information in new contexts. As a result of this process, research could provide developers with experimentally tested knowledge about the benefits of applying different development technologies and their application conditions.

The application of experimental methods to rigorously evaluate and compare methods, techniques and tools used in software development assures that the results of the evaluations and comparisons are reliable. Different types of empirical studies can be applied depending on the set goals. Experiments are typically conducted in laboratory contexts. Case studies are used to analyse the effect of a particular technique or method in industry. Surveys are used to gain an overview of some facet of software development.

Artificial Intelligence

Artificial Intelligence Group (Alfonso Rodríguez-Patón Aradas)

Four of the five group talks will focus on introducing different branches of natural computing. Natural computing can be described as a scientific field with two goals: (1) understand the computational processes that take place in nature (particularly, biology) and (2) develop nature-inspired computational models. The seminar sessions will be as follows:

  1. Alfonso Rodríguez-Patón will present a summary of all five talks focusing especially on in vivo biomolecular computing, biomolecular automata, systems biology and research work in these areas.
  2. Andrei Paun will address "in info" biomolecular computing (specifically, the model referred to as P system or membrane cell computing) and his recent research in these areas.
  3. Petr Sosík will explain his view of the interrelationship between the following areas: biomolecular computing, computational biology, bioinformatics and biological computing. He will also discuss open problems and future lines of research.
  4. Daniel Manrique Gamo will give an introduction to artificial neural networks and evolutionary computing and will explain his research work in these fields.
  5. The fifth session will be led by Prof. Juan Pazos Sierra, who will expound his theory based on holons and informons

Learning from data streams

Joao Manuel Portela da Gama, Universidade de Porto

  1. Introduction to data streams
  2. Maintaining simple statistics
    2.1 Histograms
  3. Change detection
    3.1 Introduction
    3.2 Methods and algorithms
  4. Knowledge discovery
    4.1 Desiderata for learning
    4.2 Decision trees
    4.3 Clustering
    4.4 Association rules
    4.5 Multiple models
    4.6 Detecting drift using ensembles of classifiers
  5. Final words

Molecular computing

Yakoov Benenson, Harvard University

The course will emphasize experimental approaches and cover the following topics:

  1. Introduction to molecular biology for computer scientists
  2. Computer science concepts relevant to molecular computing
  3. Early theory of molecular computing (1970-1993)
  4. Solving combinatorial problems using DNA
  5. Self –assembly of DNA as computation
  6. Autonomous molecular computers and molecular automata
  7. In vivo computing systems

Computational Perception and Robotics

Computational Perception and Robotics Research Group (Luis Baumela Molina)

Privacy, security, and data mining

Yucel Saigin, Sabancı Üniversitesi

Lectures for this seminar will cover the following:

Document 1

Document 2

References

The remaining non-class hours will be spent interacting with students working on a real privacy preserving data analysis case study at the same time as database searches for knowledge and the challenges that were analysed in the theory classes.

(1 free choice credit for undergraduate computing students)

Data mining processes

Pedro Sousa, Universidade Nova de Lisboa

The professor will teach 10 class hours with the following outline:

The theory classes will be supplemented with at least 10 hours working on a case study:

(1 free choice credit for undergraduate computing students)

Publish/subscribe systems for information dissemination in dynamic distributed systems

Roberto Baldoni, Univ. La Sapienza di Roma

The publish/subscribe communication paradigm is an alternative to the more traditional client/server model. Thanks to its innate decoupling characteristics it is now recognized as a promising solution for applications requiring interactions with low coupling. During this seminar students will be introduced to this communication model and its possible implementations. In particular, the seminar will review event routing algorithms, discussed in terms of the induced message load, routing information required by each node, subscription language dependence, adaptability to dynamic changes in the underlying network. Additionally, we will specify how each class of algorithms relates to the overlay network, underlining in particular which overlay network is best for a specific events solution and why. Finally, we will look at the most representative pub/sub systems. The course is composed of 15 class hours and the professor will dedicate another 15 hours to other activities to attend to the course students.

