|Dr. Marco Ragni||Dr. Lars Konieczny|
|Dipl. Psych. Thomas Fangmeier||Dipl. Ling. Sven Brüssow|
- Prof Dr. Joachim Funke (Universität Heidelberg)
- Dr. Sabine Ohlendorff (Uniklinik Freiburg, Abt. Röntgendiagnostik)
- Prof Dr. Christoph Schlieder (Universität Bamberg)
- Prof Dr. Markus Knauff (Universität Gießen)
Currently, formal approaches and cognitive models that deal with constraint satisfaction problems for spatial domains do not have much in common. The main hindrance lies in the inaccessibility of the human reasoning processes. Therefore, no theoretical basis and no algorithmic specification on how humans reason about spatial problems have been established yet, which would be the foundation for the development of a cognitive complexity theory for human spatial problem solving. Potential applications of such a theory are in the improvement of human-machine interaction tasks.
In this project, we address the desideratum of a cognitive complexity theory from a formal and computational side by using methods from artificial intelligence (AI) flanked by behavioral experiments to identify the influence of operations in reasoning and planning.
This interdisciplinary project intends to
(i) identify factors that influence human spatial reasoning and planning processes in contrast to formal approaches in constraint satisfaction,
(ii) formally specify these with the help of formal and computational techniques, and
(iii) develop a cognitive complexity measure for human spatial reasoning.
A central objective of this project is to identify the factors that influence relational reasoning and planning difficulty (which we call /cognitive complexity/) in human spatial cognition and to model cognitive complexity by methods from artificial intelligence (AI). Furthermore, we will determine the factors that are responsible for errors and allow for an identification of fallacies during mental reasoning processes. Our starting point is a computational model for human spatial relational reasoning that has been developed in the former SFB/TR 8 project R2-[BackSpace]. This formal framework generates mental models on the basis of given (spatial) information and manipulates these mental models by executing specific operations. By assigning unit costs to each step of operation, empirical phenomena can be explained. We identified a high correlation between the ability to perform on relational tasks and the ability to follow a large number of spatial operations in the formal model. These results provide the basis for analyzing cognitive complexity as – roughly speaking – the need for determining operational power, i.e. determining the required mental model operations for constructing and manipulating mental models. In this project we will:
(i) theoretically analyze different cognitive and formal complexity measures, compile a catalogue of formal spatial reasoning problems, and investigate whether we can support our theoretical assumption of an attention focus in spatial mental models by eye tracking studies,
(ii) identify empirically what kind of operations are performed in spatial planning tasks and isolate the determinants of cognitive complexity;
(iii) complement our empirical investigations by formal methods (modal logic and conceptual graphs) to provide a formal basis for cognitive complexity and to compare formal complexity measures (e.g. Kolmogorov complexity) with our own as well as other cognitive complexity measures (e.g. Halford’s relational complexity theory).
These investigations of cognitive complexity complement and extend the general cognitive modeling approach of the project R1-[ImageSpace]. The overall goal of R8-[CSpace] is to develop – for spatial reasoning and planning problems – a cognitive counterpart of formal complexity theories in computer science.