CISUC has nine research groups:
G1: Computer Science
G2: Artificial Intelligence: Foundations and Applications
G3: Simulation and Information Technologies in Education and Training
G4: Adaptive Computation
G5: Dependable Systems
G6: Communications and Telematics
G7: Databases
G8: Information Systems
G9: Evolutionary and Complex Systems
The list of PhD Proposals is pided by each group.
Each proposal has a reference Gx.y (x is the number of the group; y is the number of
the proposal within that group). Use these references to fill in the Scholarship
Application Form.
CISUC will offer 9 scholarships in the upcoming year. Students that do not have any
other sources of funding can apply to these scholarships. The scholarship includes the
payment of the tuition fees and a monthly payment of 980€. Students should be aware
that this is a scholarship only for the first year. Financial support for the remaining
years of the PhD should be of the responsibility of the Supervisor. It is advisable that
students should first discuss with their future Supervisor about the possibilities for
financial support for the rest of the PhD program.
Candidates should choose 3 Thesis Proposals, listed by order of preference. The chosen
proposals can be from the same or different groups.
Each group of CISUC has one scholarship to offer. The best candidate in each group
will get the scholarship. The selection of the candidates that will get a scholarship is not
a global process, but rather a distributed process with nine independent queues of
selection. Each group is responsible for the selection of the best candidate, and the final
list will be approved by the Scientific Commission of CISUC.
In the following pages we present a list of proposed PhD Thesis for the next year (2007-
2009). The candidates to the scholarship should read this list carefully and choose the
3 proposals that better fit into the interests and research background.
Deadline to apply for a scholarship: 21st of July, 2006
G1: Computer Science
http://cisuc.dei.uc.pt/csg/
PhD Thesis Proposal: G1.1
Title: “Human readable ATP proofs in Euclidean Geometry”
Keywords: Automatic Theorem Proving, Axiomatic Proofs in Euclidean Geometry.
Supervisor: Prof. Pedro Quaresma (pedro@mat.uc.pt)
Summary:
Automated theorem proving (ATP) in geometry has two major lines of research:
axiomatic proof style and algebraic proof style (see [6], for instance, for a survey).
Algebraic proof style methods are based on reducing geometry properties to algebraic
properties expressed in terms of Cartesian coordinates. These methods are usually very
efficient, but the proofs they produce do not reflect the geometry nature of the problem
and they give only a yes/no conclusion. Axiomatic methods attempt to automate
traditional geometry proof methods that produce human-readable proofs. Building on
top of the existing ATPs (namely GCLCprover [5, 4, 8, 9, 10] to the area method [1, 2, 3,
7, 8, 11] or ATPs dealing with construction [6] the goal is to built an ATP capable of
producing human-readable proofs, with a clean connection between the geometric
conjectures and theirs proofs.
References:
[1] Shang-Ching Chou, Xiao-Shan Gao, and Jing-Zhong Zhang. Automated production of traditional proofs for
constructive geometry theorems. In Moshe Vardi, editor, Proceedings of the Eighth Annual IEEE Symposium
on Logic in Computer Science LICS, pages 48–56. IEEE Computer Society Press, June 1993.
[2] Shang-Ching Chou, Xiao-Shan Gao, and Jing-Zhong Zhang. Automated generation of readable proofs with
geometric invariants, I. multiple and shortest proof generation. Journal of Automated Reasoning, 17:325–347,
1996.
[3] Shang-Ching Chou, Xiao-Shan Gao, and Jing-Zhong Zhang. Automated generation of readable proofs with
geometric invariants, II. theorem proving with full-angles. Journal of Automated Reasoning, 17:349–370,
1996.
[4] Predrag Janicic and Pedro Quaresma. Automatic verification of regular constructions in dynamic geometry
systems. In Proceedings of the ADG06, 2006.
[5] Predrag Janicic and Pedro Quaresma. System description: Gclcprover + geothms. In Ulrich Furbach and
Natarajan Shankar, editors, IJCAR 2006, LNAI. Springer-Verlag, 2006.
[6] Noboru Matsuda and Kurt Vanlehn. Gramy: A geometry theorem prover capable of construction. Journal
of Automated Reasoning, (32):3–33, 2004.
[7] Julien Narboux. A decision procedure for geometry in coq. In Proceedings TPHOLS 2004, volume 3223 of
Lecture Notes in Computer Science. Springer, 2004.
[8] Pedro Quaresma and Predrag Janicic. Framework for constructive geometry (based on the area method).
Technical Report 2006/001, Centre for Informatics and Systems of the University of Coimbra, 2006.
[9] Pedro Quaresma and Predrag Janicic. Geothms - geometry framework. Technical Report 2006/002, Centre
for Informatics and Systems of the University of Coimbra, 2006.
[10] Pedro Quaresma and Predrag Janicic. Integrating dynamic geometry software, deduction systems, and
theorem repositories. In J. Borwein and W. Farmer, editors, MKM 2006, LNAI. Springer-Verlag, 2006.
[11] Jing-Zhong Zhang, Shang-Ching Chou, and Xiao-Shan Gao. Automated production of traditional proofs
for theorems in euclidean geometry i. the hilbert intersection point theorems. Annals of Mathematics and
Artificial Intelligenze, 13:109–137, 1995.
PhD Thesis Proposal: G1.2
Title: “Formal languages in the knowledge base management”
Keywords: Formal language, knowledge base, knowledge base inconsistency,
knowledge base management.
Supervisor: Prof. Maria de Fátima Gonçalves (mflfag@iscac.pt)
Summary:
Knowledge base management is intended as the acquisition and normalization of new
knowledge and the confrontation of this knowledge with existing knowledge, resolving
potential conflicts and updating it. When updating a knowledge base, several problems
may arise. Some of them are the redundancy of the updated knowledge and the
knowledge base inconsistency. Formal languages can be used with success to a
development of an innovative system to do the knowledge base management, resolving
potential updating problems.
PhD Thesis Proposal: G1.3
Title: “Medidas de semelhança e aplicações”
Keywords: Kolmogorov complexity, normalized distances, machine learning,
clustering, universal similarity metric.
Supervisor: Prof. Ana Maria de Almeida (amca@mat.uc.pt)
Summary:
Um problema recorrente em muitas estratégias que recorrem à apropriação de
conhecimento (learning) para actuar é o da identificação de "semelhanças". Como
podemos medir "semelhança" para, por exemplo, determinar a distância evolutiva
(evolucionária?) entre duas sequências como, por exemplo, 2 documentos na Internet,
ou 2 programas de computador, ou duas pautas musicais, ou 2 genomas, ou 2 linhas
eco-cardiográficas?
Este projecto pretende estudar a área emergente das medidas de semelhança (similarity
distance measures), úteis em data-mining, pattern recognition, learning e automatic
semantic extraction. Vitány et al criaram a "normalized information distance", baseda
na noção de Complexidade Algorítmica da Informação (vulgarmente conhecida como
Kolmogorov Complexity), e que mostraram ser uma medida universal para a classe em
causa, pois descobre todas as semelhanças efectivas. Mostraram ainda que é uma
métrica e que toma valores em [0,1]. Existem já vários exemplos práticos interessantes
em termos de resultados finais: semelhanças moleculares, identificação de
compositores através de pautas de música, ecocardiografias fetais, entre outros. A
teoria é geral e, como tal, aplicável a persas áreas ou colecções de objectos. Pretendese
investigar a sua adequação prática a casos particulares como learning algorithms
(kernel function) e clustering.
PhD Thesis Proposal: G1.4
Title: “A entropia (de Kolmogorov) na definição de modelos de Taxa-Distorção (R-D)
para a codificação de sinal de video”
Keywords: Kolmogorov complexity, Entropy, Rate-Distortion Theory.
Supervisor: Prof. Ana Maria de Almeida (amca@mat.uc.pt)
Summary:
Na última década, a Complexidade Algorítmica da Informação (ou Kolmogorov
Complexity) impôs-se como suporte teórico de uma série de aplicações nas mais
persas áreas, permitindo a (re)definação de modelos e conceitos. Nomeadamente, e
devido a uma interessante ligação forte entre a noção estatística da medida de Entropia
de Shannon e a determinista medida de CAI, tem-se assistido, não só a demonstrações
persas desta ligação teórica, mas também à realização de estudos e modelos que
mostram que a CAI pode substituir a Entropia como, e apenas como exemplo, em
Criptografia, com a vantagem de já não ser necessário conhecer distribuições de
probabilidade que, na verdade, não são, até agora realmente conhecidas.
Pretende este projecto, e uma vez que já é conhecido um modelo teórico, paralelo ao
estudo de Shannon, para substituir a Entropia na avaliação da Teoria de R-D (Rate-
Distortion), sobre a qual se baseia a transmissão de sinal, averiguar sobre a possivel
implementação prática deste novo modelo, à optimização de funções de R-D na
codificação de sinal de video.
G2: Artificial Intelligence: Foundations and
Applications
http://ailab.dei.uc.pt/
PhD Thesis Proposal: G2.1
Title: “Computational Aesthetics”
Keywords: Evolutionary Art, Artificial Artists, Computational Aesthetics
Supervisor: Prof. Penousal Machado (machado@dei.uc.pt)
Summary:
The use of biological inspired techniques for image generation tasks is a recent, exciting
and significant area of research. There is a growing interest in the application of these
techniques in fields such as: visual art and music generation, analysis, and
interpretation; sound synthesis; architecture; video; design; etc. In most cases, these
systems resort to a human guided evolutionary algorithm: the user provides fitness
scores for the images thus steering evolution towards images that match his/hers
aesthetic preferences.
Although these systems are interesting in their own right as computer aided creativity
tools, they lack autonomy and the capacity of making their own aesthetic judgements,
consequently, they cannot be considered artificial artists. According to the framework
proposed by Machado et al. (2003) an artificial artist can be seen as a system composed
by two modules: a creator and a critic.
The stunning imagery created with interactive evolutionary art tools indicates that an
evolutionary computation approach is suitable for the creator role. As such, the
bottleneck lies in the development of adequate artificial critics.
In Machado et al. (2004) the authors propose a generic framework for the development
of artificial art critics composed by two main modules: a feature extractor, which is
responsible for the “perception” of the artwork, collecting a set of low level aesthetically
relevant features; an evaluator, which performs an assessment of the aesthetic merits of
the artwork based on the output of the feature extractor module.
The proposed approach was tested in several author and style identification tasks
achieving high success rates (]95%), which shows the feasibility of the approach.
Following the previous work of Machado et al. (e.g. 2003, 2004) the present thesis will
focus on the development of artificial art critics and on their integration with existent
interactive art tools. The key issues to be addressed are the further development of the
current feature extractor, the analysis of the relevance of the incorporated features, and
the online training of the evaluator module.
References:
- Penousal Machado, Juan Romero, Bill Manaris, Antonino Santos, Amilcar Cardoso,
"Power to the Critics - A Framework for the Development of Artificial Art Critics",
IJCAI'2003 Workshop on Creative Systems, Acapulco, Mexico, August-2003.
- Penousal Machado, Juan Romero, Maria Ares, Amilcar Cardoso, Bill Manaris,
"Adaptive Critics for Evolutionary Artists", 2nd European Workshop on Evolutionary
Music and Art, Coimbra, Portugal, April-2004.
PhD Thesis Proposal: G2.2
Title: “Stylistic Based Image Retrieval and Classification”
Keywords: Content Based Image Retrieval, Artificial Art Critics, Computational
Aesthetics
Supervisor: Prof. Penousal Machado (machado@dei.uc.pt)
Summary:
The days of the text based World Wide Web are over. Today the web is dominated by
multimedia content. Nevertheless, the most popular search engines are text-based.
Although a few search engines for finding images and video are available, these engines
are still based on text. The large commercial image providers use human indexers to
select keywords and to classify each of the images. Google and other popular search
engines allowing image or video search base the retrieval on the textual content of the
page or on the keywords inserted by a human user.
Content based image retrieval is a complex task. Taken to the limit it is a generic vision
problem, requiring object recognition, image understanding, concept formation, object
classification, content analysis, etc. Computer vision is probably one of the hardest
problems in Artificial Intelligence and one that has been baffling computer scientists
during the last decades. Although significant progress has been made, it is not likely to
be solved in the near future. This poses limits on what content Based Image Retrieval
can accomplish; however it also creates opportunities.
In Machado et al. (2004) that authors presented a system composed by two main
modules: a feature extractor, which is responsible for the “perception” of the image; an
evaluator, which performs an assessment of the image based on the output of the
feature extractor module. The feature extractor is composed by a set of low-level
features that proved to be relevant to the stylistic classification of images. The proposed
approach was tested in several author and style identification tasks achieving high
success rates (]95%), which shows the feasibility of the approach and its potential for
the development of stylistic based image retrieval engines.
The main objective of this thesis is the further development of the current feature
extractor, focusing on the incorporation of features proved to be relevant for content
based image retrieval tasks. Unlike other content based image retrieval systems, which
focus on the identification of images containing a given set of objects, we are primarily
interested on the retrieval of stylistically similar images. Therefore, particular emphasis
will be given to features that are able to capture stylistic characteristics. Additionally,
the development of a fully working prototype and its empirical testing are key aspects of
this thesis.
References:
- Penousal Machado, Juan Romero, Bill Manaris, Antonino Santos, Amilcar Cardoso,
"Power to the Critics - A Framework for the Development of Artificial Art Critics",
IJCAI'2003 Workshop on Creative Systems, Acapulco, Mexico, August-2003.
