Cognitive Systems in FP6, scope and focus

Cognitive Systems
in FP6
scope and focus
Colette Maloney
DG Information Society
Slide no 1
Outline
why cognitive systems?
IST vision
what research?
focus of call
where are we starting from?
EOIs, current activities
Slide no 2
IST Vision: Ambient Intelligence
ability of computationally empowered devices to
interconnect with each other and with us
sensors provide a window from world of interconnected
computation into real physical world
using sensors, these devices will sense the world around
us and respond by interacting with the world or by
communicating with us
they need to ‘see’, ‘hear’, …. ‘understand’ and ‘act’
Ambient Intelligence requires perceptual systems
capable of cognition
Slide no 3
Multimodal interfaces - towards cognitive
systems?
Through sensors, machines can identify and recognize
patterns in human behaviour. They can learn how best to
respond and adapt their behaviour accordingly.
– future step: machines
communication
‘understand’
human
forms
of
– context: how to know and how to represent it
– multimodal sensing: one approach to dealing with context
Emergence of interfaces from ‘constrained’ or ‘known’
environments to the real-world environments 
cognitive systems
Slide no 4
Knowledge technologies - towards
cognitive systems?
Semantic Web effort: binding – through ontologies - Web
content to formal description of itself. Content can be
‘understood’ by the computer.
– processing of formalised knowledge by computer: reduces need
for human intervention in tasks such as classification,
transaction processing and decision-making in Web-based
applications
– Shortcoming of current efforts: the annotation process – ie. the
binding of content to descriptors - has to be done manually
Next step: automating this process  cognitive systems
Slide no 5
Cognitive Systems: focus in FP6
methodologies and construction of:
physically instantiated systems integrating
perception, reasoning, representation and
learning
capable of interpretation, interaction and
communication in the real world
to perform goal-directed tasks
Slide no 6
Cognitive Systems: focus in FP6
challenge: context (real-world)
emphasis: build complete systems
approach: multidisciplinary research
expect as output:
methods & approaches for constructing
robust & adaptive systems capable of
cognition
Slide no 7
Cognitive Systems: focus in FP6
applications:
– in image recognition, in behavioural
interpretation, in video annotation, speech
recognition, in automatic categorisation and
classification, in goal-specification and
decision support,…
– can serve to demonstrate and measure
progress
– but applications NOT main target of research
Slide no 8
Cognitive Systems: EOIs
– Cognitive capabilities in multimodal
interfaces
– Interfaces for ubiquitous computing
environments
– projects, not programmes
– integration, not development from scratch
Slide no 9
Cognitive Systems: short history in FP5
cognitive vision systems 2000 - 2002
– robust image recognition
– from application-specific to generalised solutions
– from focus on low-level processing & robustness of
individual components to systems approach where every
component (incl. high-level cognitive functionalities) has a
role to play in assuring robust behaviour of system
– longer-term, largely academic, interdisciplinary effort
8 RTD projects & 1 Network: www.ecvision.info
Slide no 10
Cognitive Systems - related topics
• Multimodal Interfaces, call 1
intuitive multimodal interfaces ... autonomous and capable
of learning and adapting … in dynamically changing contexts
• Beyond Robotics, call 1
focus on open-ended lifelong learning systems
• Disappearing Computer, call 1
focus on open architectures allowing arbitrary combinations
of ‘building blocks’ for ‘universal application’
• Embedded Systems, call 2
focus on networked systems, distributed control
Slide no 11
Cognitive Systems: relevant info.
• Cognitive Vision Systems, workshop report,
June 2000
• Cognitive Systems, workshop report, July
2002
• ECVISION network of excellence,
www.ecvision.info
Slide no 12