General description, Bengt Aspvall

piloturuguayanAI and Robotics

Oct 15, 2013 (3 years and 7 months ago)

71 views

C
omputer
S
cience

Teachers available for thesis supervising and their field of interest

Autumn 2011

Supervisor

Slots

Keywords

Bengt
Aspvall

2

Analysis of algorithms, Data structures, Complexity theory, Performance,
Discrete mathematics

Stefan
Axelsson

0

Security, Anomaly detection, Intrusion detection, Information
visualization (as applied to these areas), Network security, Digital
forensics, unix/linux

Hussein
Aziz

0

Video streaming, Adaptive the video streaming, video conferencing, QoE
for video commun
ication

Martin Boldt

2

Computer security, Information security, Malicious software, Software
security, Reputation/recommendation systems, Mac
hine learning, Social
networks, Social engineering

Bengt
Carlsson

1

Social computing, Security, Privacy tasks, Information ecosystems,
Cognitive behaviors, Social security, Social nets, Malware, AI
-
techniques

Jeanette
Eriksson

0

1a: End
-
user tailorable

software, User centric design, Cooperative design,
End
-
user development, usability.

1b: Tailorability in games, Tailorable avatars, Tailorability in game
-
based
learning

Håkan
Grahn

0

Multicore systems, Parallel programming, Computer Architecture,
Performance Evaluation, Operating System

Bo Helgeson

2

Interaction design, Computer supported cooperative work, Ethnography,
HCI, Mobile computing, User experience, User centered design,
Telemedicine and eHealth

Johan
Holmgren

0

Agent Technology, Optimization, Simulation, Agent
-
based simulation,
Machine learning, Data mining

Stefan
Johansson

0

Artificial potential fields, Multi
-
agent systems, Game AI, Board games

Hans
Kyhlbäck

1

Interaction as situated action in the world, Human computer interaction,
Computer supported collaborative work, Participatory design,
Developmental work research, Activity theory, Object
-
oriented
system
development

Niklas
Lavesson

2

Machine learning, Computational learning theory, Artificial intelligence,
Data mining, Knowledge disc
overy, Pattern recognition, Natural
language processing, Supervised learning, Clustering, Classification,
Parameter tuning, Complexity analysis, Evaluation methods, Evaluation
measures

Craig
Lindley

2

Robotics; Cognitive architectures

New paradigms for synthetic Intelligence; Convergence of computing
science, Molecular science, Biotechnology and robotics

Autonomy; Self
-
assembling,
-
organising
and
-
replicating systems

Digital fabrication of physical systems; Human
-
robot interaction

Lars
Lundberg

2

Multiprocessor scheduling,
Real
-
time scheduling, High performance
scheduling, Performance issues, Cloud computing

Andrew
Moss

1

Program analysis, Compilers, Parsing,
Language design, Domain specific
languages, Cryptography, Performance measurement, Parallel processing,
Algorithms, GPU

Marie
Persson

2

Heal
thcare management, Operating room scheduling, Waiting list
management, Simulation, Optimization, Agent technology, Machine
learning

Kari Rönkk
ö

0

Child friendly computing, User experience, Wicked design problems,
Usability, Work practice studies, Human aspects and communication,
Flexible research studies, Techniques, Methods and Processes

Charlotte
Sennersten

0

Game
-

and simulation development, Cognitive science, Human Factors,
Playability, Game design, Gameplay, Eyetracki
ng and physiological in
-

and output, the Arts, Sensor interaction, 3D scanning and printing

Veronica
Sundstedt

0

Computer graphics (Mod
eling, Animation, Rendering), Novel gaze
interaction (Games and interfaces), Experimental psychology (User
studies), Perceptual models, Visualisation, Image quality metrics, Virtual
environments

Hans Tap

2

IT and work practice, Interaction design, Interaction aspects within: HCI,
UbiComp, CSCW etc, Interaction analysis, Innovative interfaces
,
Appliance design, Serious gaming

Johanna
Törnquist

0

Transport and traffic scheduling, Railway traffic, Optimization,
Mathematical
modeling, Heuristics, Simulation, Decision support
systems, Data sensitivity analysis


General description, Bengt Aspvall

Algorithms and data structures:

In
Analysis of Algorithms

one studies fundamental aspects of algorithms as well as the
interplay between data structures and algorithms. Decisions about structuring data cannot be
made without knowledge of the algorithm, and vice versa, the choice of algorithms depends
on the stru
cture of the underlaying data. New and faster hardware will help to speed up
computations. Finding better combinations of algorithms and data structures might result in
even larger performance increase. Parallel and distributed computing adds power
---
and
additional challenges.