- Penousal Machado, Juan Romero, Maria Ares, Amilcar Cardoso, Bill Manaris,
"Adaptive Critics for Evolutionary Artists", 2nd European Workshop on Evolutionary
Music and Art, Coimbra, Portugal, April-2004.
PhD Thesis Proposal: G2.3
Title: “Image Representation for Evolutionary computation”
Keywords: Evolutionary Art, Programmatic Compression
Supervisor: Prof. Penousal Machado (machado@dei.uc.pt)
Summary:
The use of evolutionary computation techniques for image generation tasks is a recent,
exciting and significant area of research. Following the footsteps of Karl Sims, most
systems resort to the evolution of symbolic expressions. Once evaluated these
expressions result in images. The tree-like nature of the expressions allows their
meaningful manipulation through conventional Genetic Programming operators, and
the results attained by several researchers during the past years show that this is a
powerful image generation method. Although expression based representations prove to
be adequate in the context of an evolutionary approach, they suffer from a major
problem. The generation of an image from a given expression is straightforward,
however the inverse problem – finding the symbolic expression for a given image – is
(NP) hard. The goals of this thesis are twofold: 1) explore alternative image
representation schemes, e.g. line based representations, that are adequate for
evolutionary computation but that do no suffer from the aforementioned problem; 2)
Following the work of McGuire, Nordin, and others explore the feasibility of using
programmatic image compression methods in order to find compact symbolic
expressions for a given image.
PhD Thesis Proposal: G2.4
Title: “Evolution of Dynamic Reactive Artworks”
Keywords: Evolutionary Art, Fitness Automation
Supervisor: Prof. Penousal Machado (machado@dei.uc.pt)
Summary:
The use of evolutionary computation techniques for image generation tasks is a recent,
exciting and significant area of research. During the past few years we watched the
emergence of several evolutionary art tools. Most of these tools are interactive, in the
sense that the user guides the evolutionary process, thus steering evolution towards
images that match his/hers aesthetic preferences. However, the resulting artworks are
static images and not interactive installations. The typical evolutionary art tool evolves
programs, once executed this programs result in images. The programs take no input;
as a result their output is static. The objective of this thesis is the development of and
evolutionary art tool able to produce dynamic and reactive artworks. To achieve this
task the programs being evolved receive an input signal, e.g. a hand gesture, their
output depends on the input. Therefore different “gestures” will yield different images,
which results in a dynamic reactive artwork. The programs being evolved can be seen
as mappings between the input signal and images. The difficulty lies on the
development of interesting mappings. Given the time-based nature of the domain
resorting to human guided evolution is not a viable option. Instead we wish to explore
ways of automating fitness assignment, so that uninteresting mappings can be
eliminated beforehand. Additionally, and although “gestures” are a natural choice for
input signal, others exist and should be explored in this thesis, allowing the evolution of
artworks that react to sound, lighting, bio-signals, etc.
PhD Thesis Proposal: G2.5
Title: “An Affect-based Multi-Agent System”
Keywords: Affect, emotion, autonomous agents, multi-agent systems
Supervisor: Prof. Luís Macedo (macedo@dei.uc.pt)
Summary:
Emotion and motivation (merged in the broader term affect) are essential for survival,
well-being and communication in humans by, among other functions, playing a central
role on cognitive activities such as decision-making, planning and creativity. So, the
question is why artificial agents don’t take advantage of emotions and motivations as
humans do? What can emotional artificial agents do better than those that are not
based on emotion? What emotion and motivation can offer to artificial agents?
Certainly, not all advantages that humans benefit are applicable to artificial agents.
But, we might think in a series of situations in which we can see the emotional
advantage such as in text-to-speech systems by giving more intonation to speech,
entertainment, preventive medicine, helping autistic people, in artificial pets,
personalized agents that can act on the behalf of someone by selecting news, music,
etc., according to someone’s mood, consumer feedback by measuring the emotions of
consumers when dealing with a specific product, etc. Such applications require the
abilities to recognize, express and experience emotions. Research in Artificial
Intelligence has almost ignored this significant role of emotions on reasoning, and only
recently this issue was taken seriously mainly because of the recent advances in
neuroscience, which have given evidence that cognitive tasks of humans, and
particularly planning and decision-making, are influenced by emotion .
The research question/thesis statement is that those tasks mentioned above can be
robustly performed by affective agents. The approach comprises the development of a
multi-agent system comprising affective agents [Macedo & Cardoso, 2004] so that this
framework can be used to build agent-based applications.
References:
- Macedo, L. and A. Cardoso (2004). Exploration of Unknown Environments with
Motivational Agents. Proceedings of the Third International Joint Conference on
Autonomous Agents and Multiagent Systems. N. Jennings and M. Tambe. New York,
IEEE Computer Society: 328 - 335.
- Macedo, L. The Exploration of Unknown Environments by Affective Agents. PhD
Thesis, Universidade de Coimbra, 2006
PhD Thesis Proposal: G2.6
Title: “Collaborative Multi-Agent Exploration of 3-D Dynamic Environments”
Keywords: Collaborative Multi-Agent Exploration of 3-D Dynamic Environments
Supervisor: Prof. Luís Macedo (macedo@dei.uc.pt)
Summary:
Exploration gathers information about the unknown. Exploration of unknown
environments by artificial agents (usually mobile robots) has actually been an active
research field [Macedo & Cardoso, 2004]. The exploration domains include planetary
exploration (e.g., Mars or lunar exploration), search for meteorites in Antarctica, volcano
exploration, map-building of interiors, etc. Several exploration techniques have been
proposed and tested either in simulated and real, indoor and outdoor environments,
using single or multiple agents. The main advantage of multi-agent approaches is to
avoid covering the same area by two or more agents. However, there is still much to be
done especially in dynamic environments as those mentioned above. Besides, real
environments, however, consist of objects. For example, office environments possess
chairs, doors, garbage cans, etc., cities comprise several kinds of buildings (houses,
offices, hospitals, churches, etc.), cars, etc. Many of these objects are non-stationary,
that is, their locations may change over time. This observation motivates research on a
new generation of mapping algorithms, which represent environments as collections of
objects. At a minimum, such object models would enable a robot to track changes in
the environment. For example, a cleaning robot entering an office at night might realize
that a garbage can has moved from one location to another. It might do so without the
need to learn a model of this garbage can from scratch, as would be necessary with
existing robot mapping techniques. This thesis addresses the problem of finding multiagent
strategies to address the problem of collaborative exploration of unknown, 3-D,
dynamic environments. The strategy or strategies should be tested against other
exploration strategies found in the literature.
References:
- Macedo, L. The Exploration of Unknown Environments by Affective Agents. PhD
Thesis, Universidade de Coimbra, 2006.
- Macedo, L. and A. Cardoso (2004). Exploration of Unknown Environments with
Motivational Agents. Proceedings of the Third International Joint Conference on
Autonomous Agents and Multiagent Systems. N. Jennings and M. Tambe. New York,
IEEE Computer Society: 328 - 335.
PhD Thesis Proposal: G2.7
Title: “Case-Based Hierarchical-Task Network Planning”
Keywords: HTN Planning, Case-based Planning, Decision-theoretic planning
Supervisor: Prof. Luís Macedo (macedo@dei.uc.pt)
Summary:
Hierarchical-Task Network (HTN) planning is a planning methodology that is more
expressive than STRIPS-style planning. Given a set of tasks that need to be performed
(the planning problem), the planning process decomposes them into simpler subtasks
until primitive tasks or actions that can be directly executed are reached. Methods
provided by the domain theory indicate how tasks are decomposed into subtasks.
However, for many real-world domains, sometimes it is hard to collect methods to
completely model the generation of plans. For this reason an alternative approach that
is based on cases of methods has been taken in combination with methods. Real-world
domains are usually dynamic and uncertain. In these domains actions may have
several outcomes, some of which may be more valuable than others. Planning in these
domains require special techniques for dealing with uncertainty. Actually, this has been
one of the main concerns of the planning research in the last years, and several
decision-theoretic planning approaches has been proposed and used successfully, some
based on the extension of classical planning and others on Markov-Decision Processes.
In these decision-theoretic planning frameworks actions are usually probabilistic
conditional actions, preferences over the outcomes of the actions is expressed in terms
of an utility function, and plans are evaluated in terms of their expected utility. The
main goal is to find the plan or set of plans that maximizes an expected utility function,
i.e, to find the optimal plan. In this thesis a planner that combines the technique of
decision-theoretic planning with the methodology of HTN planning should be built in
order to deal with uncertain, dynamic large-scale real-world domains [Macedo &
Cardoso, 2004]. Unlike in regular HTN planning, methods for task decomposition
shouldn’t be used, but instead cases of plans. The planner should generate a variant of
a HTN - a kind of AND/OR tree of probabilistic conditional tasks - that expresses all the
possible ways to decompose an initial task network.
References:
- Macedo, L. and A. Cardoso (2004). Case-Based, Decision-Theoretic, HTN Planning.
Advances in Case-Based Reasoning: Proceedings of the 7th European Conference on
Case-Based Reasoning. P. Calero and P. Funk. Berlin, Springer: 257-271.
- Macedo, L. The Exploration of Unknown Environments by Affective Agents. PhD
Thesis, Universidade de Coimbra, 2006.
PhD Thesis Proposal: G2.8
Title: “Ontology Learning from Text in Portuguese”
Keywords: Ontology Learning, Information Extraction, Natural Language Processing
Supervisor: Prof. Paulo Gomes (pgomes@dei.uc.pt)
Summary:
Much of today’s Knowledge is gathered in a textual format (e.g. imagine the amount of
knowledge available on the web), as such, the extraction and mining of knowledge from
texts becomes an important dimension of current research. One way of representing
knowledge is through the use of an ontology. Simply put, an ontology is a shared
understanding of some domain of interest where hidden connections between concepts
are made explicit. The use of such a structure allows communities and computer
systems to share a consistent understanding of what information means; in other words
its semantics.
The main objective of this thesis is the development of a set of methodologies that allow
the extraction of important concepts (relative to a certain domain) from text along with
the implicit and explicit relations that hold between them. Another aspect the thesis
should address pertains to the evaluation of the knowledge extracted and its
applicability in information systems.
PhD Thesis Proposal: G2.9
Title: “Intelligent Knowledge Management using the Semantic Web”
Keywords: Semantic Web, Knowledge Management, Artificial Intelligence, Ontologies
Supervisor: Prof. Paulo Gomes (pgomes@dei.uc.pt)
Summary:
Nowadays, companies gather and store big amounts of information in databases. This
information presents potential high value knowledge for a company. But most of this
information or data is not transformed in knowledge, remaining lost in data bases or
document repositories. Software development is a knowledge intensive activity involving
several types of know-how and skills. Usually development teams have several
members, which makes sharing and dissemination of knowledge crucial for project
success. One evolving technology that can be used with the purpose of building
knowledge management tools for the software development area is the semantic web.
Semantics are the lost chain between information/data and knowledge, and the
semantic web provides the infrastructure needed for making a true sharing of
knowledge possible.
The semantic web is an infrastructure providing semantics associated with words in
web resources. But, by itself it does not provide a tool for knowledge management. What
are needed, are tools that enable the usage of the semantic web in an intelligent way, so
that users can take advantage of knowledge sharing. The main problem to be dealt with
in this thesis is how a team of software development engineers can be aided by a tool, or
a set of tools, that enable them to reuse knowledge in a more efficient way, thus
increasing their productivity.
The main objective of this thesis is to develop a set of tools based on the semantic web.
These tools are intended to have a set of intelligent characteristics, such as: learning,
proactive reasoning, semantic searching and retrieval of knowledge, representation of
knowledge, knowledge acquisition, personalization, and others. Several reasoning
methods have been developed in Artificial Intelligence and are ideal candidates to be
used in this research work. Some of the results of this research work are new
algorithms and methodologies for knowledge management.
PhD Thesis Proposal: G2.10
Title: “Automatic Document Indexation in Portuguese”
Keywords: Document Indexing, Natural Language Processing, Semantic Web
Supervisor: Prof. Paulo Gomes (pgomes@dei.uc.pt)
Summary:
The main goal of this thesis proposal is to develop a system capable of indexing
documents in an ontology. The application area for this thesis is the domain of software
development. The main research contribution of this thesis is the development of
classification and indexation methodologies for documents related with software
engineering. Target documents are written in Portuguese and are related with software
development (manuals, reports, papers ...). The approach to be followed is based on a
repository of knowledge objects, which is structured by a ontology. The main
infrastructure for the storage and indexing of these objects is based on the languages
developed for the Semantic Web. New classification algorithms will also have to be
developed, so that they can cope with this document persity. Another major challenge
of this thesis is the correct disambiguation of document topics. This work is to be
integrated into ReBuilder (rebuilder.dei.uc.pt) – a software tool for reuse of UML
diagrams.
PhD Thesis Proposal: G2.11
Title: “Automatic Named Entity Recognition”
Keywords: Named Entity Recognition, Natural Language Processing, Case-Based
Reasoning
Supervisor: Prof. Paulo Gomes (pgomes@dei.uc.pt)
Summary:
Automatic named entity recognition is the identification and classification of linguistic
expressions that refer to a specific entity. For example, “Coimbra University” is a named
entity, which comprises a sequence of words referring to the University of Coimbra.
