SPATIAL REASONING | METHODS AND SYSTEMS
Currently, we develop practical methods (implemented as general purpose tools) for declarative spatial representation and reasoning based on the following declarative programming paradigms:
- Rule-based inference with Constraint Logic Programming (CLP)
- Structured spatio-temporal relational learning with Inductive Logic Programming (ILP)
- Non-monotonic spatial reasoning based on answer-set programming (ASP) and ASP Modulo Theories (ASPMT)
We are currently especially interested and concerned with firmly building-on and ensuring compatibility with the state-of-the-art in the knowledge representation and reasoning (KR) community.
RULE-BASED INFERENCE
RELATIONAL LEARNING
EXPLANATION
The ability to declaratively specific a domain of interest using a rule-based programming framework such as logic programming, mixing knowledge types, handling spatio-temporal objects as first-class entities etc.
The ability to learn relational knowledge founded in spatio-temporal relations (and possibly additional knowledge sources) in a manner fully compatible with the rule-based programming framework.
The ability to perform non-monotonic reasoning (e.g., by abduction) in order to support explanatory reasoning (e.g., about space & motion) in a dynamic spatial systems setup.
CLP(QS)
Rule-based
declarative spatial reasoning with
Constraint Logic Programming (CLP).
ILP(QS)
Structured spatio-temporal
relational learning with
Inductive Logic Programming (ILP).
ASPMT(QS)
Non-monotonic spatial reasoning with
Answer Set Programming (ASP)
Modulo Theories (ASPMT).