<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> <HTML> <HEAD> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1"> <META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7654.12"> <TITLE>RE: [Nnhep] Proposal for the meeting</TITLE> </HEAD> <BODY>  Hello Lev, i will be busy tomorrow from 9 till 5pm and then i have a workshop dinner from 19:30 in Meyrin. Friday looks better, i am busy until 4pm and then free. Thus i would suggest Friday evening and we could continue with a pint. All the best, Tomas -----Original Message----- From: Lev Dudko [<A HREF="mailto:dudko@fnal.gov">mailto:dudko@fnal.gov</A>] Sent: Wed 5/27/2009 11:51 AM To: Tomas Lastovicka Cc: nnhep@projects.hepforge.org; Konstantin.Toms@cern.ch Subject: RE: [Nnhep] Proposal for the meeting     Hello Tomas,  What time is appropriate for you? I will have shift 28.05 till 16.30, after that I am able to meet. On Friday I will have meeting at 9.30 till 11 (11.30) and after that I can meet in any time appropriate for you. It is also possible to meet on Saturday.        Best regards,             Lev ? ???, 27/05/2009 ? 07:31 +0100, Tomas Lastovicka ?????: > > Dear all, > > i am traveling to CERN today from Montenegro. Unfortunately, i will be > busy with the CLIC workshop on both Thursday and Friday the whole day > and thus i am not sure i could join you at 29.05.09 11.00-12.00. I > will contact you once at CERN. > > Best, > > Tomas > > > > -----Original Message----- > From: nnhep-bounces@projects.hepforge.org on behalf of Lev Dudko > Sent: Tue 5/26/2009 12:02 PM > To: nnhep@projects.hepforge.org > Cc: Konstantin.Toms@cern.ch > Subject: [Nnhep] Proposal for the meeting > > >      Dear Colleagues, >   we have discussed privately, the possible time of the first meeting > dedicated to NNHEP proposal. Looks like 3 people are in CERN now. > Sasha > can be connected with skype. >     The proposed time is 29.05.09 11.00-12.00  If you want to change > the > time, lets discuss it. > >     The first version of the questions to discuss is the following: > > 1. We have to define the structure of interfaces between different > modules and understand what should be the input of the module and the > output. >    Let me remind, what I mean the module, this is one independent > component e.g. one neural network or some other algorithm which has > some > input and some output, internal set of the parameters and some way of > graphical representation of what is happened with the input in this > module. > > 2. We can start to discuss what modules (algorithms) we would like to > see as the components of the system and what parts we can take from > the > available external code. > The first set could be the following: > Feed-Forward Neural Network, Bayesian Neural Network, Boosting > algorithm, Decision trees, Self Organizing Maps, Likelihood > functions, ... >       Some of the modules are realized already in the exist packages > (e.g. Feed-Forward NN training with perfect optimizations and training > methods are realized in MLPfit package; BNN is realised in RuleFit, > etc.) > > 3. Practical realization of the discussed plans and how we plan to > proceed with the code development. > >         Please, correct/add the set of questions above and provide > your > opinion. >               Best regards, >                     Lev > > > > -- > >  Lev V. Dudko            e-mail:dudko@fnal.gov >  t. +7(495)9395881      <A HREF="http://top.sinp.msu.ru">http://top.sinp.msu.ru</A> > > > > --  Lev V. Dudko            e-mail:dudko@fnal.gov  t. +7(495)9395881      <A HREF="http://top.sinp.msu.ru">http://top.sinp.msu.ru</A> </BODY> </HTML>