Theoretical research in the field is today quite specialized and often requires advanced
mathematics. Application areas are rich sources of algorithmic challenges that can lead to
practical (and possibly theoretical) contributions. A

good starting point for thesis work on
algorithms and data structures is a well
-
selected and interesting application area.

General information, Stefan Axelsson

My main interest is the application of information visualization to the problem of anomaly
dete
ction (esp. machine learning/data mining algorithms) so that the operator can gain an
operational understanding of what the anomaly detector is saying and why. This technique is
of course more generally applicable.

Other interests are digital forensics, es
p. file fragment analysis, memory fragment analysis
etc. possibly applying the same methods as above. I also have a background (from industry
and personal interest) in networking and the unix/linux plattform.

Contact information:
stefan.axelsson@bth.se

General description, Martin Boldt

My main research interest involves the study of Privacy
-
Invasive Software (PIS), loosely
labelled spyware. The main research goal in this setting is to create a classification of
PIS, and
thereby investigate how PIS relate to both legitimate and traditionally malicious software. I
focus on the process of informing users about software implications as they are about to
install new software on their system. One method we are working
on is automatic analysis of
the End
-
User License Agreements (EULA) as well as the use of a collaborative reputation
systems. Apart from this I'm also interested in topics within the additional areas specified by
the keywords.

Contact information:
martin.boldt@bth.se

General description, Bengt Carlsson

Most security and privacy related problems are connected to net activities. When we do a
Google search, stay in contact with friends through a Facebook connection or
store our files
on the web we may put our privacy at risk or be a target for malicious activities. Investigating
the computer cloud and especially social computing, e.g. social security and social nets are an
increasing area of interest. Models of informat
ion ecosystems, cognitive skills including AI
-
applications and management issues together form a complex virtual environment where
security and privacy play important roles. By combining traditional malware behavior
/prevention, analysis of software in com
mercial use and AI
-
techniques with activities within
the computer cloud we may describe and find solution to the state
-
of
-
the art social computing
problems.

Contact information:
bengt.carlsson@bth.se

General des
cription, Jeanette Eriksson

1.In most business today, competition is hard and it is a matter of company survival to
interpret and follow up changes within the business market. The margin between success and
failure is small. Possessing suitable, sustainabl
e information systems is an advantage when
attempting to stay in the front line of the business area. In order to be and remain competitive,
these information systems must be up
-
to
-
date, and adapt to changes in the business
environment, which is a challeng
e. One way to approach this challenge is trough flexibility in
systems. The power of flexibility is that it keeps the system usable and relevant and allows it
to evolve. End
-
user tailorability and end user development is a kind of flexibility where it is
p
ossible for the end
-
user to evolve an application better to fit altered business requirements
and tasks. Tailorability automatically introduce limitations of what the end
-
user can do, and to
overcome the limitations cooperative design processes are needed,

where the end
-
user and the
developers work together. Another perspective on tailorable software is personalization.

Personalization occurs in most commercial software, and is also an important part of
computer games. For example game
-
based learning is an
area where tailorability is an
essential factor. The relationship (in theory and practice) between tailorable business systems,
commercial software, and the tailorability in computer games is an interesting area to
explore.Contact information:
jeanette.eriksson@bth.se

General description, David Erman

Large
-
scale, real
-
time multimedia distribution over the Internet has been the subject of
research for a substantial amount of time. A large number of mechanisms,
policies, methods,
and schemes have been proposed for media coding, scheduling, and distribution. IP multicast
was expected to be the primary transport mechanism for this, though it was never deployed to
the expected extent. Recent developments in overlay
networks have reactualised the research
on multicast, with the consequence that many of the previous mechanisms and schemes are
being re
-
evaluated. BTH have extended the popular BitTorrent protocol to become suitable as
a solution for streaming video, and
implemented this in both a simulator as well as in a actual
BitTorrent client. However, much work remains to be done to provide real
-
time streaming
support, more efficient peer selection schemes, as well as a more comprehensive incentive
mechanism.

IMS is
a very important part of the next generation telephony and communication services,
VoIP as well as land
-
line and mobile. However, earlier tests have shown that some of the
implementations of IMS available on the market today suffer from prohibitively large

delays.
It is presumed that this is due to both the execution environment of the IMS installation, as
well as the protocols used in their interaction, primarily SIP. More work is needed to
investigate these issues, and suggest improvements.

