We present a novel approach to handle uncertainty, by assuming an irrational view of the world thus breaking with the tradition of taking a rational view. We propose a solution that handles irrationalities by generating different hypotheses of the environment and testing these for their validity thereby fully automating the abductive loop. By assuming the irrational view, we focus on a fundamental change in relation to existing formal techniques that have been developed over the past decades for dealing with uncertain information. While other methods assume a set of axioms to describe the basic building blocks of the problem domain and use these to build concepts. The new approach, unlike the requirements of a rational system, has no predefined problem domain concepts and any concept can be generated and selected based on their ability to be useful in fulfilling the purpose. This means a completely dynamic ontology is generated and maintained. In this way the proposed system could probably cope with any type of problem domain.

Combining the results of classifiers has shown much promise in machine learning generally. However, published work on combining text categorizers suggests that, for this particular application, improvements in performance are hard to attain. Explorative research using a simple voting system is presented and discussed in the light of a probabilistic model that was originally developed for safety critical software. It was found that typical categorization approaches produce predictions which are too similar for combining them to be effective since they tend to fail on the same records. Further experiments using two less orthodox categorizers are also presented which suggest that combining text categorizers can be successful, provided the essential element of ‘difference’ is considered.

The combination of the Content Addressable File Store (CAFS®; CAFS is a registered trademark of International Computers Limited) and an extension of relational analysis is described. This combination allows a simple and compact implementation of a database query and update language (FIDL). The language has one of the important properties of a “natural” language interface by using a “world model” derived from the relational analysis. The interpreter (FLIN) takes full advantage of the CAFS by employing a unique database storage technique which results in a fast response to both queries and updates.