Qualitative Representations for Robots
Papers from the 2014 AAAI Spring Symposium
Nick Hawes Program Chair
Technical Report SS-14-06
112 pp., $30.00
ISBN 978-1-57735-646-2
[Add to Cart] [View Cart]
The fields of AI and robotics have many approaches to representation and reasoning. This symposium focuses on one approach, which has been growing in popularity in recent years: qualitative representations. Such representations abstract away from the quantitative features that underlie many physically situated systems, providing compact, structured representations, which omit (unnecessary) detail. Qualitative representations have many advantages, including naturally encoding semantics for many systems, being accessible to humans, providing smaller state spaces for learning, allowing to build robust and complex applications and also suitability for communication. These advantages have seen them being increasingly used in intelligent, physically-grounded systems. This work is being done across many different subfields of AI including knowledge representation and reasoning, planning, learning, and perception. We strongly believe that the time is now right to bring these disparate groups together to share experiences and technical knowledge. We also wish to connect recent robotics work on qualitative representations to the rich history of related ideas in AI.