Network codin
g is a promising technique for, among other things, decreasing network load in
multi
-
path wireless networks and improving effciciency in multicasting. NC extends the
capabilities of a typical router (forwarding on a single outgoing link or replicate to sev
eral
outgoing links) with processing capabilities. This assumption is also one made in the
implementation of P2P overlay systems such as Gnutella, Skype, BitTorret etc. However,
little work has been done on looking into the implications and possible uses o
f NC in a P2P
context.

Contact information:
david.erman@bth.se

General description, Håkan Grahn

Today we are in the middle of a paradigm shift in the computer industry. Current and future
processor generations are
based on multicore architectures, and multicore processors are the
main computing platform in all types of system from small
-
scale embedded systems to large
-
scale server systems. The hardware performance increase will mainly come from an
increasing number
of cores on each chip, and hardware manufacturers predict that the number
of cores will double every second year. In order to harvest the performance potential of future
multicore processors, the software also need to be parallel and scalable. However, wri
ting
parallel programs is not trivial. Therefore, a number of areas needs to be addressed in order
achieve both correct and high
-
performance parallel software. Suitable topics for master thesis
projects in this area are, e.g., correctness and testing issue
s, performance evaluation, general
parallel programming issues and languages, GPU computing, compilation techniques,
computer architecture, and operating system support.

Contact info:
hakan.grahn@bth.se

General des
cription, Bo Helgeson

My main interest is oriented towards how people actually do things, and how we could design
IT
-
based tools to support people in doing so.

I have been working within the field Computer Supported Cooperative Work (CSCW) for
more than 20

years, and I still think this is an extraordinary exciting area for research.
Interaction Design is another research area, which closely overlaps CSCW. Also here
knowledge about how people actually are carrying out things is crucial for design and
develop
ment. Of course, issues related to usability and interfaces are central for my interests
in the area of people and computers.

Contact information:
bo.helgeson@bth.se

General description, Johan Holmgren

My research
mainly concerns the integration of agent technology, simulation and
mathematical optimization with applications to decision support and policy analysis in freight
transportation. My research has, for instance, resulted in TAPAS (
www.bth.se/tapas
), which is
a simulation tool for analyzing transport
-
related policy and infrastructure measures. My main
application domain is freight transport modeling, but since the emphasize in my research is on
technologies, I am also ab
le to supervise master thesis projects in which the techniques listed
above are applied in application areas other than freight transport modeling.

Contact information:
johan.holmgren@bth.se

General description,

Stefan Johansson

Artificial potential fields (APF) is a technique mainly used in robot navigation systems to
avoid obstacles in dynamic environments. Lately we have shown its applicability also in a
number of other domains such as Real Time Strategy (RTS)

games, and car racing
simulations. We have e.g. a competitive bot that we have developed for the Starcraft bot
competitions and there are good opportunities to make interesting investigations using that
platform and the same applies to the car racing comp
etition.

Another option is to make a project using Multi
-
agent concepts to implement a bot playing
some sort of board game and test its performance against other solutions. So far we have
covered the games Diplomacy, Risk and Settlers of Catan.

In both the
se two project options (APFs and Bord games), we have a strong research
background and publish our results on a regular basis in journals and at conferences. All
projects require quite a bit of programming, so if you are looking for a non
-
programming
proje
ct, kindly move on to the next description. Post
-
project co
-
writing with students on
papers that report on the results of the project is encouraged if the results are interesting
enough.

Contact information:
s
tefan.johansson@bth.se

General description, Hans Kyhlbäck

Studies of interaction as situated action in the world often include elements of both computer
and sork science. Focus on real word activities can have a mainly computer science character,
while hum
an interaction also relate to other research areas. Fieldwork with an ethnographic
approach is a way to gather data about human interaction through th use of computers. A
theoretical frameworksuch as the Activity Theory, support analysis of data generated
with
ethnographic means. Communities of Human Computer Interaction (HCI), Computer
Supported Collaborative Work (CSCW), Participation Design (PD), and Developmental Work
Research (DWR) contribute to researchin different ways.

However, observation of real w
orld activities are fundamental to many of those communities.
Analysis of such fieldwork can provide a good ground for a changed work organization
and/or design of an information and communication technology (ICT) system. With anobject
-
oriented approach to

software development, changed computer tools and changed life
activities can become a shared interest among people with different work roles and various
preferences.