These entities (represented by a sequence of words) possess their own linguistic
properties. Natural language processing systems and other applications dealing with
natural language text must be able to identify and classify these entities, in order to use
the associated semantics, which is very different from using the words inpidually.
There are several international contests that compare and evaluate systems for named
entity identification, for example MUC (Message Understanding Conference), later the
ACE (Automatic Content Extraction). More recently the first contest for Portuguese
HAREM (http://poloxldb.linguateca.pt/harem.php). This goal of this proposal is to
develop a named entity recognition system for the Portuguese language, and participate
in the HAREM contest. A Case-Based Reasoning (CBR) approach is suggested. CBR can
be defined as a way of reasoning based on past experiences, which from our point of
view CBR can be applied successfully to this problem. CBR also enables the integration
of other approaches making a good framework for solving this problem.
PhD Thesis Proposal: G2.12
Title: “Converting Text into UML Diagrams”
Keywords: Natural Language Processing, Case-Based Reasoning, Software Reuse,
UML
Supervisor: Prof. Paulo Gomes (pgomes@dei.uc.pt)
Summary:
Language is the most common form of communication between humans, both written
and spoken. Software developers are no exception, with natural language text being an
important part of the software specification documents. In the last decade, software
modeling languages, such as UML, have been developed and used in the specification of
software systems. The main idea of this proposal is to develop an approach for the
conversion of natural language software specifications into UML diagrams, both use
cases and class diagrams. This work is to be integrated into ReBuilder
(rebuilder.dei.uc.pt) – a software tool for reuse of UML diagrams.
PhD Thesis Proposal: G2.13
Title: “Reusing Software Design Patterns”
Keywords: Software Design Patterns, UML, Case-Based Reasoning, Software Design
Reuse
Supervisor: Prof. Paulo Gomes (pgomes@dei.uc.pt)
Summary:
Software engineers and programmers deal with repeated problems and situations in the
course of software design. Software design patterns were developed to deal with this
type of situation, where the same abstract solution is used several times. Software
design patterns can be defined as a description of an abstract solution for a category of
design problems. One of the main advantages of patterns is design reusability. Another
main advantage is that the application of design patterns improves and makes software
maintenance easier – design for change. Software design patterns are described in
natural language, not having a formalization. This is due to the abstract level of the
design patterns, which makes the application of design patterns a human dependant
task.
Existing approaches to pattern application using computer tools need the help and
guidance of a human designer. This is especially true in the selection of the design
pattern to apply. It is difficult to automate the identification of the context in which a
pattern can be applied. Human designers must also identify which are the objects
involved in the pattern application. The automation of this task opens the possibility of
CASE design tools to provide complete automation in applying design patterns. This
new functionality can help the software designer to improve systems, and do better
software reuse. Case-Based Reasoning can be defined as a way of reasoning based on
past experiences. From our point of view CBR can be applied successfully to the
automation of software design patterns. The aim of this thesis proposal is to develop an
approach that addresses this problem. The proposed CBR framework must be able to
select which pattern to apply to a target design problem generating a new design. It can
also learn new cases from the application of design patterns. This work is to be
integrated into ReBuilder (rebuilder.dei.uc.pt) – a software tool for reuse of UML
diagrams.
PhD Thesis Proposal: G2.14
Title: “Adaptation and Reuse Mechanisms for UML Diagrams”
Keywords: Case-Based Reasoning, Ontologies, Software Reuse, UML
Supervisor: Prof. Paulo Gomes (pgomes@dei.uc.pt)
Summary:
Management and reuse of UML diagrams is an important aspect for any software
development company. The productivity increase that can be obtained from an effective
reuse of software development knowledge is crucial for market survival. This proposal
intends to develop new mechanisms for adaptation and reuse of UML diagrams. This
work is to be integrated into ReBuilder (rebuilder.dei.uc.pt) – a software tool for reuse of
UML diagrams. ReBuilder will work as a development platform, in which the developed
approaches will be integrated. This enables an easy testing and experimentation of the
developed approaches.
PhD Thesis Proposal: G2.15
Title: “Algorithms for Semantic Annotation of Positioning Information”
Keywords: locations, places, positioning systems, location based services
Supervisors: Prof. Francisco Câmara (camara@dei.uc.pt)
Prof. Carlos Bento (bento@dei.uc.pt)
Summary:
Although we find today a myriad of positioning technologies (from the “common” GPS to
Wireless, GSM cell or Ultra Wide Band positioning algorithms), the interpretation of
what exactly position means is still cumbersome. For example, the information that “we
are at latitude 4,234W and longitude 30,123N” or “my current GSM cell ID is 1098” is
poor in terms of meaning for a user. Informations such as “I am in Morrocco”, “my
current location is in Coimbra” or “I am at work” are clearly richer and useful for a
wealth of applications. This is known as the “From Position to Place” problem
(Hightower, 2003) and is currently a hot topic in the Ubiquitous Computing area. The
primary goal of this PhD project is to study and develop methodologies that can
contribute to solving the problem just described. The approach expected will likely take
into account the user model, context and social interaction. This work is one of the
central topics of research of the Ubiquitous Systems Group of the AILab and has a high
potential of applicability in a range of state-of-the-art ubiquitous systems.
PhD Thesis Proposal: G2.16
Title: “Improvement of Algorithms for Indoor Location Supported on GSM and WiFi
Signatures”
Keywords: indoor location, GSM and WiFi signatures, location based services,
ubiquitous computing
Supervisors: Prof. Francisco Câmara (camara@dei.uc.pt)
Prof. Carlos Bento (bento@dei.uc.pt)
Summary:
Information on location is a main concern for ubiquitous computing. Many applications
of ubiquitous systems depend on location to achieve their goal. Although the problem of
positioning a device or person outdoor is reasonably solved with GPS technologies, the
problem of indoor location is much more challenging. Various technologies exist for
indoor location, but in general they need a dedicated infrastructure. Some researchers
followed a promising approach that consists on using the information provided by GSM
and WiFi equipment on the level of the signals received from the base stations
(signature) to infer the current location of the equipment. At our group we have a
consolidated research on this direction using GSM and WiFi signatures to infer current
location using case-based reasoning techniques. The theme for this thesis focus on
improving these algorithms, in terms of precision and accuracy, following concurrent
inference approaches, not necessarily restricted to case-based reasoning.
PhD Thesis Proposal: G2.17
Title: “Context Modelling from Data on Ubiquitous Computational Devices”
Keywords: context awareness, context models, inference of context models,
proactivity, ubiquitous devices
Supervisors: Prof. Francisco Câmara (camara@dei.uc.pt)
Prof. Carlos Bento (bento@dei.uc.pt)
Summary:
The growing availability of a wide range of sensors and communication services (such
as camera, gps, gsm or wireless) in devices such as PDAs or mobile phones is opening a
new world of possibilities in Ubiquitous Computing. The relatively new area of Context
Awareness is dedicated to model this information as well as to the design of possible
applications. In this PhD project, it is expected to explore the predictability associated
to the use of Ubiquitous devices, in other words, we believe it is possible to build user
models from use of context (and other) data and make inference about future
interactions from these models. This will bring intelligence and more ease of use to
Ubiquitous devices. This project is part of a central topic of research within the
Ubiquitous Group of the AILab: Context Awareness.
PhD Thesis Proposal: G2.18
Title: “Emotion Expression with Music”
Keywords: Affective computing, AI and Music, Computational Creativity
Supervisor: Prof. Amílcar Cardoso (amilcar@dei.uc.pt)
Summary:
Music assumes a central role in driving the emotional experience in a perse range of
situations, from computer games to cinema, theatre and many other artistic,
entertainment and educational setups. Composing a soundtrack for such scenarios
requires the ability to align the expected emotional effect of the music and sound effects
with the intended emotional experience. The task is particularly hard when a predefined
fixed script doesn’t exist and the course of action is decided in a dynamic way, like in
most computer games and other interactive applications.
Roughly, soundtrack production includes the composition of a set of music sections
and the collection of sound and music effects, which constitute inputs for a production
phase where they are sequenced/mixed/blended. All these phases are typically
conducted by the soundtrack composer, although for computer games several
techniques exist to provide real-time sequencing solutions.
The ultimate goal of this work is to investigate computational approaches for the
problem of generating a soundtrack given a variable “emotion spectrum” as input.
Several branches of the problem may be explored, depending on the student
background, including but not limited to: music classification and retrieval by
emotional effect, generation of emotionally affecting music, music alignment (e.g.,
according to rhythm, dynamics, emotional effect).
G3: Simulation and Information Technologies in
Education and Training
http://cisuc.dei.uc.pt/sitg/
PhD Thesis Proposal: G3.1
Title: “Alternative specification and visualization representations in initial
programming learning”
Keywords: computer science education; programming learning; alternative
representations.
Supervisor: Prof. Maria José Marcelino (zemar@dei.uc.pt)
Summary:
The main objective of this thesis is to study, propose and validate, on one hand, new
alternative forms of representation for algorithm and program specification and, on the
other, new alternative ways of algorithm and program visualization to support initial
programming learning and evaluate their impact on the quality of the achieved
students’ learning.
Initial programming learning is quite hard for the majority of students. It is usually
supported by one (or more) of three typical modes of algorithm/program representation:
pseudo code, flowcharts and code in a specific programming language. In what
concerns algorithm/program visualization several approaches have also been used:
variable log, debugging helps, simulated algorithm/program animation. Each student
has her/his own preferences about these representation and visualization metaphors.
There are particular types of programming problems that are mandatory in initial
programming learning and for which typical students’ solutions (good as well as
erroneous) have been identified. We believe that, although final programs must be
coded in one particular programming language, during initial learning stages many
programming students could benefit from the study and implementation of perse
alternative solution representations as well as visualizations, especially if they are more
close to students’ previous experience and context.
In the scope of this thesis student preferential alternative representations both at the
level of algorithm and program specification and of results visualization will be
identified and evaluated. After new forms will be developed and proposed in order to
cope with students more commonly found difficulties that traditional approaches can
not deal with. These new forms will be afterwards the object of thorough evaluation.
PhD Thesis Proposal: G3.2
Title: “Learning communities to support initial programming learning”
Keywords: computer science education; programming learning; learning
communities.
Supervisor: Prof. António José Mendes (toze@dei.uc.pt)
Summary:
Initial programming learning is known as a hard task to many novice students at
college level, leading to high failure and drop out in many courses. Many reasons can be
found for this scenario and several approaches have been proposed to facilitate
students’ learning. However, problems continue to exist and it is necessary to
investigate new solutions that may help programming students and teachers.
Learning communities’ concept exists for some time. It has been presented as a way to
create rich learning contexts where teachers, students and other people, namely
experts, can coexist and collaborate in the production of knowledge, consequently
leading to learning enhancement.
This thesis proposal includes first the study of representative learning communities’
successful cases and characteristics, and after the study, proposal and creation of a
learning communities support platform specially adapted to the needs of students
during programming learning. The platform and its utilization will undergo a full
evaluation, in order to access its success in promoting programming learning. It is
expected that this platform includes innovative characteristics, for example the
inclusion of virtual members that may interact with real members when necessary and
specially tailored features and tools that may improve the quality of programming
learning.
PhD Thesis Proposal: G3.3
Title: “Problem solving patterns and remediation strategies in programming learning”
Keywords: computer science education; programming learning; learning
communities.
Supervisors: Prof. António José Mendes (toze@dei.uc.pt)
Prof. Maria José Marcelino (zemar@dei.uc.pt)
Summary:
Initial programming learning is known as a difficult task to many novice students at
college level. In those courses it is common to use a set of typical problems to introduce
students to basic programming concepts and also to stimulate them to develop their
first programs and programming skills. This work is essential, since it should allow
beginners to develop the basic programming problem-solving skills necessary to be
further developed and refined later. So, this first learning stage is crucial to students’
performance in all programming related courses.
This thesis proposal includes a study about the different ways students approach these
typical basic problems, leading to the identification of common problem solving
patterns. Some of these patterns will be adequate, while others will not lead to the
development of correct solutions, being considered wrong or erroneous patterns that
must be identified and corrected in student’s strategies knowledge. Based on this
information, the thesis main objective will be the proposal, implementation and
evaluation of methods and/or tools that may identify novice students’ strategies,
categorize typical wrong patterns and common errors, and interact with them giving
personalized remediation feedback when necessary. The forms of this feedback must
also be studied, so that it becomes effective not only to help students to solve the
current problem, but mainly to help them to develop better approaches that may lead to
correct solutions in later problems and learning stages.
PhD Thesis Proposal: G3.4
Title: “Using Design Patterns in Modeling and Simulation”
Keywords: Model Reuse, Design Patterns, Parallel & Distributed Simulation
Supervisor: Prof. Fernando Barros (barros@dei.uc.pt)
Summary:
Software Design Patterns (SDPs) are a widely used technique in software development.
Time-based SDPs have been developed for building real-time software and its use is a
promising approach to build modeling and simulation software. This proposal intends
to develop new SDPs to help the development of reusable simulation models and
reusable simulation kernels able to deal with both conservative and optimistic parallel
& distributed approaches.
PhD Thesis Proposal: G3.5
Title: “Modeling and Simulation of Adaptive Sampling Systems”
Keywords: Numerical Simulation, Sampled-Based Systems
Supervisor: Prof. Fernando Barros (barros@dei.uc.pt)
Summary:
Adaptive step-size numerical methods permits to improve simulation performance while
yielding the same accuracy of fixed step-size methods. The use of asynchronous
numerical solvers permits to concentrate computation power in most demanding
models enabling larger systems to be represented. Digital Control and Digital signal
processing areas are currently exploiting multirate sampling and adaptive sampling
techniques as a more efficient alternative to conventional fixed sampling rate
approaches. In this proposal, we intend to develop new algorithms based on adaptive
sampling and make their application to numerical solvers, event detectors, and signal
and control systems.
G4: Adaptive Computation
http://cisuc.dei.uc.pt/acg/
PhD Thesis Proposal: G4.1
Title: “Intelligent Data Mining in GRID Technology”
Keywords: Machine Learning techniques such as Neural Networks, Support Vector
Machines and Clustering in data mining, prediction and recognition
Supervisor: Prof. Bernardete Ribeiro (bribeiro@dei.uc.pt)
Summary:
Grid technology emerged from distributed computing with the goal of generating
processing power for meeting users’ needs. To increase computing power, computing
resources are gathered across physical places. The idea was to match the unused
computing cycles with the needs created by applications and other users. This notion is
now a ubiquitous solution practised all around the world. It ensures continuous
computing availability despite scheduled maintenance, power outages, and unexpected
failures. The main idea of research is to prepare selected data mining algorithms,
preferably those originating in soft-computing, to be able to run in distributed
environments like clusters and global grids. Different techniques will be used, from
architectural parallelization of the models, data parallelization of the models and
parallel parameter-search methods for sequential models. An analysis will be given
focusing on a suitability of the techniques and particular distributed environments in
combination with the implemented data mining models. The implementations of
methods on distributed computing resources will be tested on wide variety of data bases
emerging from industry, medicine and internet in order to investigate their efficiency
and robustness.
PhD Thesis Proposal: G4.2
Title: “Learning from Side-Information and From Heterogeneous Data Sources”
Keywords: Machine Learning techniques such as Neural networks, Support vector
Machines and Clustering in data mining, prediction and recognition
Supervisor: Prof. Bernardete Ribeiro (bribeiro@dei.uc.pt)
Summary:
In the last decade we have witnessed an impressive growth of the on-line information,
mostly available through the Web, the intranets and other sources. It is estimated that
the Information density doubles every 12 to 15 months but the capacity of reading and
analysing it remains constant. This is not only due to better technology that allows for
fast acquisition, but also due to faster computer technology that allows exploiting the
data in machine learning tasks. It seems however that the growth of the data is more
explosive than the boost in computing power, and this evolution is improbable to
change when Moore’s law will saturate once the limits of electronics are approached.
Then only the algorithmics can make further speed-ups possible. Along with the
explosive growth of data availability, an increasing persity of the data types can be
observed. The questions how to deal with this heterogeneity and how to weight the
importance of different sources of data and information remains to be solved. Learning
from heterogeneous sources of data, and learning in semi-supervised learning settings
is the main theme of research. Applications of these settings abound, in the broad
domain of bioinformatics (mainly learning the heterogeneous information), machine
vision (mainly learning from side-information, for image and video segmentation), many
generic classification problems in the field of text mining and information extraction
and many others.
PhD Thesis Proposal: G4.3
Title: “Assigning Confidence Score in Page Ranking for Intelligent Web Search”
Keywords: Ranking, Text Mining, Machine Learning
Supervisor: Prof. Bernardete Ribeiro (bribeiro@dei.uc.pt)
Summary:
Web has become the main centre of research around the globe. Users face themselves
with an overload of data when a simple search is fed into Google or a similar web search
engine. A recurrent problem is to unveil the desired information from the wealth of
available search results. Ranking, which can be achieved by providing a meaningful
score for each classification decision, is important in most practical settings. Text
retrieval systems typically produce a ranking of documents and let a user decide how
far down that ranking to go. Several Learning Machine techniques allow the definition of
scores or confidences coupled with their classification decisions. The main idea of the
current proposal is to explore ranking systems based on Bayesian approaches to the
web learning problem that allow a refinement of existing systems. Moreover
classification systems can be improved by enriching information and information
representation with external background information, such as, ontology-related data.
Evaluation can be done on benchmarks, but also with real users defining the goals and
assessing the final results, including score changes in final ranking.
PhD Thesis Proposal: G4.4
Title: “Homecare Diagnosis of Pediatric Obstructive Sleep Apnea”
Keywords: homecare; obstructive sleep apnea; reduction of complexity; biosignals
processing; computational intelligence; automatic diagnosis.
Supervisor: Prof. Jorge Henriques (jh@dei.uc.pt)
Summary:
The main goal of this work is to investigate homecare solutions that could stratify
normal and apnea events for diagnostic purposes in children suspected for the presence
of obstructive sleep apnea syndrome.
Obstructive sleep apnea syndrome (OSAS) is a condition whereby recurrent episodes of
airway obstruction are associated with asphyxia and arousal from sleep. It is estimated
to affect between 1 and 3% of young children and its potential consequences include
excessive daytime somnolence, behavioral disturbances and learning deficits,
pulmonary and systemic hypertension, and growth impairment. The currently accepted
method for diagnosis of OSAS is overnight polysomnography (PSG), done in sleep
laboratories, where multiple signals are collected by means of face mask, scalp
electrodes, chest bands etc. It monitors different activities, including brain waves (EEG),
eye movement (EOG), muscle activity (EMG), heartbeat (ECG), blood oxygen levels and
respiration. However, the diagnosis of OSAS from these huge collection of data is
sometimes not straightforward to clinicians, since major relations between features and
consequents are most often very high dimensional, non-linear and complex. These
requirements impose the necessity of innovative signal processing techniques and
computational intelligent data interpretation methodologies, such as neural networks
and fuzzy systems. One of the main goal of this work is to provide clinicians with the
tools that can help them in their diagnosis.
Although PSG is considered the gold standard for diagnosis of OSAS, given the relatively
high medical costs associated with such tests and the insufficiency number of pediatric
sleep laboratories, PSG is not readily accessible to children in all geographic areas.
Thus, analysis of the validity of alternative diagnostic approaches should be done, even
assuming their accuracy is suboptimal. The second goal of this work points in this
direction. It aims investigating the viability to reduce the number and complexity of
measurements in order to make possible the stratification of OSAS in children natural
environment.
PhD Thesis Proposal: G4.5
Title: “Architectures and algorithms for real-time learning in interpretable neurofuzzy
systems”
Keywords: on-line learning; neuro-fuzzy systems; interpretability; machine learning
Supervisor: Prof. António Dourado (dourado@dei.uc.pt)
Summary:
The development of fuzzy rules to knowledge extraction from data acquired in real time
needs new recursive techniques for clustering to produce well designed fuzzy-systems.
For Takagi –Sugeno-Kang (TSK) systems this applies mainly to the antecedents, while
for Mamdani type it applies both for the antecedents and consequents fuzzy sets. To
increment pos-interpretability of the fuzzy rules, such that some semantic may be
deduced from the rules, pruning techniques should be developed to allow a humaninterpretable
labelling of the fuzzy sets in the antecedents and consequents of the rules.
For this purpose convenient similarity measures between fuzzy sets and techniques for
merging fuzzy rules should be developed and applied. The applications envisaged are in
industrial processes and medical fields.
PhD Thesis Proposal: G4.6
Title: “Intelligent Monitoring of Industrial Processes with application to a Refinery”
Keywords: intelligent process monitoring; multidimensional scaling; computational
intelligence; clustering
Supervisor: Prof. António Dourado (dourado@dei.uc.pt)
Summary:
High dimensional data in industrial complexes can be profitably used for advanced
process monitoring if it is reduced to a dimension where human interpretability is easily
verified. Multidimensional scaling may be used to reduce it to two or three dimensions if
appropriate measures of similarity/dissimilarity are developed. The measures express
the distance between attributes, the essence of the information, and a similar difference
should be guaranteed in the reduced space in order to preserve the informative content
of the data. Research of appropriate measures and reduction method is needed.
In the reduced space, classification of the actual operating point should be dome
through appropriate recursive clustering and pattern recognition techniques. The
classification is intended to evidence clearly the quality level of the actual and past
operating points in such a way that the human operator finds in it a useful decision
support system for the daily operation of the mill. The work has as applications the
process of visbreaker in the Galp Sines Refinery.
G5: Dependable Systems
http://cisuc.dei.uc.pt/dsg/
PhD Thesis Proposal: G5.1
Title: “Grid Computing Support for Distributed Unreliable Networks”
Keywords: Grid computing, parallel programming, distributed computing, desktop
grids
Supervisor: Prof. Paulo Marques (pmarques@dei.uc.pt)
Summary:
Nowadays there is a huge need for parallel computing. Researchers in areas like
biology, physics, computer networks, among others, need to perform a huge number of
computer experiments in order to gather results. At the same time, there is an increase
need for on-demand-experiments (e.g. “I want to know this result now!”). Researchers
what to quickly run an experiment, which may involve thousands of calculations, in
order to know how to set “yet another parameter” or which part of a search space to
explore. They are not willing to wait days or weeks for getting simple answers that just
guide the direction of their research – they want them at the touch of a button.
Although many frameworks for parallel computing exist (e.g. MPI, PVM, OpenMP), they
are typically thought for cluster computing. This raises a problem because not all
researchers have a readily available cluster for performing experiments. Even when they
do have access to a cluster, in most cases, the relative size of the cluster is small to the
number of researchers wanting to use the resources, which increases the turn-around
time for running experiments. At the same time, cluster environments are not so
compatible with the increasing requests from researchers for on-demand-computing.
The alternative is trying to run scientific applications in non-dedicated machines, also
known as desktop grids (e.g. classroom/office PCs), which abound in organizations.
But, in that case, the use of MPI and similar frameworks is very inappropriate (e.g. the
fault-model of MPI implies that if one process crashes, the whole computation dies –
which is incompatible with the unreliability of those computers!).
Although many frameworks have been developed for running parallel applications on
non-dedicated machines (e.g. Condor, BOINC, Alchemi, etc.), in order to coupe with the
unreliability of the machines, they typically don’t allow communication between nodes
and the computations are based on bounded work units assigned to the nodes. For
instance, using most of these frameworks, it’s quite difficult to write grid-based
algorithms (e.g. for calculating the air flow that passes through an airplane wing) or a
global back-tracking algorithm (e.g. for optimizing a path throw a network).
Finally, researchers in other areas than computer science are now coming to terms with
more modern programming environments and easier to use computer languages. For
instance, Python and Numeric Python, as well as Matlab and Mathematica, are quite
popular with biologists, physicists and even social scientists. It’s now time to move
parallel programming beyond C and Fortran.
Clearly, research is needed on programming models and infrastructures that allow
parallel programs to run on unreliable distributed networks, and at the same time
allows them to be written in modern, easy-to-use and productive computer languages.
Proposal
This PhD dissertation will consist in investigating, implementing and evaluating a
programming model that allows parallel programs to be easily written and reliably
execute on desktop grids. It is also a specific objective of the thesis to deviate from the
traditional message passing paradigm and from the remote method invocation models
of creating distributed applications.
The framework to be developed will address questions as (but not necessarily all): a)
programming easiness; b) global reliability of the computation in presence of failures; c)
access to stable storage for reading and writing results; d) security; e) deployment and
monitoring; f) visualization.
In this context, some technologies will be interesting to consider and explore:
distributed hash tables; P2P routing and service discover; erasure codes; distributed
map/reduce programming models; self-organization; consensus; group membership
and election algorithms; mobile code, among others.
PhD Thesis Proposal: G5.2
Title: “Sensor Networks for Space Exploration”
Keywords: Sensor networks, Ad-hoc networking, Space exploration
Supervisor: Prof. Paulo Marques (pmarques@dei.uc.pt)
Summary:
Sensor networks are currently a hot topic in distributed systems research. A sensor
network consists in tens or hundreds of inexpensive sensing devices, typically not much
larger than a coin, that are able to gather information from the environment, coordinate
among themselves, and relay that information to a remote location. This type of systems
has a huge number of application areas, like earth observation, environment
monitoring, security, medical healthcare, among others. Typical deployment scenarios
include scattering devices through a forest for detecting the start of wildfires; placing
devices throughout a river basin for detecting pollutant dumping; or even equip cars
with such devices for detecting and warn about immediate collision danger.
One extremely interesting application scenario for sensor nets is space exploration. It is
quite easy to imagine that deploying hundreds of sensors over some kilometers while a
space probe is descending to Mars can be quite beneficial. Instead of being limited to
gather data where the spacecraft lands or where their rovers can move, truly distributed
data acquisition can take place. The same thing applies, for instance, in orbit for
gathering data and performing distributed experiments in Earth’s upper atmosphere, or
even to gather data from orbiting probes (the current largest European satellite is the
size of a TIR truck!).
Using sensor networks in space presents unique and challenging problems. Space is
quite a harsh place: radiation abounds causing software and hardware failures,
temperature is typically well below zero, electromagnetic interference makes radio links
quite unreliable. Since normally these devices are small, cheap and disposable,
typically they are quite limited in terms of computational power, energy and
transmission bandwidth. This makes engineering this type of networks for reliability
difficult, especially for space applications.
Proposal
This PhD dissertation will consist in investigating, implementing and evaluating
algorithms for fault-tolerance in sensor networks for harsh, Byzantine, environments,
for space exploration. In particular, the thesis will focus on exploring the spectrum of
possibilities for achieving different degrees of reliability in computationally limited
devices, when this reliability comes at cost of spending energy and having to
communicate with other sensing nodes.
Currently, it is envisioned that this work will be carried out in the context of a research
project where other partners with develop the sensing platform hardware and also
provide a realistic context for fault injection and testing, possibly in a collaboration with
the European Space Agency (to be confirmed).
PhD Thesis Proposal: G5.3
Title: “Self-Healing Techniques for Application Servers”
Keywords: Autonomic computing, self-healing, software aging, rejuvenation,
dependability.
Supervisor: Prof. Luís Moura e Silva (luis@dei.uc.pt)
Summary:
One of the actual big-challenges of the computer industry is to deal with the complexity
of the systems. The Autonomic Computing initiative driven by IBM defined the following
functional areas as the cornerstone of an autonomic system: self-configuration, selfhealing
, self-optimization and self-protection. The self-healing property refers to the
automatic prediction and discovery of potential failures and the automatic correction to
possibly avoid downtime of the computer system. This leads to the vision of “computers
that heal themselves” and do not depend so much on a system manager to take care of.
While there has been some interesting work on self-healing techniques for missioncritical
systems there is a long way to achieve that goal in commercial off-the-shelve
(COTS) servers running Apache/Linux, Tomcat, JBoss, Microsoft .Net. The purpose of
this PhD is to study and propose low-cost and highly-effective self-healing techniques
for these application servers.
One of the potential causes of failures in 24x7 server systems is the occurrence of
software aging. The phenomena should be studied in detail together with high-level
techniques for application-level failure detection. Some mathematical techniques should
be applied to detect software aging and to forecast the potential time for the failure of
the server system. When the aging is detected the server system should apply proactively
a software rejuvenation technique to avoid the potential crash and to keep the
service up and running. Techniques for micro-rejuvenation should be further studied to
avoid downtime of the server. The final result of this PhD should be a set of software
artifacts and the refinement of data analysis techniques to apply in COTS application
servers in order to predict failures and software aging in advance and to apply some
corrective action to avoid a server crash.
Proposal
This PhD will comprehend the following initial tasks: (1) State-of-the-art about
Autonomic Systems, Self-healing, Software Aging, Software Rejuvenation, Microrebooting
and Dependability Benchmarking; (2) Machine learning techniques to forecast
the failures and software aging; (4) Application-level techniques for failure prediction
and early detection; (5) Micro-rejuvenation techniques for application servers; (6)
Extension of the techniques SOA-based and N-tier applications; (7) Dependability
benchmarking; (8) Implementation of an experimental framework; (9) Analysis of
experimental results.
PhD Thesis Proposal: G5.4
Title: “Dependable and Self-Managed VoIP Infrastructures”
Keywords: Autonomic computing, Self-healing, Dependability, Peer-to-Peer, VoIP,
QOS.
Supervisor: Prof. Luís Moura e Silva (luis@dei.uc.pt)
Summary:
Peer-to-peer techniques have been widely applied for decentralized file-sharing in the
internet, distributed computing, content distribution and to support applications like
Voice-over-IP. The most popular example is Skype that is based on a supernode-based
P2P network. Since in those P2P networks some of the server-based services can be
provided by the client machines there is mandatory need to provide the P2P
infrastructure with techniques for self-configuring, self-healing and self-management,
in the line of Autonomic Computing systems.
The goal of this thesis is to devise and study some software techniques to enhance the
dependability and autonomic computing capabilities of VoIP infrastructures. In a first
step, the goal is to provide fault-tolerant mechanisms for VoIP servers, by studying the
reliability of staged-event servers, the use of software fail-over mechanisms, studying
failure-analysis and prediction to anticipate the occurrence of software aging inside a
server, the use of micro-rebooting of service components and complementary
techniques to enhance the self-healing capabilities of a VoIP server. In the second phase
of the thesis, there should be interesting to study the usage of high-level reconfiguration
techniques that will require architectural changes in the VoIP infrastructure and some
reconfigurable usage of SIP/RTP protocols with the ultimate goal of providing a higher
QOS and transparency of failures for the end-user of the application. The result should
be a highly-reliable P2P infrastructure that can be used in to enhance the QOS of VoIP
applications like Skype.
Proposal
This PhD will comprehend the following initial tasks: (1) State-of-the-art about Peer-to-
Peer networks, VoIP infrastructures, STUN/TURN servers, SIP and RTP protocols, Selfhealing
techniques, Staged-event servers, Load-balancing and Server Reconfiguration;
(2) Application-level techniques for failure prediction and early detection; (3) Prediction
of server availability; (4) Software techniques for load-balancing (wackamole project); (5)
Study of self-healing techniques for stagged-event servers (SEDA project); (6) Microrebooting
techniques for VoIP servers; (7) Reconfiguration techniques for VoIP servers;
(8) Support for protocol reconfiguration; (9) Construction of the framework and results;
(10) Data analysis.
PhD Thesis Proposal: G5.5
Title: “Wired self-emerging ad hoc network”
Keywords: peer-to-peer, ad hoc, distributed hash table.
Supervisor: Prof. Filipe Araújo (filipius@dei.uc.pt)
Summary:
In recent years computer communication is departing from the client-server
architecture and moving increasingly more toward a peer-to-peer architecture. One
aspect that characterizes this kind of interaction is the opportunistic participation of
many of the peers: they connect to the network for only a few moments, just to discover
and download (or not) what they are looking for and then they disconnect. Interestingly,
mobility and battery exhaustion can reproduce this same trend in wireless ad hoc
networks, comprised of devices that use radio broadcast to communicate.
While wired peer-to-peer and wireless ad hoc networks share a number of common
features, like self-configuration, decentralized and fault-tolerant operation, they have
however an important difference: wired peer-to-peer networks run as overlay networks
on top of the IP infrastructure. This raises the following question: can we take the
paradigms from wireless networks and create IP-less self-organizing wired networks?
Our goal is to plug-in and out new devices or even entire networks from the wired
infrastructure in a scalable and decentralized way and without the need for any a priori
configuration. In contrast, current IP networks can only scale, because they are highly
hierarchical and they require a considerable amount of human assistance. As a
consequence they are often highly congested, expensive to maintain and unreliable.
The fundamental difference between the solution we seek and wireless ad hoc networks
has to do with available bandwidth. In fact, the most important constraint that makes
collection of routing information so challenging and that limits the pace of change of
topology in wireless ad hoc networks is the (lack of) available bandwidth. Available
bandwidth is a very scarce resource, because it is shared among all the nodes. This
makes it theoretically impossible to create a wireless ad hoc network that scales with
the number of nodes. As a consequence, algorithms for wireless ad hoc networks are
often localized or have, at most, very limited information of distant regions of the
network. This is very unlike the situation in wired networks: for the same pace of
topological change, the supply of bandwidth is not shared and it is much larger. This
paves the way for better and more powerful solutions, which, we believe are largely
unexplored in literature.
Proposal:
This PhD work encompasses the following tasks: (a) review of the state-of-the-art; (b)
design of the architecture; (c) evaluation of the scalability of the architecture (admissible
number of nodes and topological changes versus available bandwidth); (d) exact and
range-based lookup algorithms that leverage on previous work on distributed hash
tables and peer-to-peer file-sharing applications; (e) design of an interconnection
infrastructure, to connect islands of wired ad hoc networks with the IP network.
PhD Thesis Proposal: G5.6
Title: “Fast Moving Wireless Ad Hoc Nodes”
Keywords: peer-to-peer, wireless ad hoc, wireless infrastructured, Wireless Access
for the Vehicular Environment (WAVE).
Supervisors: Prof. Filipe Araújo (filipius@dei.uc.pt)
Prof. Luís Moura e Silva (luis@dei.uc.pt)
Summary:
In recent years we have assisted to an increasing interest in wireless networks. While
most current applications seem to be set for sensor networks, we can foresee many
other applications for mobile ad hoc or mixed ad hoc/infrastructured networks, where
nodes are mainly mobile and communication goes beyond simple data gathering of a
sensor network. For instance, applications can enhance the behavior of a crowd by
providing additional services to users holding mobile wireless devices, like search for a
given person that is momentarily lost, search for a person that matches some social
interests, exchange of perse information, of a product, etc. Another context that is
extremely promising is that of a spontaneous network formed by cars in a road,
enriched with some infrastructure that is able to provide traffic, weather and other
information to drivers. By letting cars share their information, it may be possible to save
significant costs in the infrastructure and still considerable improve the quality and
quantity of information.
In this PhD work we want to leverage on some existing routing algorithms for wireless
ad hoc networks and make them work on particular environments with specific
patterns of mobility. Interestingly, in networks with a high degree of mobility it is often
possible to increase the speed of the flow of information, because mobility creates more
opportunities to exchange this information. In particular, we want to consider a
scenario where the network is comprised of fast-moving cars equipped with IEEE
802.11p network adapters (Wireless Access for the Vehicular Environment – WAVE).
This is a case where part of the information is created and sent to some points of the
infrastructure through a chain of nodes, while at the same type, cars can also introduce
new information in the network, for instance by signaling their presence to cars in
front, in the rear or to cars traveling in the opposite direction. In particular, the
information shared with cars going in the opposite direction first and then with the base
stations located along the road is of paramount utility as this has the potential to
propagate very accurate data of traffic jams or accidents at virtually no cost. We expect
to use similar principles to more complex but slower-moving networks comprised of
people with handheld or other wireless devices walking in crowds.
Proposal:
This PhD work encompasses the following tasks: (a) review of the state-of-the-art; (b)
design of routing algorithms for environments with high mobility; (c) design of
information-sharing applications for environments with high mobility; (d) simulation in
real environments.
PhD Thesis Proposal: G5.7
Title: “Detecting Software Aging in Database Servers”
Keywords: Software aging, software rejuvenation, autonomic computing, database
management systems, dependability benchmarking
Supervisors: Prof. Marco Vieira (mvieira@dei.uc.pt)
Prof. Luís Moura e Silva (luis@dei.uc.pt)
Summary:
One of the main problems in software systems that have some complexity is the
problem of software aging, a phenomenon that is observed in long-running applications
where the execution of the software degrades over time leading to expensive hangs
and/or crash failures. Software aging is not only a problem for desktop operating
systems: it has been observed in telecommunication systems, web-servers, enterprise
clusters, OLTP systems, spacecraft systems and safety-critical systems.
Software aging happens due to the exhaustion of systems resources, like memory-leaks,
unreleased locks, non-terminated threads, shared-memory pool latching, storage
fragmentation, data corruption and accumulation of numerical errors. There are several
commercial tools that help to identify some sources of memory-leaks in the software
during the development phase. However, not all the faults can be avoided and those
tools cannot work in third-party software modules when there is no access to the
source-code. This means that existing production systems have to deal with the
problem of software aging.
The natural procedure to combat software aging is to apply the well-known technique of
software rejuvenation. Basically, there are two basic rejuvenation policies: time-based
and prediction-based rejuvenation. The first applies a rejuvenation action periodically,
while the second makes use of predictive techniques to forecast the occurrence of
software aging and apply the action of rejuvenation strictly only when necessary.
The goal of this PhD Thesis is to study the phenomena of software aging in commercial
database engines, to devise and implement some techniques to collect vital information
from the engine and to forecast the occurrence of aging or potential anomalies. With
this knowledge the database engine can apply a controlled action of rejuvenation to
avoid a crash or a partial failure of its system. The ultimate goal is to improve the
autonomic computing capabilities of a database engine, mainly when subjected to high
workload and stress-load from the client applications.
Proposal:
The PhD work will comprehend the following initial tasks: (a) overview of the state-ofthe-
art about software aging, rejuvenation and dependability benchmarking; (b)
development of a tool for dependability benchmarking of database engines; (c)
development of a workload and stress-load tool for databases; (d) infrastructure of
probes (using Ganglia) to collect vital information from a database engine; (e)
development of mathematical techniques to forecast the occurrence of software aging
(time-series analysis, data-mining, machine-learning, neural-networks);(f) experimental
study. Analysis of results; (g) adaptation of rejuvenation techniques for database
engines; (h) writing of papers;
PhD Thesis Proposal: G5.8
Title: “Security benchmarking of COTS components”
Keywords: Software reliability, Security benchmarking, Experimental evaluation,
Dependability benchmarking
Supervisors: Prof. Henrique Madeira (henrique@dei.uc.pt]
Prof. João Durães (jduraes@dei.uc.pt)
Summary:
One of the main problems in software systems is the vulnerability to malicious attacks.
Complex systems and systems that have high degree of interaction with other systems
or users are more prone to be successfully attacked. The consequences of a successful
attack are potentially very severe and may include the theft of critical-mission
information and trade secrets. Given the pervasive nature of software systems in
modern society, the issue of security and testing for vulnerability to attacks is an
important research area.
The vulnerability of software systems are caused by several factors. Two of these factors
are the integration of third-party off-the-shelf components to build larger components,
and bad programming practices. The integration of third-party generic-purpose
components may introduce vulnerabilities in the larger system due to interface
mismatch between the components that may be exploited for attacks. Bad programming
practices may lead to weaknesses that may be exploited by tailored user inputs. Testing
a system or component against malicious attacks is a difficult problem and is currently
an open research area. Testing for vulnerabilities to malicious attacks can not be
performed as traditional testing because there is no previous knowledge about the
nature of the attacks. However, these attacks follow a logic based on exploiting possible
weaknesses inside the software and this logic can be used to forecast the existence of
vulnerabilities.
