Dr. Diarmuid O'Donoghue's research is on analogical reasoning and its role in scientific computational creativity. We have developed computational models of how people solve a variety of problems using analogical comparisons. This work is greatly informed by cognitive science. In pure computer science terms however, our analogy models are based on isomorphic structure mapping between the source and target analogs - represented as a pair of k-edge & j-node coloured graphs. We have also adapted these algorithms to use analogous information in domains like: topographic maps, geometric proportional geometric analogies and C# source code implementations. These analogies were used to suggest solutions to problems in each of the problem areas - identifying composite structures like universities, solving GPA problems and generating specifications for given source code implementations.
Another thread of my research is focused on Evolutionary Algorithms with Genetic Repair for constrained optimisation. We are exploring biologically inspired models of Genetic Repair that enforce constraints on the evolutionary search process. Of special interest is the genetic repair process proposed in the plant Arabidopsis thaliana (Nature 434, 505-509), suggesting a "cache" of ancestral genomic information to help the repair process - acting as an extra-Mendelian inheritance mechanism. Our work makes two distinct contributions. First we are developing very efficient general-purpose algorithms for solving constraint problems. Secondly, we are exploring the plausibility of the proposed repair mechanism through our computational experiments.