1. Provisional draft 1 ICT Work Programme 2007-08 Challenge 2 Cognition, Interaction, Robotics NCP meeting 19 October 2006, Brussels Colette Maloney, PhD Head of Unit, INFSO E5, Cognition European Commission

2. Provisional draft 2 Cognitive Systems, Interaction, Robotics: motivation software-based systems need to be more… 1.adaptive - to changing tasks and situations; 2.robust – against surprise; 3.effective - improve through anticipating or predicting the future; 4.natural - in dealing with humans and human-like behaviour. In a nutshell: software-based systems should be able to function effectively in circumstances that were not planned for explicitly when the system was designed … an engineering goal

3. Provisional draft 3 Cognitive Systems, Interaction, Robotics: 2007-08 3 intertwined perspectives  development of engineering principles for intelligent integrated systems – cognitive systems – combining many aspects of human competence, and based on research in several disciplines or sub–disciplines.  design of robots or agents that operate autonomously in teams or in cooperation with humans, for example, in exploration, in emergency situations, in productivity improvement, in control, in delivering assistance.  technologies for human–machine-interaction based on a well-grounded understanding of sensor data and human language, the ability to generate concepts and to translate across languages.

4. Provisional draft 4 Cognitive Systems, Interaction, Robotics: 2007-08 Target outcome 1) Engineering principles for artificial systems that:  can achieve general goals in a largely unsupervised way, and persevere under adverse or uncertain conditions; adapt, within reasonable constraints, to changing service and performance requirements, without the need for external re-programming, re- configuring, or re-adjusting. and/or  communicate and co-operate with people or each other, based on a well-grounded understanding of the objects, events and processes in their environment, and their own situation, competences and knowledge.

5. Provisional draft 5 Cognitive Systems, Interaction, Robotics: 2007-08 Focus on one of 3 areas  robots handling different objects and operating autonomously or in cooperation with people  systems (stand-alone or networked) monitoring and controlling material or informational processes  multimodal interfaces and interpersonal communication systems understanding language, gestures emphasis is scientific and technological advance -not application development; applications serve to provide research questions and to demonstrate the impact of conceptual or technical innovation.

6. Provisional draft 6 Cognitive Systems, Interaction, Robotics: 2007-08 Challenges  a key issue is how these systems should work – greatly improving robustness etc. requires rethinking the way systems are engineered. Work should: develop and apply engineering approaches,.. and criteria for benchmarking  what we need are theories (systems theories, software architectures, control theories, modelling theories,... ) that will allow us to build these types of systems. Work should: contribute to theory: … learning in artificial systems, interpretation of data, design and architectures…

7. Provisional draft 7 Cognitive Systems, Interaction, Robotics: 2007-08 Challenges  engineering progress will depend on advancing scientific understanding of what both natural and artificial systems can and cannot do, and how and why.  integration of disciplines: artificial intelligence, computer vision, natural language, robotics,… computational neuroscience, cognitive science, psychology,…philosophy of mind,…mathematics,… Work should explore:  use of advanced components, biomimetics,..  new/bio-inspired information processing paradigms, models of natural cognition, emergent behaviour,..  new approaches to language understanding and generation

8. Provisional draft 8 Cognitive Systems, Interaction, Robotics: 2007-08 Target outcomes 2) and 3): 2)development of a principled methodology for analysing what is needed to achieve our long term goals; development of methods and experimental scenarios for evaluating performance, eg, how to demonstrate generality, how to measure robustness, etc;  creation of a requirements and capability-led understanding of natural systems that can be utilized across multiple engineering and application domains  creation of a common language to discuss natural and artificial systems in terms of performance - bridges to adjacent disciplines (cognitive neuroscience, psychology) 3)coordination at EU-level  gather momentum and focus efforts of the many and varied relevant research communities in support of this challenge

9. Provisional draft 9 Cognitive Systems, Interaction, Robotics: 2007-08 Proposals should  identify the type(s) of system(s) they want to build and explain why they cannot build it/them now  be highly interdisciplinary  proposals should complement and may also extend current projects – rather than try to be different (see links at end)  be ambitious – beyond incremental progress!  involve only those partners needed to achieve the goals and no more key issue is how these systems should work – rethinking the way systems are engineered

10. Provisional draft 10 Links More (FP6 project descriptions, presentations, consultations, reports) at  http://www.cognitivesystems.eu http://www.cognitivesystems.eu  http://cordis.europa.eu/ist/ic/index/htm http://cordis.europa.eu/ist/ic/index/htm See also  http://www.eucognition.org http://www.eucognition.org  http://www.roboticsplatform.com http://www.roboticsplatform.com Contact Points:  Cecile.Huet@ec.europa.eu Cecile.Huet@ec.europa.eu  Colette.Maloney@ec.europa.eu Colette.Maloney@ec.europa.eu Challenges 2 & 4 info day in Lux: Jan 24-26, 2007 Cognitive Systems, Interaction, Robotics: 2007-08