The goal of this PhD Thesis is to study the phenomena of attacks to software systems
and devise a methodology to assess the vulnerability to these attacks. This includes the
proposal of experimental techniques to test systems and components following the logic
of dependability benchmarking and experimental evaluation. It is expected that at the
conclusion of the Thesis there is case study with practical results of assessment
security and vulnerability forecasting for comparison purposes. Web servers are
suggested as on of possible case studies. Fault injection techniques and robustness
testing techniques should be considered as enabling techniques for the purposes of the
Thesis.
Proposal:
The PhD work will comprehend the following initial tasks:
(a) overview of the state-of-the-art about software security, software vulnerabilities,
software defects, validation methods, robustness testing and dependability
benchmarking;
(b) development of methods and tools for analysis of the identification of patterns
related to vulnerabilities and the automated testing of the possible vulnerabilities (case
studies include web-servers)
(c) proposal of generic test methodologies for evaluation of software vulnerability to
malicious attacks based on software defects and program pattern analysis for system
comparison purposes;
(d) proposal of formal methodologies for experimental assessment of security and
vulnerability forecasting on third-party (black-box) software components;
(e) development of experimental infrastructure of tools for practical demonstration of the
above to real systems (case studies include web-servers);
(f) experimental study. Analysis of results;
(g) writing of papers;
Target conferences to publish papers:
- Dependable Systems and Networks (DSN)
- International Conference on COTS-Based Software System (ICCBSS)
- International Conference on Computer Safety, Reliability and Security (SAFECOMP)
- International Symposium on Software Reliability Engineering (ISSRE)
G6: Communications and Telematics
http://cisuc.dei.uc.pt/lct/
PhD Thesis Proposal: G6.1
Title: “Security in Wireless Sensor Networks”
Keywords: Sensor Networks, Security, Mobility
Supervisor: Prof. Jorge Sá Silva (sasilva@dei.uc.pt)
Summary:
Although in the last years, we witnessed the increase in processing capabilities and in
bandwidth of communication systems, several researchers consider that, in a near
future, an inversion of trends will occur. These new computational systems will not
consist of devices with higher processing power, but simply networks of sensors.
Wireless Sensor Networks (WSNs) are composed by a high number of nodes, each one
equipped with a microprocessor, low memory and a basic communication system.
The integration of WSNs in the Internet will revolutionize several concepts and it will
require new paradigms. More recently, a new group was created at the IETF, the
6LoWPAN, which is responsible to produce problem statements, assumptions and goals
for network elements with restricted requirements such as limited power, in which the
WSNs can be included.
However, security mechanisms for these networks are scarce and inefficient. The
research work of this PhD program will comprise the study and the proposal of new
models for the security of these new Internet elements. This is particularly important in
mobile environments, as the research community generally assumed sensors as static
nodes. The new WSN protocols should consider security issues to protect against
eavesdropping and malicious behavior.
PhD Thesis Proposal: G6.2
Title: “Multicast in Next Generation Networks”
Keywords: Multicast, 4G Networks, Mobility
Supervisor: Prof. Jorge Sá Silva (sasilva@dei.uc.pt)
Summary:
Multicast communication in the Internet has deserved an increasing attention in the
last few years. Nowadays, there are more and more applications that require
communication systems with multipoint communication capabilities. Multicast
communication reduces both the time it takes to send data to a large set of receivers
and the amount of network resources required to deliver such data.
The appearance of such new applications, with multicast requirements, evidenced the
need of truly multicast protocols in the IP layer. However, traditional solutions were too
complex to implement and only few network equipments support them.
The purpose of this research work is to study new multicast paradigms that offer simple
solutions for the Next Generation Internet, that will include elements such laptops,
PDAs, mobile phone and sensors. To date, the design goals for multicast protocols in
wired or wireless environments haven’t included sensor nodes. However, this will be
crucial as sensors will be preponderant elements in the Next Generation Internet.
PhD Thesis Proposal: G6.3
Title: “Routing for Resilience in Ambient Networks”
Keywords: Routing, resilience, ambient networks
Supervisors: Prof. Edmundo Monteiro (edmundo@dei.uc.pt)
Prof. Marilia Curado (marilia@dei.uc.pt)
Summary:
The new types of applications and technologies used for nowadays communication
among users, and the persity of types of users have shown that traditional routing
paradigms are not capable of coping with these recent realities. Therefore, the role of
routing in IP networks has shifted from single shortest path routing to multiple path
routing subject to multiple constraints such as Quality of Service requirements and
fault tolerance. Moreover, traditional routing protocols have several problems
concerning routing information distribution which compromises routing decisions.
Namely, routing decisions that are based on inaccurate information due to bad routing
configurations either caused by faulty or malicious actions will cause severe disruption
in the service that the network should provide.
These are particularly important issues in networks that involve different types of
communication devices and media, such as happens in ambient networks. Ambient
networks pose an additional challenge to routing protocols, since network composition
changes very often when compared to traditional IP networks, and networks are
expected to cooperate among each other on-demand without relying on previous
configuration. Moreover, associated with the dynamic structure of ambient networks,
traffic patterns in ambient networks also change very often due to the composition and
decomposition of network structure.
The work proposed for this thesis aims at studying the existing vulnerabilities of actual
routing protocols used in the Internet and to propose a resilient routing scheme that
overcomes these weaknesses in order to improve network availability and survivability.
The work will comprise the study of the state of the art of routing protocols for
resilience, the characteristics of ambient networks, and the proposal of enhancements
to existent routing schemes in order to improve the contribution of the routing protocol
for the resilience of ambient networks.
The research work of the PhD candidate will be included in the European Union
Integrated Project WEIRD (WiMAX Extension to Isolated Research Data networks -
http://www.ist-weird.eu).
PhD Thesis Proposal: G6.4
Title: “Community networks connectivity and service guarantees”
Keywords: Mobility, nomadicity, community networks
Supervisor: Prof. Fernando Boavida (boavida@dei.uc.pt)
Summary:
A number of recent technological developments have enabled the formation of wireless
community-wide local area networks. Dispersed users (residents or moving users)
within the boundaries of a geographical region (neighbourhood or municipality) form a
heterogeneous network and enjoy network services such as Internet connectivity. This
environment, named Community Networks, is well suited for both traditional Internet
access and the deployment of peer-to-peer services.
Achieving and retaining connectivity in this highly heterogeneous environment is a
major issue. Although the technology advances in wireless networks are fairly mature,
one further step in the management of Community Networks is to provide mobility and
nomadicity support. Nomadicity allows connectivity everywhere while mobility includes
the maintenance of the connections and sessions while the node is moving from one
place to another. Mobility and nomadicity in community and home networks still place
several challenges, as these environments are highly heterogeneous.
Seamless handover between different layer two technologies is still a challenge.
Seamless multimedia content distribution to the home may include several network
technologies, such as WLAN, power-line, GRPS or UMTS. Thus, the inter-layer issues
involved are complex and a lot of work is required to match MIP6 with them in order to
provide seamless mobility for multimedia information. The ability to support sessions
on multiple access networks is another open issue. These issues will be the central
concern of the proposed PhD work.
The research work of the PhD candidate will be included in the European Union IST
FP6 CONTENT Network of Excellence (CONTENT – Content Networks and Services for
Home Users, http://www.ist-content.eu) and will be carried out in close cooperation
with a foreign institution.
G7: Databases
http://gbd.dei.uc.pt/
PhD Thesis Proposal: G7.1
Title: “The Optimization Problem (OP) for QoS-compliance in Systems Engineering”
Keywords: QoS-Broker, Distributed Programming, SLAs, Optimization
Supervisor: Prof. Pedro Furtado (pnf@dei.uc.pt)
Summary:
The capacity to monitor and control QoS-parameters and their lifecycle is an important
aspect for today’s mature systems and software. Quality-of-Service and automatic
adaptation are also at the centre of most current developments in systems and
software. Another frequent issue in today’s parallel, distributed, mobile and in generic
networked systems is the optimization of content placement and replication and its
applications. Our group has been working on both issues and their relationship and
has two projects running. We have interesting and innovative proposals that make very
good PhD works. We also have some basic algorithmic and solver pilot prototypes for
(OP) and for QoS functionality as starting points and a detailed knowledge of current
and promising work. Besides these issues, the proposal is an opportunity for the PhD
candidate to learn and work with technologies such as condor, java and solver software,
and to work on exciting algorithms within a helpful team.
PhD Thesis Proposal: G7.2
Title: “QoS-Brokering Lifecycle with Automated Monitoring for Generic Applications
and Networked Data Intensive Systems”
Keywords: QoS-Broker, Distributed Programming
Supervisor: Prof. Pedro Furtado (pnf@dei.uc.pt)
Summary:
The Generic QoS-Broker is a piece of Software that can seat anywhere in any system
and whose purpose is to make contracts, monitor and react with any piece of software
in any desired manner to provide factual or optimistic Quality of Service guarantees. It
can also interact with lower level QoS-Brokers, so that in principle any desired QoS
objective can be met in any context. The questions that must be answered are: how
does it work? Given certain contexts, how is it applied? (e.g. mobile, real-time, data
management, etc). Our group has been busy working on some of these issues and has
two projects related to this issue. We have some initial prototypes for these systems. As
a result, we have quite a few interesting and innovative proposals that make very good
PhD works.
PhD Thesis Proposal: G7.3
Title: “Automatic Time-prediction and QoS in Distributed Data-Intensive Systems”
Keywords: QoS-Broker, Distributed Programming
Supervisor: Prof. Pedro Furtado (pnf@dei.uc.pt)
Summary:
The capacity to monitor and control QoS-parameters is an important aspect for today’s
mature systems and software. Quality-of-Service and automatic adaptation are also at
the centre of most current developments in systems and software. There are several
works on these issues. Our group has been working on mixing generic QoS with data
services. We would like to explore this in a typical current distributed computing
platform. The proposal is an opportunity for the PhD candidate to learn and work with
technologies such as condor and java, and to work on exciting algorithms within a
helpful team. Related Projects: Adapt-DB, current proposal in cooperation with other
CISUC group.
PhD Thesis Proposal: G7.4
Title: “Optimal Caching, Replication, Placement and Compression of Streams for
Multimedia Streaming in Mobile and Heterogeneous Systems”
Keywords: QoS, Replication, Placement, Caching
Supervisor: Prof. Pedro Furtado (pnf@dei.uc.pt)
Summary:
Given the rise and convergence of network technologies and of streaming and rich
media content delivery, the caching, replica placement and compression of streams
become increasingly important issues. Given our expertise in replica management,
caching and QoS brokering, we expect to produce innovative proposals in this context
that make very good PhD works. Related Projects: current proposal, in cooperation with
other CISUC group.
PhD Thesis Proposal: G7.5
Title: “Predictability, Portability, Auto-Adaptability and Re-Organization of Data
Services”
Keywords: RDBMS, Replication, Placement, Caching, Query Processing
Supervisor: Prof. Pedro Furtado (pnf@dei.uc.pt)
Summary:
Auto-adaptability of Services and Applications is a hot topic currently. We have
developed work on both run-anywhere data management services and generic QoSBroker
Architectures. Now we are particularly interested in providing automatic tools to
determine placement, replica and other adaptation strategies based on our monitoring
and history-analysis capacity. We have a basic prototype and several initial ideas on
these issues. As a result, we have quite a few interesting and innovative proposals that
make very good PhD works. Related Projects: Auto-DWPA.
PhD Thesis Proposal: G7.6
Title: “Data Anywhere, at Anytime with QoS in Heterogeneous Networks”
Keywords: QoS, Replication, Placement, Caching
Supervisor: Prof. Pedro Furtado (pnf@dei.uc.pt)
Summary:
The current trend towards mobile and ubiquitous computing should fundamentally
change our concepts of mostly static data storage and access. In the future, access will
increasingly be mobile from multiple devices, places and through different mediums
and the user will not define or worry about where and how the access happens or where
the contents are located. This represents a paradigm shift for the user from specifying
“where and how” to specifying “Object Properties”. UbiData is an automatic transparent
manager for user content through QoS definition. Users specify all kinds of objects,
access and resource properties and the system handles the objects automatically
(storage, consistency, availability, other QoS parameters). Related Projects: Adapt-DB,
current proposals;
PhD Thesis Proposal: G7.7
Title: “QoS and Replica-based Strategies for QoS-Brokering in Adaptable Data
Services”
Keywords: QoS, Transactional Systems
Supervisor: Prof. Pedro Furtado (pnf@dei.uc.pt)
Summary:
Auto-adaptability of Services and Applications is a hot topic currently. We have
developed work on both run-anywhere data management services and generic QoSBroker
Architectures. Now we are working in adopting interesting QoS and Replication
strategies in transactional environments. As a result, we have quite a few interesting
and innovative proposals that make very good PhD works. Related Projects: Adapt-DB;
PhD Thesis Proposal: G7.8
Title: “Timely ACID Transactions in DBMS”
Keywords: Databases, transaction processing, performance and QoS, timely
transactions, real-time databases, fault-tolerance
Supervisors: Prof. Marco Vieira (mvieira@dei.uc.pt)
Prof. Henrique Madeira (henrique@dei.uc.pt)
Summary:
On time data management is becoming a key difficulty faced by the information
infrastructure of most organizations. A major problem is the capability of database
applications to access and update data in a timely manner. In fact, database
applications for critical areas (e.g., air traffic control, factory production control, etc.)
are increasingly giving more importance to the timely execution of transactions.
