Cesena, 5 November 2010

Cesena, 5 November 2010

When

Friday 5/11/2010, 10:00 - 12:00

Where

DEIS Cesena, via Venezia 42, 2nd Floor, "Sala Riunioni".

Who

  1. Matteo Casadei
  2. Ambra Molesini
  3. Sara Montagna
  4. Elena Nardini
  5. Andrea Omicini
  6. Mirko Viroli

What

— Introduction to the project

Mirko has done a brief introduction to the SAPERE project remembering the scenario, the project's goals and the work packages structure. In particular, the goals of WP1, WP2 and WP4 have been well explained since our group is deeply involved in these WPs. 

About WP4, Mirko has noted a mismatch between the expected outputs presented in SAPERE specification and our view.
In particular SAPERE specification reported as part of the output the development of some <i>library</i>, a concept that does not fit very well our viewpoint over the infrastructure. Additionally, Mirko pointed out that we might want to develop in isolation one (or possibly more) prototypes of SAPERE infrastructure, with the goal of validating the model and experimenting on it.

— Introduction to the SAPERE idea and open issues

Mirko has recalled the SAPERE idea by means of a display example.

The example has been very usefull for discussing about our key topics:

  • LSA. The Live Semantic Annotation (LSA)  should be modelled as a sort of tuple, with an ID, a semantic content, and some dangling bonds. It may also contain a declarative specification of behaviour, to be interpreted by the infrastructure -- though in a simpler way than e.g. TOTA.
  • Eco-law. Eco-laws represent the coordination rules used by the infrastructure in order to manage LSAs. In particular these Eco-laws should be represented according to the chemical metaphor: they take few LSAs and produce some or  transform existing ones, e.g. by chemical bonding.
  • Topology. The topology represents the spatial structure of the eco-system: a concept that is very much dependent to infrastructure considerations (and should be hence deepened in WP4)<br/>

When an agent enters a SAPERE space it injects its LSA, evolved by the SAPERE infrastructure by means of the application of Eco-laws. Such Eco-laws modify an LSA e.g. by binding it with other (matching) LSAs. The agents observing their LSAs will be able to understand what to do next.
The SAPERE infrastructure should support the concept of spatial self-organisation (the local interactions should become global interaction through the creation of a net of LSAs -- to be deepened in WP2) and situation recognition (capturing certain patterns of LSAs and reify the situation as a new, possibly diffused LSA -- to be deepened in WP3).

As an open issue, Mirko has spoken about the importance of modelling ecological behaviour: e.g., if an LSA is not used it decays over the time -- do we want to address this issue from the beginning?

— Organisation and schedule of our activities

Mirko has explained the structure of the activities where we are involved, with particular regard to the leaders of the different activities:

ActivityLeaderPeople Involved
T1.1ViroliMolesini, Montagna, Omicini
T1.2NardiniOmicini, Viroli
T1.3MolesiniCasadei, Montagna, Omicini
WP2CasadeiMontagna, Viroli
WP4CasadeiOmicini, Nardini, Viroli
WP6OmiciniViroli

Note: the leader of T1.3 is tentative, for this task will start only on Month 12 (Andrea will back up Ambra if needed).

— Next steps

The next steps will be:

  • stimulating discussion with other research units, e.g. Ginevra (about diffusion and globality of eco-laws), STA (about semantic matching), Modena (about connections with middleware and agents)-at the moment we do not consider discussion about the methodological aspects. Some member of the other units will be invited to participate to our meetings in order to stimulate collaboration.
  • we note that Matteo will stay in Geneve for a week for a meeting concerning WP2.
  • our initial research activities will be targeted to month 4, when an early report on the model will be provided at the second SAPERE meeting (deepening lsa, eco-laws, interaction patterns, abstract middleware, abstract application).
    <p/>
    In addition Mirko has proposed to try to organise weekly internal meetings to bootstrap the research of this unit, and later to move to meetings taking place every two weeks.

— Notes

Some final remarks to the people involved in this unit:

  • each paper posted on the SAPERE web site has to be provided with a mini-abstract indicating why the paper (or paper sections) is important in the SAPERE context. Matteo will soon publish an initial list of papers.
  • remember to acknowledge SAPERE in your SAPERE-related papers!
  • pay attention to the bureaucratic deadlines of your research grants!
  • be prepared to attend the Awareness meeting if neither Andrea or Mirko will be able to participate 
  • pay attention to the <i>co-authoring</i> issue.

— Next meetings

Next meeting will be on 12th November 2010 in Cesena at 10:00 am.
 Proposed topics:

  • a strategy for continuous logging our work (Andrea Omicini). This presentation will be about how to document in the proper way the work we will do inside the SAPERE project with particular regard to summary of the meetings, papers, activities
  • Technical presentations:
    • connection with bio-inspired models of computation (Sara Montagna).
    • approaches to semantic matching (Elena Nardini). 
    • first sketch about methodological issues (Ambra Molesini). 

The subsequent meeting should be on 19th November 2010 in Cesena at 10:00 am.
 Proposed topics:

  • self-organisation (Matteo Casadei). Technical presentation.
  • summary of the Geneve meeting (Matteo Casadei).