Contact information:
hans.kyhlback@bth.se


Ge
neral description, Niklas Lavesson

Machine Learning is the study of programs that improve their performance through
experience. Since the birth of Artificial Intelligence and Computer Science, Machine Learning
has resided within the core of these subjects.

One of the most heavily studied problems in
machine learning is that of Classification, which can be described as the assignment of
categories to unknown instances of data. The supervised learning paradigm comprises
algorithms that try to generalize from
data with known categories to generate a Classifier that
is able to categorize previously unseen instances of data. Machine Learning techniques are
frequently applied to solve problems in other disciplines such as: natural language processing,
pattern reco
gnition, data mining, and robotics. Machine Learning is an exciting research area
with a large potential for interesting research problems in both theory and application.
Theoretical research could for example involve the study of: computational efficiency
,
learnability, generalizability, model selection, or bias plus variance decompositions. The
potential applications of machine learning are endless but typical examples are: face
recognition, voice recognition, machine translation, spam detection, and pers
onalization.

Contact information,
Niklas.Lavesson@bth.se

General description, Craig Lindley

Research in robotics is advancing rapidly, with parallel technology developments providing
many new opportunities in
advanced robotics systems concepts and new modes of human
-
robot interaction, such as eye gaze tracking and brain
-
machine interfaces. Robotic systems of
interest include robotic systems in manufacturing and field robots, such as air, land, space and
underwa
ter vehicles. At the same time, autonomy of robotics systems proceeds more slowly,
raising the possibility of new paradigms for synthetic intelligence, founded upon non
-
silicon
substrates such as molecular computing or biological neuron cell cultures. New
manufacturing methods also provide possibilities for new robot materials and synthetic
intelligence architectures. Digital Fabrication of Physical Systems is concerned with the
development of technologies for modelling products, systems and artifacts in in
teractive 3D
visualisation and simulation environments, and then using digitally controlled fabrication
technologies to directly produce the modelled items in a physical form. Examples range from
current additive rapid prototyping/manufacturing systems, to

new and emerging digital
biosynthesis technologies.

Contact information:
craig.lindley@bth.se

General description, Lars Lundberg

Multiprocessor scheduling:

Scheduling of processes (or tasks) to processors (or computers) is becoming increasing
important. In the future we will use parallel computer platforms even more than we do today;
examples of such parallel platforms are multi
-
core processors and clusters o
f computers. At
least two important aspects of multiprocessor scheduling can be identified:

-

High performance scheduling. In this case the goal is to minimize the completion time (or
makespan) of one parallel program.

-

Real
-
time scheduling. In this case
the goal is to find a schedule that guarantees that a set of
real
-
time tasks will meet their deadlines.

Cloud computing:

Cloud computing is becoming increasingly popular, and there is a lot of interest for this in
both industry and academia. There are diff
erent kinds of computer clouds, such as
Infrastructure
-
as
-
a
-
Service, Platform
-
as
-
a
-
Service and Software
-
as
-
a
-
Service. Another
dimension is the difference between public clouds, e.g. Amazon and Google versus private
clouds. There are many interesting and ch
allenging aspects of cloud computing, ranging from
technical aspects, such as virtualization, resource allocation, to business aspects, such as
charging models and revenue management.

Contact information:
lars.lu
ndberg@bth.se

General description, Andrew Moss

Program Transformation looks at how we may convert programs from one form to another. In
Program Analysis the target is to transform the program into a behavioural description, such
as how long it takes to run
, or whether or not it executes securely. In comparison Program
Compilation tries to convert a program from one language into another. Both of these areas of
Program Transformation touch on issues in Language Design, in particular Domain Specific
Languages

look at how we may design a language specifically for a particular task in order to
make it easier to compile or analyse programs. Some specific domains that are interesting in
this area are languages for cryptography and parallel processing on GPUs. In t
hese areas there
is a tight connection between how we design algorithms and the languages that we use to
express them. Research projects in this field span questions from the purely theoretical
foundations to practical issues such as how we can accurately
measure and model the
performance of these systems.

Contact information:
andrew.moss@bth.se


General description, Kari Rönkkö

Designing Child Friendly Computer Systems is an area under rapid development to which th
e
research area Human Computer Interaction (HCI) recently added much knowledge. Still there
is room for much more innovative work related to: providing information, education, and
social security; protecting from maltreatment, neglect, and exploitation; pa
rticipation and
having the right to act, being informed, and being involved in decision making
-

master
projects targeting any of these challenges are most welcome.

