Machine Learning tea

The first Research Programme in Machine Learning (RPML) tea meeting of 2010 will be held on Thursday March 25 3.30-4.30 pm in Mini-Forum 1 (level 2) in the Informatics Forum.

Tea/coffee and biscuits will be provided, and there will also be a number of posters including:
Amit Dubey
Modelling the Effect of Discourse on Syntactic Ambiguity Resolution 
Michael Fourman
LDA topic analysis 
Jakub Piatkowski
Estimating white matter tract volume in partial volume voxels with diffusion MRI 
Charles Sutton
Machine Learning and Information Retrieval for Source Code 


iDEA lab lunch

On February 23 we will host a lunchtime meeting of all iDEA Lab projects, 12.00-14.00, in the Informatics Forum.

Lunch will be provided, based on pizza (freshly cooked in our own ovens) and Irn Bru.

11:45 - 12:15 registration and preview
12:15 - 13:00 rapid-fire "elevator-pitch" introductions: 3mins per project.
13:00 - 14:00 posters and pizza

Sign up and come!


2010 call for proposals

One of your New Year resolutions might be to use iDEA lab POPP funding to move forward one of your ideas, from pipe-dream to funded project.

iDEA lab proof-of-principle-prototyping funding can support 3-6 person months of effort to carry out preliminary implementation and experimentation in order to enable new research, research impact or strategic collaborations. http://idea-lab-edinburgh.blogspot.com/2009/09/cse-popp-fund-approved.html

We have already committed to fund a dozen initial projects. Further proposals are invited for a 31st January deadline.

We expect final decisions on proposals received in January to be made on 18th February

Our current projects will be showcased at a pizza lunch in the Forum on 23rd February - further details of this event to follow early in the New Year

With best wishes for the holidays and the New Year.


iDEA Lab adverts

Assistants are needed for the following projects. These projects are at an early stage of discussion, and this is an excellent opportunity to help shape the project itself. Duration and e.g. whether full or part time are negotiable. Offers from non-academic parties to contract work on these projects will also be considered. For further discussion please contact John Lee.
1. Automated assessment of collaborative work in the humanities. [This post is now filled.]
We have observed an increase in quality of work produced by students in a wiki (where they work together) compared to their normal essays. However, we would like to quantify and qualify this rather than going with impressions. One method is comparing the 2 sources by hand using defined indicators of say evidence of using critical thinking, use of primary sources etc. These can be tested by checking for the presence of certain words or word patterns. We now seek to automate aspects of this process. Automated essay marking as such needs to be based on a large annotated corpus, which doesn't exist here, so what is proposed is a more "crude" but still valid approach of assessing quality using linguistic techniques to search for known indicators of quality in the Humanities.
Skills needed or advantageous are:
good programming ability in e.g. python and java;
some experience of text processing techniques;
some knowledge of natural language structure and processing.
2. Faster methods for algorithmic music synthesis, including use of GPUs.
The project is geared towards an implementation of a simulation-based physical modelling sound synthesis framework in a highly parallel form, using the resources at the Parallel Computing Centre (EPCC). The framework itself allows a user (a composer) to create his own, possibly very complex percussion instrument, consisting of a network of connected objects, such as bars and plates, and then to "play" it, by generating a score of hit times, locations, strengths, and then "listen" to it by taking multichannel output from various locations on the instrument. The environment itself has already been prototyped in Matlab, and used by a professional composer in order to generate a short piece of music. Sound quality is excellent, but run times are quite slow---on the order of several minutes to generate a second of sound, for a reasonably complex instrument configuration. It would be extremely useful to have a fast implementation. Another possibility would involve looking at GPUs in order to perform audio synthesis.
Skills needed or advantageous are:
a high level of skill in analysis and programming;
familiarity with MATLAB;
knowledge and experience of programming applications on GPUs.


Project development and proposal

We're working on a drupal site and associated tools that will support the project lifecycle from brain-storming to reporting. This will be launched, real soon now, at http://idea.ed.ac.uk/

We've been waiting for Google Wave with eager anticipation. It may be just what we need for early-stage brainstorming and for communication within a project.

We have a couple of seed Wave accounts. Google is letting each new user make a limited number of nominations, then trickling out invitations as they build capacity.

If you think you may want to use Wave for IDEA lab engagement, and don't have a Wave account, let us know by adding a comment below. We'll find you a nomination as soon as we can (FIFO), provided you agree to use at least two of your nominations to add others below you in the list.

If you already have a Wave account, contact michael.fourman@googlewave.com (don't use this address to send email: despite appearances, this does not function as a regular email address–maybe that will come later) to join the IDEA lab wave.

Thanks to Andrew Ranicki for the, most apposite, cartoon link.


what is iDEA lab?

IDEA lab is a virtual laboratory that will facilitate prototyping of novel applications of state-of-art informatic technologies. IDEA lab will enable and support multidisciplinary "proof-of-principle prototyping" (POPP) projects that develop and use informatic tools and techniques to:

  • collect and manage data,
  • discover and visualise structures and relationships;
  • model, simulate and analyse processes,
  • facilitate collaboration and create communities.

POPP projects will generate new collaborations across a range of interdisciplinary areas, as a key stimulus to an entrepreneurial e-research culture.

  • IDEA lab is an open community of developers and innovators from within and outwith the university, with wide-ranging interests, skills and know-how.
  • It provides a expert pool from which we can rapidly bring together agile teams for POPP projects making novel use of informatic applications and technologies.
  • It supports matchmaking, provides advice, and maintains a repositories of code, data, and expertise – open source, open data, open innovation.
  • It will help researchers to refine proposals, to identify appropriate software, data, and expertise, and to manage POPP projects that will open up new opportunities.
  • IDEA lab will administer POPP funds for specific groups: in particular, it will stimulate proposals, allocate funding, support and monitor POPP projects.

CSE POPP fund approved

Funding of £300k, to be spent in this academic year, has been earmarked by the College of Science and Engineering for POPP projects that will explore and assess new opportunities and pave the way for future funded research. This CSE POPP Fund is open to researchers from any School in the College.

Spending from the CSE POPP fund will be tracked by school. Proposals must specify the expenditure requested and its proposed allocation to schools, which must be approved by the head(s) of the relevant school(s) in the College of Science and Engineering.

Can others play too?

The lab is happy to look at and try to support any projects that make proof of principle prototypes for novel applications of informatics, within or outwith the university.

The CSE POPP fund is targetted at projects that will benefit our research - by enabling further research or tech transfer. Some Head of School has to be convinced that the project will benefit his school's research and agree to fund it from her/his school's account. The actual work could be contracted outside the University.

Everyone can play.

We intend to build a diverse community including developers, researchers and entrepreneurs. Diversity brings strength and innovation. We're interested in applications to research in the arts and humanities, as well as the sciences, and we're open to projects where the anticipated benefit is primarily cultural, social or commercial (but we still want novelty, we're not just a development agency). However, unless there is support from a School in the College of Science and Engineering, as above, the CSE POPP fund can't help: the funding has to come from elsewhere.