Funding strategies for the Semantic Web: current activities and future

trends in the European Union

Stefano Bertolo, European CommissionDG Information Society and Media, Unit E2

stefano.bertolo@cec.eu.int

European Semantic Web Conference Heraklion, 31 May 2005

Presentation outline

• Current Semantic Web activities– FP6 Call 1 and 2

• Future Semantic Web activities– FP6 Call 4– FP6 Call 5– FP6 Call 6– FP7

• Strategies– Understand what we are asking for and why– Understand how we evaluate

Current FP6 activities

• Total FP6 Call 1 and 2 funding: 137M €• Call 1 projects

– AceMedia, Aim@Shape, AgentLinkIII, Alvis, Aspic, DIP, DirectInfo, Knowledge Board 2.0, Knowledge Web, Metokis, Muscle, News, Rewerse, Simac, SEKT, VIKEF

• Call 2 projects– 3DTV, Axmedis, Content4All, GameTools, Holonics, Inccom,

Inscape, IperG, IP-Racine, M-Pipe, NM2, Peng, Polymnia, Reveal This, WalkOnWeb, Worldscreen

• ftp://ftp.cordis.lu/pub/ist/docs/directorate_e/kmcc/call1and2_project_factsheets_v2.doc

Current FP6 activities• FP6 Call 1 and 2 emphasis on

– Convergence between web services and semantic annotation

– Standards for semantic annotation– (semi)automatic extraction of semantic

annotation from text, images, video– Scalability, robustness, distributedness of KR

and reasoning– Environments for KR (methodologies, editors,

mediators)– Lack of temporal/dynamic KR

Future FP6 activities: Call 4

• 152 proposals requesting 530M€• 1360 participants from 39 countries

(reaching beyond EU)• 34.8% public/gov, 33.5% industry, 21.7%

not-for-profit• Selection close to end: expect to fund ~25

projects, ~124Mۥ Continued but not exclusive focus on

semantic web technologies

Future FP6 activities: call 5

• Closing 21 september 2005• http://fp6.cordis.lu/fp6/call_details.cfm?CA

LL_ID=208• No strategic objective from DG INFSO/E2• Infrastructure-oriented (Grid)• Opportunities under:

– Semantic grid– Large test-beds– Large cultural/scientific archives

Future FP6 activities: call 6

• Likely to be published at the end of 2005

• Relatively modest budget: ~10M€

• Focus: search technologies for audiovisual materials

• Rationale: trying to integrate into usable infrastructure components that are presently being developed separately

FP7 activities

• FP7 to run 2007-2013• EC proposal published April 2005

http://www.cordis.lu/fp7• Public consultations Fall 2005• First call early 2007, project start end 2007 • ICT remains top priority, budget may

increase substantially (12B€ requested)• DG INFSO/E2 to continue SW research

support

FP7 novelties

• Projects vs Ideas

• European Research Council

• investigator-driven research projects

• http://www.cordis.lu/fp7/ideas.htm

DG INFSO/E2 in FP7 looking for:

• More multimedia knowledge extraction and reuse:– DG Information Society and MEDIA– Our portfolio is ‘text-heavy’

• Dissemination = shared components/data– Want to see content/media producers use

tools built by SW engineers– Greater emphasis on building communities

beyond researchers,including IT integrators, early adopters

DG INFSO/E2 in FP7 looking for:

• Solutions = things YOU would want to use– Don’t assume you can start from scratch: show

integration of legacy systems– Don’t scare away users with steep learning curve– ‘Eat your own dog food’: if you are not using your

technology, why should others?– Don’t just build individual tools, architect a system in

which they cooperate reliably– Don’t just design standards, show how they come

together in a coherent stack to solve independently motivated problems that can’t be solved with existing tools/standards

DG INFSO/E2 in FP7 looking for:

• Persistence = tools/data and research communities that thrive after EC funding is over– Don’t just expect it will happen: plan for it– Open Source whenever appropriate– Distribute data: people learn from your data as much

as from your code (and data is more expensive to collect)

– Plan for transportable test-beds • Third parties should be able to replicate your experiments• Prepare your data and protocols as you would want to find

them if you were to replicate somebody else’s experiment

Strategies• Understand what we are asking for and why

– why: make EU economy more competitive through use of advanced technology (Lisbon objectives)

• mid-term return on research investment (not to the exclusion of basic research)

• Help us recognise/define FP7 strategic objectives • Scale of ambition is to solve problems, not climb citation index

– what: make sure what you are proposing would be recognized by a reasonable person as part of our published strategic objectives

– what: solve EU-sized problems, not the integration problems of a single SME

Strategies

• Understand how proposals are evaluated– The problem is stated in the SO: from you we want to

hear the solution;– If you have a solution please explain your plan, i.e. all

the technical steps: reviewers can be expected to be knowledgeable

– If you have a plan, explain how you will manage it:• Who will do what, why, with what resources and on what

time-scale• Project fail when things don’t go as planned: do credible

contingency planning

Strategies

• Understand how proposals are evaluated– Your proposal has competition: make sure the

main ideas are exciting and clearly differentiate you

– Avoid echoing keywords from call’s SO– Reviewing is hard work:

• help the reviewers get a very concrete idea of what you want to do;

• Ambition is excellent if backed up by credible plans; unsubstantiated exagerations alienate reviewers.

Strategies

• Look for convergence/reuse– Build tools that will speed progress in other

projects (solve someone else’s problem)– Use data that comes out of other projects

(exploit clean semantics)– Create data that can be used by other

projects (provide clean semantics)