Database applications with timeliness requirements have to deal with the possible
occurrence of timing failures, when the operations specified in the transaction do not
complete within the expected deadlines. For instance, in a database application
designed to manage information about a critical activity (e.g., a nuclear reactor), a
transaction that reads and store the current reading of a sensor must be executed in a
short time as the longer it takes to execute the transaction the less useful the reading
becomes. This way, when a transaction is submitted and it does not complete before a
specified deadline that transaction becomes irrelevant and this situations must be
reported to the application/business layer in order to be handled in an adequate way.
In spite of the importance of timeliness requirements in database applications,
commercial DBMS do not assure any temporal properties, not even the detection of the
cases when the transaction takes longer than the expected/desired time. The goal of
this work is to bring timeliness properties to the typical ACID (atomicity, consistency,
integrity, durability) transactions, putting together classic database transactions and
recent achievements in the field of real time and distributed transactions. This work will
be developed in the context of the TACID (Timely ACID Transactions in DBMS) research
project, POSC/EIA/61568/2004, funded by FCT.
Proposal
The PhD work will comprehend the following initial tasks: (a) overview of the state-ofthe-
art about timely computing and real-time databases; (b) characterization of timed
transactions; (c) analysis of DBMS core implementations; (d) infrastructure to support
timely execution of ACID transactions; (e) development of mathematical techniques to
forecast transactions execution times; (f) Implementation and evaluation; (g) writing
papers;
PhD Thesis Proposal: G7.9
Title: “Security Benchmarking for Transactional Systems”
Keywords: Security, benchmarking, database management systems, transactional
systems
Supervisors: Prof. Marco Vieira (mvieira@dei.uc.pt)
Prof. Henrique Madeira (henrique@dei.uc.pt)
Summary:
One of the main problems faced by organizations is the protection of their data against
unauthorized access or corruption due to malicious actions. Database management
systems (DBMS) constitute the kernel of the information systems used today to support
the daily operations of most organizations and represent the ultimate layer in
preventing unauthorized access to data stored in information systems. In spite of the
key role played by the DBMS in the overall data security, no practical way has been
proposed so far to characterize the security in such systems or to compare alternative
solutions concerning security features. Benchmarks are standard tools that allow
evaluating and comparing different systems or components according to specific
characteristics (e.g., performance, robustness, dependability, etc.). In this work we are
particularly interested in benchmarking security aspects of transactional systems.
Thus, the main goal is to research ways to compare transactional systems from a
security point-of-view. This work will be developed in the context of a research
cooperation with the Center for Risk and Reliability of the University of Maryland, MA,
USA. During this work the student will have the opportunity to visit the University of
Maryland in order to carry out joint work with local researchers.
Proposal
The PhD work will comprehend the following initial tasks: (a) overview of the state-ofthe-
art about security, security evaluation, and dependability benchmarking; (b)
definition of a security benchmarking approach for transactional systems; (c) study of
attacks for security benchmarking; (d) definition of a standard approach for security
evaluation and comparison; (e) implementation and evaluation; (f) writing papers;
PhD Thesis Proposal: G7.10
Title: “Dependability Benchmarking for Distributed and Parallel Database
Environments”
Keywords: Databases, distributed systems, parallel systems, dependability
benchmarking, fault-tolerance
Supervisors: Prof. Marco Vieira (mvieira@dei.uc.pt)
Prof. Henrique Madeira (henrique@dei.uc.pt)
Summary:
The ascendance of networked information in our economy and daily lives has increased
the awareness of the importance of dependability features. In many cases, such as in ecommerce
systems, service outages may result in a huge loss of money or in an
unaffordable loss of prestige for companies. In fact, due to the impressive growth of the
Internet, some minutes of downtime in a server somewhere may be directly exposed as
loss of service to thousands of users around the world.
Database systems constitute the kernel of the information systems used today to
support the daily operations of most organizations. Additionally, in recent years, there
has been an explosive growth in the use of databases for decision support decision
support systems. The biggest differences between decision support systems and
operational systems, besides their different goal, are the type of operations executed
and the supporting database platform. While operational systems execute thousands or
even millions of small transactions per day, decision support systems only execute a
small number of queries on the data (in addition to the loading operations executed
offline).
Advanced database technology, such as parallel and distributed databases, is a way to
achieve high performance and availability in both operational and decision support
systems. However, although distributed and parallel database systems are increasingly
being used in complex business-critical systems, no practical way has been proposed so
far to characterize the impact of faults in such environments or to compare alternative
solutions concerning dependability features. The fact that many businesses require very
high dependability for their database servers shows that a practical tool that allows the
comparison of alternative solutions in terms of dependability is of utmost importance.
In spite of the pertinence of having dependability benchmarks for distributed and
parallel database systems, the reality is that no dependability benchmark has been
proposed so far. A dependability benchmark is a specification of a standard procedure
to assess dependability related measures of a computer system or computer
component. The awareness of the importance of dependability benchmarks has
increased in the recent years and dependability benchmarking is currently the subject
of strong research. In a previous work the first know dependability benchmark for
transactional systems has been proposed. However, this benchmark focuses singleserver
transactional databases. The goal of this work proposal is to study the problem of
dependability benchmarking in distributed and parallel databases. One of the key
aspects to be addressed is to figure out how to apply a faultload (set of faults and
stressful conditions that emulate real faults experienced by systems in the field) in a
distributed/parallel environment. Several types of faults will be considered, namely:
operator faults, software faults, and hardware faults (including network faults).
Proposal
The PhD work will comprehend the following initial tasks: (a) overview of the state-ofthe-
art about parallel and distributed databases, dependability assessment and
dependability benchmarking; (b) definition of a dependability benchmarking approach
for distributed and parallel databases; (c) study of typical faults in distributed and
parallel databases environments; (d) definition of a standard approach for dependability
evaluation and comparison in distributed and parallel databases; (e)implementation
and evaluation; (f) writing papers;
PhD Thesis Proposal: G7.11
Title: “Techniques to Improve Performance in Affordable Data Warehouses”
Keywords: Data warehousing, on-line analytical processing (OLAP), performance,
parallel and distributed databases
Supervisors: Prof. Jorge Bernardino (jorge@isec.pt)
Prof. Henrique Madeira (henrique@dei.uc.pt)
Summary:
In recent years, there has been an explosive growth in the use of databases for decision
support. These systems, generically called Data Warehouses, involve manipulations of
massive amounts of data that push database management technology to the limit,
especially in what concerns to performance and scalability. In fact, typical data
warehouse utilization has an interactive characteristic, which assumes short query
response time. Therefore, the huge data volumes stored in a typical data warehouse and
the queries complexity with their intrinsic ad-hoc nature make the performance of
query execution the central problem of large data warehouses. The main goal of this
work is to investigate ways to allow a dramatic reduction of the hardware, software, and
administration cost when compared to traditional data warehouses. The affordable data
warehouses solution will be built upon the high scalability and high performance of the
DWS (Data Warehouse Stripping) technology. Starting from the classic method of
uniform partitioning at low level (facts), DWS includes a new technique that distributes
a data warehouse by an arbitrary number of computers. Queries are executed in
parallel by all the computers, guaranteeing a nearly optimal speedup.
This work will focus various aspects related to:
• Automatic data balancing: As each node in the cluster may have different processing
capabilities, it is important to provide load balancing algorithms that automatically
provide the best data distribution. With these mechanisms the system will be able to
reorganize the data whenever it is needed, in order to make the load in each node as
balanced as possible, allowing similar response times for every nodes.
• Auto administration and tuning: by using a cluster of machines the administration
complexity and costs tend to increase dramatically. Although the several nodes have
normally similar configurations, some discrepancies are expected due to the
heterogeneous nature of the cluster. To achieve the best configuration we need to tune
each node inpidually. Thus, we have to develop a solution for automatic
administration and tuning in distributed data warehouses that allows a reduction of the
administration cost and an efficient use of the system resources.
Proposal
The PhD work will comprehend the following initial tasks: (a) overview of the state-ofthe-
art; (b) characterization of affordable data warehouses requirements; (c)
infrastructure for improved performance in affordable data warehouses; (d)
implementation and evaluation; (e) writing papers;
PhD Thesis Proposal: G7.12
Title: “Towards High Dependability in Affordable Data Warehouses”
Keywords: Data warehousing, on-line analytical processing (OLAP), fault-tolerance,
data security, parallel and distributed databases
Supervisors: Prof. Marco Vieira (mvieira@dei.uc.pt)
Prof. Henrique Madeira (henrique@dei.uc.pt)
Summary:
In recent years, there has been an explosive growth in the use of databases for decision
support. These systems, generically called Data Warehouses, involve manipulations of
massive amounts of data that push database management technology to the limit,
especially in what concerns to performance and dependability. The huge data volumes
stored in a typical data warehouse make performance and availability two centrals
problem of large data warehouses. An affordable data warehouses solution is being built
upon the high scalability and high performance of the DWS (Data Warehouse Stripping)
technology. Starting from the classic method of uniform partitioning at low level (facts),
DWS includes a new technique that distributes a data warehouse by an arbitrary
number of computers. The fact that the data warehouse is distributed over a large
number of computers raises new challenges as the probability of failure of one or more
computers greatly increases. The main goal of this work is to investigate ways to
achieve high-dependability in the affordable data warehouses solution while allowing a
dramatic reduction of the hardware, software, and administration cost when compared
to traditional data warehouses. Thus, this work will focus various aspects related to:
– Data security: the affordable data warehouses solution will be based on open-source
database management systems (DBMS). However, these databases do not provide the
security mechanisms normally available in commercial DBMS. In addition, the
distributed database approach increases the data security requirements. The goal is to
investigate the security needs for distributed data warehouses over open source DBMS
and to propose advanced mechanisms that improve the overall system security.
– Data replication and recovery: comparing to a single-server database, one of the
consequences of the use of a cluster of affordable machines is the increase of the
probability of failure. This way, one of the goals is to research a new technique for data
replication and recovery that allows the system to continue working in the presence of
failures in several nodes and facilities the recovery of failed nodes.
Proposal
The PhD work will comprehend the following initial tasks: (a) overview of the state-ofthe-
art; (b) characterization of affordable data warehouses dependability requirements;
(c) infrastructure to support high-dependability in distributed data warehouses; (d)
Implementation and evaluation; (e) writing of papers;
G8: Information Systems
http://isg.dei.uc.pt/
PhD Thesis Proposal: G8.1
Title: “Learning Contexts and Social Networking”
Keywords: e-learning 2.0, e-learning, learning contexts, learning, LMS, social
networking, technology enhanced learning (TEL), web 2.0
Supervisor: Prof. Antonio Dias de Figueiredo (adf@dei.uc.pt)
Summary:
This thesis concentrates on the design, implementation and management of learning
contexts. The major tenet of our Learning Contexts Project, which has been gaining
strength in the course of over thirty years devoted to ICT and Education, is that the
future of learning is not to be found just on content, but also, and very much, on
context, that is, on making learning happen within activity rich, interaction rich, and
culturally rich social environments that never existed, that the intelligent use of
technology is making possible, and where different paradigms apply. Many of the most
dynamic fields of research in learning and education, such as computer supported
cooperative learning, situated learning, or learning communities relate to learning
contexts. Hundreds of expressions used in education – such as project based learning,
action learning, learning by doing, case studies, scenario building, simulations, role
playing – pertain to learning contexts. The advantage of concentrating on context, as a
whole, rather than on the multiplicity of its manifestations studied by disparate
research groups is that, by doing so, we can articulate that multitude of theories and
practices into a single, coherent, organic, and operational worldview. The proposed
thesis pushes forward our current efforts in this field by exploring the relationship
between Learning Contexts and Social Networking. This may include collaboration with
another of our projects, the Serendipty Project, centred on the development of a
serendipitous social search engine for which we hold a US patent application. In order
to stimulate the creativity of the candidates, plenty of leeway will be given to them, so
that they may choose to concentrate on theoretical aspects, on practical educational
issues, or on the specification and design of the ideal, and yet inexistent, learning
context management system (LXMS).
Prospective candidates whishing to clarify the research implications of learning contexts
may download from the journal Interactive Educational Multimedia our paper “Learning
Contexts: a Blueprint for Research”. Further information can be obtained by
downloading Chapter 1, “Context and Learning: a Philosophical Framework”, of our
book Managing Learning in Virtual Settings: the Role of Context, published by
Information Science Publishing (Idea Group). Successful candidates will have, or be
willing to develop throughout their PhD, a mixed profile of educational technology and
educational and social researcher.
