Research

The following overview describes shortly different research directions and their fields of applications. The publications listed may serve as entry point for more detailed information.

My main interests include various research questions relating to:

Please click on one of the research directions or scroll down to find further information.

 

 

Computational Cognition of Spatial Objects

Spatial cognition differs among people and situations and is highly influenced by the context in which a spatial object is perceived. Gestalt psychology postulates that we experience things as an integral, meaningful whole, which is more than just the sum of its parts. A computational model which is intended to reflect human spatial cognition must be flexible enough to adapt a flat representation of singular elements and produce a structured representation of the spatial object reflecting the perception of the meaningful whole. We investigate experimentally how humans perceive geometric figures and apply these findings to develop a model for computational cognition of spatial objects. Particular focus lies on the Gestalt-based perception of geometric figures in the context of analogies. While we concentrated on simple, artificial stimuli thus far, future work shall examine complex forms and line drawings of real-world spatial objects.

The computational cognition approach is based on the symbolic analogy model "Heuristic-Driven Theory Projection" (HDTP). Research covers also general research about analogy and the further development of HDTP. We investigate how different cognitive abilities such as learning abstract principles or creativity in problem solving can be explained via analogy and propose a generic architecture for analogical learning. At present, research concentrates on the development of heuristic-driven theory projection as a mechanism for analogy making. Future work will center around a combination of analogical learning with other learning strategies such as inductive learning.

Related Publications from us:

Related Publications from other researchers:

 

 

Analogical Reasoning on Spatial Temporal Patterns

The Brazilian Amazon is the world’s largest tropical rainforest tract and plays an important role for the equilibrium of the global carbon cycle and the global climate. Since the 1970s, deforestation has drastically increased due to human settlement and development of land. This research develops a computational approach to analyze and model land-use change patterns, particularly deforestation in the Amazon. The interplay of socio-economic and ecologic parameters allow for determining different spatio-temporal change patterns. Similar patterns are assumed to underlie similar cause-and-effect relations. Our research develops a model to compare deforestation at different locations in the Amazon for similarities and analogue developments. This approach is based on analogical reasoning for pattern comparison. Parallels between similar previous deforestation processes help to predict the development of current deforestation.

While we first concentrate on deforestation change pattern in the Brazilian Amazon, this research aims at a more general theory of change. Such patterns of change appear in remote sensing data (land cover change) and in urban cadastre (changes in parcels and roads). An analogical comparison of change patterns across different domains will extract and condense the central structures of change which hold domain-independently. One of the important scientific challenges is to devise a general theory for spatial patterns of change.

This research is a cooperation with National Institute for Space Research (INPE), in San José Dos Campos, Brazil.

 

 

Spatial Cognition and Human Computer Interaction

Sketch map is an important representation of spatial information often used in human-human communication. Compared with verbal or textual language, sketch map is more intuitive and supports better human spatial thinking and thus is a more natural way to reflect how people perceive properties of spatial objects and their spatial relations.

However, during the period of sketch map drawing, errors due to human spatial cognition in mind may occur: e.g. distance judgments for routes are judged longer when the route has many turns and direction get straightened up. Similarly, buildings and streets with different shapes are often simplified and depicted as schematic figures like blobs and lines. These errors are not random; rather they appear to be a consequence of ordinary perceptual and cognitive processes.

We aim to develop an approach that analyzes sketch maps formally accounting for distortion and schematization effects mentioned above. We aim to develop a system which allows a user to formulate a spatial query by drawing the desired spatial configuration and get it translated into a symbolic representation to be processed against a geographic database.

Related Publications from us:

Related Publications from other researchers:

 

 

Semantics and Usability

The research focuses on capturing and representing the semantics of spatial information in order to enable effective and accurate information processing. Intelligent methods are required to provide optimal support of users' needs and overcome semantic interoperability problems. Our research focuses on semantic annotationof spatial data and the development of models to measure semantic similarity of natural language expressions.

Consistent and flawless communication between humans and machines is the precondition for a computer to process instructions correctly. While machines use well-defined languages and formal rules to process information, humans prefer natural-language expressions with vague semantics. I investigate experimentally the meaning of natural-language spatial relations and develop a computational model to specify the semantics and reason on spatial relations. Natural-language relations and cognitively plausible operations shall improve query languages of geographic information systems and increase the usability for humans.

Publications from us: