Disussing ALAMAS proposal(s) in Barcelona



Date view Thread view Subject view Author view Attachment view

From: D. Kazakov (kazakov@cs.york.ac.uk)
Date: Thu 30 Jan 2003 - 13:16:01 GMT


Dear colleagues,

here are some ideas about the discussion on EU proposals at the ALAD SIG
meeting in Barcelona next week.

There are three decisions that would always have to be made in the
preparation of a successful proposal:

1. Choose the team of partners and scientific content

2. Choose the thematic area to which to submit the proposal and decide
on the type of proposal (STREP, IP, NoE, Thematic Network). These two
issues are highly correlated, as, for instance, submitting to the FET
area would mean preparing either an IP in one of the 3 areas described
in the call (Complexity Science, the disappearing computer, Robots) or a
much smaller STREP or Thematic Network proposals to the open call in the
same area.

3. Find the most closely related initiative (project proposal) in the
area of choice that has been successful in the past and achieve at least
some level of integration in the proposals (highly recommended by the
Commission).

We, the SIG meeting organisers, expect points 2 and 3 to be actively
discussed, and I have prepared to give some background and food for
thought on these two issues. However, no final decision could be made
without
more information about some of the other initiatives (e.g., AgentLink
III NoE), and direct enquiries with the European Commission Officers (EU
has encouraged contacts with them at the early stages of drafting the
proposal). It will also be good to know if we can cluster the potential
partners around a few topics of interest.

All this leads to the suggestion to approach point 1 above in a
bottom-up fashion, i.e., 

* assume that the partners will be split in a number of STREP projects
(1-2M Euro, 4-5 partners each) and start discussing the content of (and
partners in) these proposals. This will lead to the creation of a strong
core of partners and the development od proposals, which could at later
stage be combined in an IP or a network proposal if we feel or/and EU
recommends that as a good idea.

* The initial two topics around which two STREPs could start to emerge
are:

Topic 1. Reinforcement Learning for Agents and MAS (discussion will be
lead by D. Kudenko and M. Strens (QinetiQ) )

Topic 2. Evolutionary Learning for Agents and MAS. (discussion will be
lead by D. Kazakov) 


An integrated project proposal could emerge by including Topics 1-2 as
workpackages. Also, a third workpackage integrating the two approaches
is likely to be added. Further, the proposal should show a degree of
vertical 
integration between the theory and technology, with, possibly, some high
profile applications to demonstrate the merits of the research.

Other suggestions for STREP proposals/topics that are likely to attract
substantial interest are also welcome. The person/institution
proposing the topic should be ready to get actively involved in the
coordination of the proposal.

This is all for now. See you in Barcelona.

Best greetings,

Dimitar


-- 
Dr Dimitar Kazakov
Lecturer in Machine Learning

AIG, Dept. of Computer Science
University of York
Heslington, York 
YO10 5DD, UK
tel. +44 (0)1904 43 4775
fax: +44 (0)1904 43 2767
http://www.cs.york.ac.uk/~kazakov/

Date view Thread view Subject view Author view Attachment view