From Ken Forbus' survey for
the CRC Handbook of Computer Science and Engineering (paper):
Qualitative reasoning is the area of AI which creates
representations for continuous aspects of the world, such as space,
time, and quantity, which support reasoning with very little
information. Typically it has focused on scientific and engineering
domains, hence its other name, qualitative physics. It is motivated by
two observations. First, people draw useful and subtle conclusions about
the physical world without differential equations. In our daily lives we
figure out what is happening around us and how we can affect it, working
with far less data, and less precise data, than would be required to use
traditional, purely quantitative methods. Creating software for robots
that operate in unconstrained environments and modeling human cognition
requires understanding how this can be done. Second, scientists and
engineers appear to use qualitative reasoning when initially
understanding a problem, when setting up more formal methods to solve
particular problems, and when interpreting the results of quantitative
simulations, calculations, or measurements. Thus advances in qualitative
physics should lead to the creation of more flexible software that can
help engineers and scientists.
Current research spans all aspects of the theory and applications of
qualitative reasoning about physical systems.
- Cognitive modeling (e.g., cognitive theories of reasoning
about physical systems, theories and experiments concerning human
reasoning and learning of mental models, QR models for spatial
reasoning, cognitive maps, cognitive robots);
- Techniques (e.g., qualitative simulation, ontologies,
management of multiple models, reasoning over time and space,
mathematical formalizations of QR, qualitative algebras, qualitative
dynamics, qualitative kinematics, qualitative optimization);
- Task-level reasoning (e.g., design, planning, monitoring,
diagnosis and repair, explanation, tutoring and training, process
control and supervision);
- Applications (e.g., engineering, education, business,
biology, chemistry, ecology, economics, social science, environmental
science, medicine, and law);
- Intersection with other modeling approaches (e.g., system
dynamics and bond-graphs, signal processing, numerical methods,
statistical techniques, differential equations);
- Knowledge acquisition methods (e.g., model building tools and
techniques, automated model construction and machine learning,
acquisition of models from data).
- Theoretical foundations of qualitative reasoning techniques.
Here are a few ways to get started learning more about what's going on
in QR these days:
- The 2005 workshop: QR 05
- Archive of previous QR workshops: 1987-2004
- Special issue of AI Magazine: Winter
- Current Topics in Qualitative Reasoning, Editorial Introduction by
Bert Bredeweg and Peter Struss.
- Model-Based Systems in the Automotive Industry, by Peter Struss
and Chris Price.
- Qualitative Modeling in Education, by Bert Bredeweg and Ken
- Qualitative Spatial Reasoning Extracting and Reasoning with
Spatial Aggregates, by Chris Bailey-Kellogg and Feng Zhao.
- Model-Based Programming of Fault-Aware Systems, by Brian C.
Williams, Michel D. Ingham, Seung Chung, Paul Elliott, Michael
Hofbaur, and Gregory T. Sullivan.
- Qualitative Reasoning about Population and Community Ecology, by
Paulo Salles and Bert Bredeweg.
- Mathematical Foundations of Qualitative Reasoning, by Louise
Trave-Massuyes, Liliana Ironi, and Philippe Dague.
- Learning Qualitative Models, by Ivan Bratko and Dorian Suc.
- Model-Based Computing for Design and Control of Reconfigurable
Systems, by Markus P. J. Fromherz, Daniel G. Bobrow, and Johan de