(1 free choice credit for undergraduate computing students)

Robustness programming patterns

Christof Fetzer, Technische Universität Dresden

The purpose of this seminar is to introduce robustness patterns that are useful for preventing software failures and are fault tolerant at runtime. The topics addressed in this seminar will include: process pairs, communication models between processes, checkpointing, assertions, retries, exception handling, model checking, logging and software mutations. This seminar is based on the fault tolerant software postgraduate course that he now teachas at the TU Dresden. The seminar will be accompanied by some exercises that students will have to do as homework. Students will have access to a system of meta-aspects and associated tools to gain some practical experience with the topics covered by the seminar. The course is composed of 15 class hours, and the lecturer will spend another 15 hours working on other activities with the course students.

(1 free choice credit for undergraduate computing students)

Scalable storage systems: efficient remote block-level I/O & reliability

Angelos Bilas, University of Crete/FORTH

Modern storage systems have very strict scalability requirements and have to scale to high-capacity storage and effective I/O throughput. For this reason,  a growing number of these systems are being built with standard hardware, mainly PCs equipped with a lot of disks and interconnected by means of high-performance system area networks. In this first seminar we will describe the open problems in this field. We will outline our approach to addressing the architectural limitations of current storage systems and will present the current work on block-level I/O storage systems performance compression on standard hardware and RDMA-capable networks and interface cards. Then we will look at how the interface exported by network storage systems can be reliably extended. The course is composed of 15 class hours, and the lecturer will spend another 15 hours working on other activities with the course students.

(1 free choice credit for undergraduate computing students)

Decision support systems

Decision Analysis and Statistics Group (Alfonso Mateos Caballero)

Many of the problems we have to deal with are extremely complex due to the presence of several sources of uncertainty, conflicting objectives and goals, possible long-term impacts of decisions on different population groups... Even though they can sometimes be solved through experience and intuition, it has been found on repeated occasions that such approaches to complex problems can lead to poor solutions. As an alternative, decision support systems provide structured problem-solving models taking into account all their key aspects.

The objective of this seminar is to show students how decision support systems can be very useful tools for making the best decision about a problem. Specifically, we focus on two decision support systems that are available free on the Web and have been built for academic and practical purposes. The first system is called GMAA (Generic Multiattribute Analysis) and is available at http://www.dia.fi.upm.es/%7Eajimenez/GMAA. This system is the fruit of research carried out by members of the Decision Analysis and Statistics Group at the Department of Artificial Intelligence of the Universidad Politécnica de Madrid's School of Computing. The second is called GeNie. It is a graphical modelling (influence diagrams) environment for representing decision problems and evaluating optimal policies. It is available at http://genie.sis.pitt.edu/ together with examples, technical documentation and references.

The seminar is divided into two stages: system presentation and system use in case studies.

Distributed Systems

LSD: Distributed Systems Lab (Ricardo Jiménez Peris)

  1. Introduction to distributed systems and LSD activities
  2. Research into Scalable and Fault-Tolerant Databases
  3. Research into Edge Computing
  4. Research into Application Server Clustering
  5. Research into Distributed Computing Theory

Static analysis, program verification, and automated theorem proving

Deepak Kapur, University of New Mexico

This short course will give a bird's eye view of current attempts in deriving program invariants/properties. The main focus will be on recent research by Prof. Kapur on automatically generating loop invariants using methods based on abstract interpretation and polynomial ideal theory, quantifier-elimination and solving recurrence relations. The course will also discuss the importance of decision procedures and automated theorem-proving techniques in these activities. Time permitting, techniques based on counter-example guided abstraction refinement framework will be discussed for deriving program properties.

Validation and Business Applications Group

Validation and Business Applications Group (Jesús Cardeñosa Lera)

  1. Group lines of research and related research work.  Speaker: Jesús Cardeñosa
  2. Machine translation: achievements and problems. ETAP System. Speaker: Igor Boguslavsky
  3. Lexical resources: Lexical database design. Speaker: Prof. Eugenio Santos
  4. New lines of information extraction from texts. Speaker: Carolina Gallardo
  5. Multilingual solutions for problems in constrained domains. Speaker: Adriana Toni.

Web usage mining

Bettina Berendt, Humboldt-Universität zu Berlin

(1 free choice credit for undergraduate computing students)

Applying ICT to teaching/learning processes

Natural Computing

Development and Implementation of Languages with Declarative Technology

Mathematical Foundations of Soft Computing

Computing applied to Signal and Image Processing

Biomedical Informatics

Data Mining Engineering

Ontological Engineering

Orthogonal Polynomials and Fractal Geometry

Process Security and Improvement

Computer and Communications Technology