King's College London
M.Sc. in Information Processing and Neural Networks |  |
M.Sc. projects for Summer 2007
Below are the projects as allocated. Everyone got their first
choice. Please note that each candidate will have to pass at least
four of the IPNN examinations, including any course listed
specifically as a prerequisite for the project allocated, in order to
be allowed to start the project in the Summer.
For projects that are not supervised internally, i.e. by a member of
staff of the Mathematics Department, a second ``internal'' supervisor
has been allocated in addition to the first, ``external'' one. The
role of the second supervisor is to provide guidance on what is
required for an M.Sc. IPNN project and for the project report, and to
ensure that the project makes full use of the theoretical material in
the taught components of the M.Sc.
project 1: Dr R Kuehn (reimer.kuehn@kcl.ac.uk)
Online learning in a multilayer architecture with
fixed pre-processing in the first layer
student: GUY LEVER
project 2: Dr R Hepburn (robert.hepburn@erg.kcl.ac.uk), Dr R Kuehn
(reimer.kuehn@kcl.ac.uk)
Identifying invalid measurements in air quality data
student: RICHARD PETER
project 3: Dr I Perez-Castillo
(isaac.perez_castillo@kcl.ac.uk)
Message passing algorithms in densely connected neural networks
NOT TAKEN
project 4: Dr L Carlin (leo.carlin@kcl.ac.uk), Prof ACC Coolen
(ton.coolen@kcl.ac.uk)
Bayesian information processing in optical proteomics
student: MARK ROWLEY (part-time)
project 5: Prof ACC Coolen (ton.coolen@kcl.ac.uk)
Information geometry with inner products
student: CASPER CLEMENCE
project 6: Dr R Heintzmann
(Rainer.Heintzmann@kcl.ac.uk),
Prof ACC Coolen (ton.coolen@kcl.ac.uk)
Bayesian reconstruction of images upon structured illumination
student: SEYED AZARBADEGAN
project 7: Dr S Alty (steve.alty@kcl.ac.uk), Dr R Kuehn
(reimer.kuehn@kcl.ac.uk)
Support Vector Machines for Cardiovascular Disease Prediction
NOT TAKEN
- You should meet as soon as possible with your supervisor
to discuss details of the project and start preparing for the project
outline; for projects with external and internal supervisors a joint
meeting with both should be held. It is up to you, the student,
to organize this meeting; contact emails for all supervisors are given above.
- Points to be discussed at this preparation meeting include:
- Selection of initial background literature for the project (and
in particular for the project outline). Discuss means of obtaining
further literature (see also the research methods seminars).
- If specific data are required, are they available already? If
not, how can they be obtained? In what format are the data, and how
are they to be converted if necessary?
- If specific software is required, discuss access to this (e.g. from which computers?) as well as any necessary training and/or
background information; what skills will you need to learn or brush up
on before the project starts?
- Agree on how your supervisor(s) can be contacted if you have
further queries about the project up until the preparation of the
project outline.
- Agree on the format and frequency of meetings with the
supervisor(s) once the project has started; a typical suggestion would
be roughly once a week. Establish whether supervisor(s) will be away
for any extended periods between June and September and plan for this.
If it can sensibly be done at this stage, draw up a rough schedule/
workplan for the project.
- Discuss also whether your first supervisor will be happy to give
feedback on an early draft of the project report, and how long before
the submission deadline (s)he would need the report to do this.
- Project outlines need to be submitted to Ton Coolen by
Friday 30 March 2007. Please give 3 hardcopies to Rebecca Cullen
in 432 or otherwise email me a PDF, postscript (or, if absolutely
necessary, Word) file. We will provide feedback on outlines by email,
provisionally by April 6th 2007. If your outline is
not satisfactory, we will ask you to revise it as necessary. Outlines
should be no more than 2 pages in length and should contain:
- Project title, your name and your supervisor's name
- A clear statement of the problem or question which the project
aims to tackle - this can be based on the project proposal produced
by the supervisor but should be in your own words (1 or 2 paragraphs).
- A brief summary of relevant existing work (1 or 2 paragraphs),
with references to a few literature sources (to be listed at the end).
- A brief but specific description of the methods to be used. For
theoretical projects, discuss e.g. what analytical techniques will be
used, how the predictions will be evaluated numerically, how
simulations - if any - will be performed (what programming language,
is there existing code), whether data for comparisons already
exist. For more applied projects, discuss e.g. what data sets are to
be analysed, whether further data will need to be obtained and how,
what learning algorithms are to be used, how they are to be
implemented (what software, will new codes need to be written), what
parameters will be varied, how performance will be assessed.
- If you haven't already done so, refresh your mind on the points
covered in the research methods seminars, e.g. regarding research
tools and scientific word processing.
- Detailed guidelines on how project reports will be marked (see
here), as well as the actual submission deadline
(likely to be around 10 Sept 2007) will be circulated in
June/July. Arrangements for the oral presentations that take place
after submission will be finalized during the summer. It is expected
that oral presentations will be held during the week 17-21 Sept
2007. Please mark these provisional dates in your diaries now. Since the presentation is a mandatory requirement for award of the
M.Sc., it is important that you remain in London until you have given
your presentation; if you are planning to leave the country, make sure
to plan your travel dates accordingly.