To successfully design and develop software products and services for targeted users is
be
sides the technical effort also a human effort and challange where the results are heavily
depending on the involved actors interpretations, interactions and means for co
-
constructing
of the future artifact under development. Understanding the influence of

human aspects and
communicaiont can increase the success rate significantly, approaches and theories borrowed
from other research areas can contribute here. Especially how to understand and handle soft
values is a challange, be it related to software deve
lopment methods and processes, or
designing for the user experience
-

master projects that contributes with knowledge here are
most welcome.

Contact information:
kari.ronkko@bth.se

General description, Charlotte Se
nnersten

From a Computer Science (CS) perspective I operate in a subdiscipline

which is called
Human Computer Interaction (HCI), more precisely though a discipline called Game
Development. The focus is the actual interaction between the computer computing and a
person. Game Development is covering technical as well as psychophysiolo
gical challenges
when it comes to
‘ how
-

to
’ design games for players (reward and motivation mechanisms),
how to run them, how to design them both for entertainment as well as for more serious
survival issues, update frequencies, quality prestanda, optimiza
tion etc. In what we
conceptualize being serious games we mean often training for special purposes where the
learning aspect and transfer to real circumstances are crucial, this happens often via
simulation engines and constructed scenarios. To design game
s, simulations and other kinds
of interactions there has to be a close collaboration concerning task issues for individuals or
groups to be able to deliver an useful application at the end. The time and timing concerning
decisions in these scenarios are of

highest importance. Theoretically but also practically most
of the work is also associated to cognitive psychophysiological processes (Cognitive Science
(CogSc), Psychology and Physiology)) , frankly how we as humans can update information
via our senses
in relation to computing but also vice verse and how we may design for support
because of our own shortcomings or if information may has to be allocated to many people at
the same time and how to meet those requirements. Additional interests and work lays
within
embodied cognition, perception with visual attention as specialty, eyetracking as method used
as one physiological in
-

and output (but also heart rate variability, brain activity and
sweating), sensor triggers for music, speech, emotions and movemen
t (dance). My
background is also in the visual (drawing, painting, sculpture)
-

and performing arts (dance
and theater).

Contact information:
charlotte.sennersten@bth.se

General description, Veronica Sunds
tedt

Computer graphics is divided into three main areas: 3D modelling, animation, and rendering.
Interesting projects with modelling include procedural generation of content for applications
such as entertainment, architecture or archeology. This overlaps
with possible projects in
visualisation, and questions in human perception. Current research topics in animation involve
how to model motion to avoid perceptual anomalies such as the 'uncanny valley' effect.
Perceptually realistic rendering requires accura
te simulation of the physical properties of
materials and the behaviour of light. As realistic computer graphics imagery is
computationally expensive interesting topics include local reflection models, global light
transport, and the whole realistic image
synthesis pipeline through to image display.

The human observer does not use all the information computed in the scene. By removing
perceptually unnecessary details we can explore how to reduce overhead in rendering
algorithms. In order to make this impro
vement we must increase our understanding of models
of human visual perception. We validate this approach using psychophysical experiments,
objective difference metrics, and eye tracking technology. Pervasive eye tracking technology
enables novel input mod
alities in many applications. One area of research is the impact on the
next generation of games, measuring its viability in comparison to traditional input devices,
while other areas include accessibility for disabled users, and fusion with other input
mo
dalities such as voice.

Contact information:
veronica.sundstedt@bth.se

and
http://www.bth.se/com/vsu.nsf/


General description, Hans Tap

How people use technolo
gy in their everyday activity is point of departure for my research
interest. A close look and understanding of peoples context is important and demands a
qualitative approach based on different ethographical methods. The interactional aspects of
interest
are mainly concerning the functional aspects of a design. How different functions can
be integrated in every day activities in order to support the activity at hand. This includes
taking cooperation and coordination into account in the analysis and concept

development.
The design space is mainly within areas that look at interfaces beyond traditional desktop set
-
up.

Contact information:
hans.tap@bth.se


General description: Johanna Törnquist

The area of transport and t
raffic scheduling contains a wide selection of interesting and real
life problems which the industry find challenging to cope with. Many of these problems are
resource allocations problems

which are so complex that a decision support often is needed
(or wo
uld be highly beneficial) to assist the planning staff. The problems may be complex in
the sense that there are many alternatives to compute and evaluate, but also because certain
decisions have to be made within a short time (seconds or minutes) and takin
g into account a
lot of information.

Contact information:
johanna.tornquist@bth.se