- Figueiredo, A. D. (2005) “Learning Contexts: A Blueprint for Research”, Interactive Educational
Multimedia, No. 11, October 2005
http://www.ub.es/multimedia/iem/down/c11/Learning_Contexts.pdf
- Figueiredo, A. D. and Afonso, A. P. (2005) “Context and Learning: a Philosophical Framework”,
in Figueiredo, A. D. and A. P. Afonso, Managing Learning in Virtual Settings: The Role of Context,
Information Science Publishing (Idea Group), October 2005.
http://www.idea-group.com/downloads/excerpts/Figueiredo01.pdf
PhD Thesis Proposal: G8.2
Title: “Quality management and information systems: getting more than the sum of
the parts”
Keywords: Quality Management, Information Systems
Supervisor: Prof. Paulo Rupino (rupino@dei.uc.pt)
Summary:
Quality Management of products, services and business processes is, today, a key issue
for the success of most companies operating in global contexts. In fact, holding a
Quality certification, such as established by ISO 9001:2000 standards, is becoming a
basic requirement for companies to play in several international markets. On the other
hand, the design and deployment of information systems is another key aspect to
consider when modern organizations define their business models and strategy.
It is quite surprising, thus, that in spite of the fact that both quality management and
information systems architecting require intensive strategic analysis and the extensive
involvement of staff and managers in the examination and redesign of business
processes, the two endeavors are still treated as completely distinct. They are usually
conducted as separate projects, handled by different teams, equipped with unconnected
methodologies.
The integrated design of these two pillars of modern organizations, in a manner that
they depend on, support, and reinforce each other, enables a quantum leap, as it lets
organizational tasks be reengineered in the light of: (i) effectiveness, consistency and
evidence of compliance, as required by quality systems; and (ii) efficiency, harnessing
the power of digital information storage, processing, and communication in the renewed
business processes. Typical criticisms to traditional implementations of Quality
Management Systems can also be alleviated, namely by reducing the added
bureaucracy and overhead imposed on users by traditional implementations. The
economic impact on organizations can be considerable, not only at the initial planning
stage but, more importantly, throughout the lifecycle of operation of this unified system.
The likelihood of synergy between quality management and IT infrastructure has been
suggested by a few authors, but no systematic processes for leveraging those synergies
can be found.
A successful Ph.D. in this unexplored field will arm its holder with the skills and tools
to act in an increasingly appealing consulting arena.
PhD Thesis Proposal: G8.3
Title: “Visualizing and Manipulating Work Load Control over Business Networks”
Keywords: Information Systems, Work Load Control Methodology, Human-Computer
Interaction, Information Visualization, Direct Manipulation Interfaces, Delegation,
Interface Agents
Supervisor: Prof. Licínio Roque (lir@dei.uc.pt)
Summary:
In a previous project we designed a web-based planning and control system for Small
and Medium Enterprises that operated on Make-To-Order clusters (a case in the Mouldmaking
Industry) that implemented and adaptation of the Lancaster proposed Work
Load Control planning methodology. The Work Load Control methodology enables
production management across shopfloors by controlling pending work load levels
across workcentres, thus effectively managing overlapping “windows of opportunity” for
completing every task at each production unit. We have developed a system where the
user sets the planning conditions and delegates in the system the generation of plan
proposals. Unmet restrictions are then iteratively resolved by taking management
decisions that adapt a candidate plan to actual production conditions and vice-versa.
Some of the conclusions of the project relative to the adoption of the new planning tool
were: a) the Work Load Control methodology while particularly flexible and adaptable
for SME running MTO operating models poses a learning obstacle for people trained to
think in time-sliced models (like those depicted in Gantt diagrams); b) a web-based
system while easy to deploy and manage poses a heavy cognitive load as the principal
interaction mode is linguistic (using menus, dialogs, forms, with some graphics for
visualizing resulting plans); c) current business globalization increasingly involves high
levels of subcontract work that needs to be managed across enterprise networks with
only partial knowledge of production conditions, which makes it difficult to use
methodologies that assume full knowledge and control over production units.
With clients, we have come to the conclusion that a planning tool to the Work Load
Control methodology needs a visualization and direct manipulation tool to reduce the
cognitive overhead posed by the complexity and “non-intuitiveness” of the methodology
and enable the person to dynamically envision and track events across networks of
enterprises. This case provides an ideal opportunity to attempt an integration of
linguistic, direct manipulation and delegation modes of interface, develop novel
visualizations and test usability evaluation techniques. The research implies acquiring a
knowledge of the methodology, conceiving and studying appropriate solutions for the
case study by designing innovative interaction techniques. Relevance is met in Decision
Support Systems, Human Computer Interaction and, more generally, in the Information
Systems academic and business communities.
PhD Thesis Proposal: G8.4
Title: “Designing Games as Learning Contexts”
Keywords: Information Systems, Human-Computer Interaction, Context
Engineering, Learning Games, Learning Contexts, Social Constructivism
Supervisor: Prof. Licínio Roque (lir@dei.uc.pt)
Summary:
Several authors have argued the idea that videogames could be exploited as learning
environments. James Paul Gee has written extensibly on the subject of learning from
computer games, noticing that we can hardly ignore the learning that takes place with
this new medium. Mark Prensky argued this idea on the simulation aspects of games.
Raph Koster, a designer and consultant, adopts the perspective that games are actually
meant to be learned, and makes playing as learning the basis for game design. Simon
Engenfeldt-Nielsen produced a PhD thesis on the educational potential of computer
games, by analyzing cases that used commercially available games.
Seeing games in the light of socio-technical theories such as Actor-Network we have
come to an interpretation of games as constructions designed and built to enforce
specific “programs of action” while its storytelling and underlying rules provide the
basis for stable translation regimes by the player. Games as simulations can also be
understood as embodied theories of physical or social reality. By providing specific
embodied concepts and representations, physical or other abstract rules, characters
with recognizable behavior, the designer builds a learning context with conditions of
engagement that concur to enable playful experiences. These can be more or less
flexible or open to interpretation through player choices, but always enforcing
underlying worldviews or inscribed theories that are meant to be learned in the course
of playing the game if the player is to achieve the game’s goal. It is this activity
conditioning that we suppose to be the usable basis for explicitly conceiving games as
learning devices.
An alternative approach would be to take game design itself as the learning activity and
explore the learning potential inherent in the design activity. Either way, the research
should focus on the methodological problem of explicitly modeling and building games
as learning contexts. The Context Engineering approach can be used to frame
development of specific contexts, by prototyping on available multiplayer game
development technology and focusing on design aspects and their relation to the
proposed problem. Adequate evaluation techniques should also be a consideration in
the studied contexts if they are to be socially accepted as effective learning alternatives.
Relevance is expected for the Learning Sciences, Human and Social Sciences,
Information Systems and Human Computer Interaction, Game Studies, Media Studies,
and society at large.
PhD Thesis Proposal: G8.5
Title: “Visual Modeling Language for Game Design”
Keywords: Information Systems, Human-Computer Interaction, Visual Modeling
Language, Games and Design
Supervisor: Prof. Licínio Roque (lir@dei.uc.pt)
Summary:
Game design is a complex activity that deals with multiple and heterogeneous
concurrent constraints. A game designer has to consider combined effects of multiple
elements effectively requiring a trans-disciplinary background that can range, with
variable intensity, from Humanities to Media Studies, to Psychology and Sociology, to
Aesthetics, to Informatics and Economics. While the advent of a common design
language still seems far in the future, some design patterns can be readily recognized
from the 30+ year history of videogame development. An example of a systematization is
Björk and Holopainen’s “Patterns in Game Design”. Nonetheless, the idea of a visual
language for game modeling and design seems not only possibly and relevant, but a
pressing research goal.
In pursuing the goal of a visual modeling language for game design it is expected that a
knowledge of socio-technical studies of science and technology will give useful insights
into the problems of trans-disciplinary and in particular the use of Actor-Network
Theory constructs as a basis language for the analysis of game contexts. A requisite for
such a language would be that we could build a prototype modeling tool that could be
used to sketch and generate game coding to be used on a current software game
platform. Another basic requisite would be that it could serve as a basis for design
dialogue between people with perse disciplinary backgrounds.
This research project would involve the trans-disciplinary background study and
drafting of o a language prototype that could then be evaluated and evolved by
successive design iterations, against updated usability requirements. Relevance and
adequacy would be judged based on actual empirical design experience and historical
design accounts, in the fields of Information Systems, Human Computer Interaction,
Game Development, Design, Games Studies and Media Studies.
PhD Thesis Proposal: G8.6
Title: “Programming Games by Demonstration (and Learning to Program)”
Keywords: Information Systems, Human-Computer Interaction, Games and Design,
Programming by Demonstration, Programming by Example
Supervisor: Prof. Licínio Roque (lir@dei.uc.pt)
Summary:
Game programming and development, from scratch, requires advanced skills and
specialties often unavailable to a game designer or an artist, less alone to the proverbial
man-on-the-street. The advent of general purpose game engines and game generation
environments made it simpler and more affordable, or at least a less specialized task, to
be able to develop and deploy complex game scenarios. Yet, the simpler game design
attempt still requires if not the domain of complex programming at least some skills
with a scripting language.
Towards democratizing this technology and the videogame medium, and taking the
historical lesson from what happened with the appearance of personal filming cameras
and the development of the cinema, again with video and the TV, it seems interesting to
work on a solution that would enable a wider public to become an active participant in
the creation of interactive content. Resorting to and evolving programming by
demonstration or programming by example techniques could play a significant part in
lowering the learning barrier to achieve useful effects analogous to behavior scripting.
Building a system that would enable reflexive action by letting the user inspect and
animate the programming results of demonstrative actions, could serve as a basis for
semi-autonomous learning of programming concepts and skills.
It is intend that the candidate researcher pursues this goal by designing and building a
prototype system on top of an existing software game platform or virtual environment
and proceed to evaluate the generated concept through actual empirical cases with
targeted user segments. Relevance and adequacy would be judged based on actual
empirical design experience, with results published in the fields of Information Systems,
Human Computer Interaction, Game Development, Design, Games Studies and Media
Studies.
G9: Evolutionary and Complex Systems
http://cisuc.dei.uc.pt/ecos/
PhD Thesis Proposal: G9.1
Title: “Evolving Representations for Evolutionary Algorithms”
Keywords: Evolutionary Computation, Gene Regulatory Networks, Self-Organization,
Supervisor: Prof. Ernesto Costa (ernesto@dei.uc.pt)
Summary:
Evolutionary Algorithms typically approach the genotype - phenotype relationship in a
simple way. As a matter of fact, conventional EAs consider the genotype as the complex
structure and rely on more or less simple mechanisms to do the mapping from genotype
to the phenotype. Some work is being done on the relation between the user-designed
representations used by the EA (the genotype) and the fitness landscape induced by the
problem (the phenotype). The idea is to understand the role played by representations
for improving evolvability. This is important but we can advance a step further.
Exploring ideas from developmental biology in the context of evolutionary algorithms is
not new. Notwithstanding, the challenge here is to understand better how we can
combine the theory of evolution with embryonic development in an unified framework
and explore it computationally aiming at evolving the representations to be explored by
an EA instead of design them offline, attaining a self-organized evolutionary algorithm
(SOEA). To achieve that goal we have to identify the building blocks for representations
as well as the transformational rules that end up in the definition of an adapted
inpidual.
PhD Thesis Proposal: G9.2
Title: “Harnessing Dynamic Environments: the problem of prediction”
Keywords: Evolutionary Computation, Dynamic Environments, Prediction
Supervisors: Prof. Ernesto Costa (ernesto@dei.uc.pt), Prof. Juergen Branke
Summary:
Most real-world optimization problems dynamically change over time (e.g. scheduling,
routing, transportation or robot’s navigation problems). In such cases, the task of an
optimization algorithm changes from finding an optimal solution to continuously adapt
an existing solution to the changing environment. Nature-inspired optimization
algorithms have proven to be successful candidates for such problems.
When changes occur repeatedly it should be possible to learn from the experience and
predict future changes of the environment. Such predictions, even if uncertain, could
help the algorithm to make decisions that prepare it for what is to come, allowing an
even faster adaptation and avoid getting stuck in "dead ends". In the context of dynamic
environments all work has focused on enabling the algorithm to adapt quickly after a
change. Only very few papers have attempted to anticipate changes. So far, there is no
fundamental and general investigation on the importance and possibilities to integrate
prediction into nature-inspired optimization. This thesis would aim at making some
fundamental investigations into the role and potential of prediction in dynamic
environments, and at developing some new and effective ways to integrate prediction
into various nature-inspired optimization heuristics.
PhD Thesis Proposal: G9.3
Title: “Evolutionary Hybridization with State-of-the-art Exact Methods ”
Keywords: Evolutionary computation, optimization, hybridization.
Supervisor: Prof. Francisco Pereira (xico@dei.uc.pt)
Summary:
Standard Evolutionary Algorithms (EA) often perform poorly when searching for good
solutions for complex optimization problems and may benefit if they are combined with
other techniques. Broadly speaking we can consider two large classes of hybrid
architectures: the EA can be complemented with other search methods or it can be
enhanced with problem specific heuristics that add explicit knowledge about the
problem being solved. A drawback associated with research conducted on this topic is
that many reported approaches are typically somewhat naive in nature and with a
limited applicability. Moreover EAs are, in most situations, combined with basic
standard procedures such as hill-climbing algorithms or simulated annealing.
This project aims at conducting an inclusive study of evolutionary hybridization to
analyze if it is possible to develop new architectures that perform better than today’s
methods. Special attention will be given to hybridization with exact algorithms, like
linear programming or gradient-based search. The challenge is to understand how the
key properties of these exact techniques, such as the capability to reduce the search
space or the effective exploration of neighborhoods, might be used by the EA to
efficiently perform a global exploration of the search space. Several examples of
optimization problems will be used to perform a comprehensive analysis of the
developed hybrid architectures.