Transcript Knowledge Technologies 2002-2006

Knowledge Technologies
Scope & focus in 2003
NCP meeting
Jan 27-28, 2003, Brussels
Colette Maloney
Interfaces, Knowledge and Content technologies, Applications &
Information Market
DG INFSO
Challenges
• information overload
• massive, heterogeneous data sets
• unstructured documents (eg e-mails)
• new forms of content
• software programs, sensors, ambient devices …
• complex work processes
• collaborative work flows
• monitoring guidelines
• corporate knowledge practices
• the “zero-latency organisation”, shared KM
• blur between content & services
• Napster: music or P2P?
Work-programme 2003-2004
Objective:
To develop semantic-based and context-aware systems to
acquire, organise, process, share and use the knowledge
embedded in multimedia content. Research will aim to
maximise automation of the complete knowledge lifecycle
and achieve semantic interoperability between Web resources
and services.
Theme
Type
ComponentFoundational
level
research
research
Systemlevel
research
Semantic-enabled
systems and services



Knowledge-based
adaptive systems



Research theme #1
• Semantic-enabled systems & services
for the next-generation Web(s)
• semantic Webs within and across organisations,
communities of interest …
• smart Web services
• automated, self-organising, robust & scaleable
• offering
– networked knowledge discovery
– multimedia content mining
– content-based retrieval across heterogeneous
databases, platforms & networks
– information visualisation
– …
Semantic-enabled systems
and services
Human
Human
Machine-Machine
Knowledge sharing
Knowledge discovery
Virtual Information and Knowledge Environments
and the
“Semantic Layer / Middleware”
Documents Databases
Email
Web
People
Other
Resources
Research theme #2
• Knowledge-based adaptive systems
• reasoning over / acting on
• large volumes of dynamic data
and information
• under uncertain or fuzzy boundary
conditions, guidelines etc
• for
1934
• automated diagnosis & decision support
• highly dynamic & time-critical
applications
• modelling & optimisation
• “anytime-anywhere inferencing”
2004
Knowledge-based adaptive
systems
Smart product development.
2000-2002 DecisionCraft Analytics Ltd.
Send the snow-cats or not?
for industry, science,
education … applications
Urban planning
Accurately predicting arrival times
for aircraft. - NASA - CTAS.
Modelling finance markets
Clinical guidelines support
KT - Basic research
• Foundational research
• formal k- models, methods & languages
• ontology lifecycle (“ecology”)
• methods & tools for creating and maintaining
extensible & interoperable ontologies
– building domain/task specific ontologies
– bootstrapping broader, upper-level ontologies
– catering for multimedia & multilingual aspects
• standards for semantic interoperability
• between Web data, services & process descriptions
• between SemWeb, metadata & multimedia coding
Ontologies can vary enormously in size. Class, Property
or Instance can range from 1-1000s...
KT – Component level research
• Component-level research into baseline
functions & toolsets
• across media / content types
• within common reference architectures
• automated knowledge acquisition
• semantic annotators
•
•
•
•
•
•
•
intelligent Web scrapers or harvesters
semantic search engines
multimedia summarizers
user-friendly editors
visual assistants
natural language tools (eg filtering & routing)
…
KT – System level research
• System integration & validation
• tying together components into innovative
end-to-end systems or services with
• enhanced reasoning capability
• over large-scale & multi-dimensional data sets
• more collaborative/community knowledge sharing
• addressing performance & effectiveness, user
acceptance, ease of integration/customisation,
impact on processes & legacy systems …
• additionality of applicative showcases
• multi-sectoral (reusability & replicability)
• multi-lingual & multi-cultural
KT – System level research
• Candidate areas - purely indicative!
•
•
•
•
•
•
•
scientific & technical resource discovery
personal & collective memory systems
multimedia content mining across the Web
business intelligence
technology watch
corporate portals & intranets
…
Should have multi-sector potential, in progressive
areas - beyond state-of-the-art
Supporting issues
• Research infrastructure
•
•
•
•
metrics & benchmarking, test-bed data sets
public domain ontologies & open source toolsets
registries & locator services
training (researchers, integrators, leading users) …
• Socio-economic issues
• usability, guides & best practice
• new business & revenue models
• awareness & user/supplier dialogue …
• Global reach
• international co-operation …
Summary
•
The vision: the Web as a semantically-annotated
resource shared by humans, software agents &
networked devices
•
Two intertwined goals:
• basic research: “understand” content, master
knowledge embedded in multimedia objects
• applied research: enable smarter, next-generation
Web applications
•
From long-term research through to exemplary
applicative showcases
•
Strong multidisciplinarity with many constituent
disciplines & technologies; significant integration
issues
What kind of project for KT?
Foundational
research
Component
level
research
System level
integration
IP
Yes
Yes
Yes
NoE
Yes
Poss.
NR
STRP
Poss.
Yes
Poss.
NR
Poss.
Yes
=
=
=
Not recommended
Possibility, if clearly justified
Highly recommended
ideal IP for KT
• the “ideal” IP should encompass
• genuine research work
• “engineering” tasks (esp. methods & tools)
• system integration & validation (“total system”
approach)
• along with
• promotion & dissemination of results
• training, awareness & best practice (researchers,
integrators, launching users)
• cooperation & exchanges with related national
and international efforts (incl. standards bodies)
• socio-economic impact & consequences
Outcome of 2003 call
• fewer, bigger projects wrt. FP5
• 55+ meuro available:
• 4-5
• 2-3
• 4-5
• 1-2
IPs
NoEs
STRPs
SSAs
• 11-13 proposals likely to be retained
for funding … highly selective process!
• proposals cutting across knowledge / content /
interface technologies are welcomed
Using the new instruments
• do not artificially create an IP!
• an IP should be THE project in the target area
• an ambitious & progressive endeavour
• with clearly defined milestones & checkpoints
• appropriate use in this sector: not 30 Meuros,
nor 3 Meuros; typically 6-12 Meuros, more
where justified by scope & impact
• an NoE should be interdisciplinary,
include an industry section and / or
a user section
Partnerships
•
consortium
• IPs
7-10 partners, from 3+ countries
• NoEs 4 “core” partners min., from 3+ countries
• STRPs 4-6 partners, from 3+ countries
•
cohesive agenda; competent, committed &
reliable partners
•
•
complementarity: cover all areas you need
duplication of competence
• Necessary for NoEs
• Acceptable for IPs where dictated by project needs
•
industry/SME/academia/NAS participation:
as dictated by project needs
Conclusion
• preserve your credibility: select one
proposal and make it win!
• ensure that the proposal brings out
key innovations
• full depth of participation rather than
long list of organisation names
• critical mass: avoid the “1 FTE per partner” trap
• check relevance of your ideas with EC
staff, at an early stage