Duality as a unifying framework for program semantics
Ingrid Rewitzky
University of Stellenbosch, South Africa

Abstract: Formal languages of many different kinds nowadays play a central role in computer science and other disciplines. Semantic models of such languages, both algebraic and relational (with or without topology), are important in reasoning about them. In particular, modal logic, modelled algebraically by Boolean algebras with operators and purely relationally by Kripke frames, has a very well-developed theory and its two semantic models are intertranslatable via duality techniques. In this talk we will illustrate the application of such duality techniques for unifying two independently developed semantic models of programming languages.

Multi-Classifier Systems have fast been gaining popularity among researchers for their ability to fuse together multiple classification outputs for better accuracy and classification. I will first present some current issues in the design of multi-classifier systems before moving on to present our multi-classifier developments for some Bioinformatics problems, mainly gene identification in DNA sequences, but not only. Our neuro-genetic based multi-classifier systems have outperformed the results of other prediction systems for some standard data sets like E.coli or Human DNA datasets.