We have all heard of the success story of the discovery of a link between the mental problems of children and the chemical pollutants in their drinking water. Similarly, we have heard of the 1854 Broad Street cholera outbreak in London, and the linking of it to a contaminated public water pump. These are two highprofile examples of bisociation, the combination of information from two different sources.

This is exactly the focus of the BISON project and this book. Instead of attempting to keep up with the meaningful annotation of the data floods we are facing, the BISON group pursued a network-based integration of various types of data repositories and the development of new ways to analyze and explore the resulting gigantic information networks. Instead of finding well-defined global or local patterns they wanted to find domain-bridging associations which are, by definition, not well defined since they will be especially interesting if they are sparse and have not been encountered before.

The present volume now collects the highlights of the BISON project. Not only did the consortium succeed in formalizing the concept of bisociation and proposing a number of types of bisociation and measures to rank their“bisociativeness,” but they also developed a series of new algorithms, and extended several of the existing algorithms, to find bisociation in large bisociative information networks.

From a personal point of view, I was delighted to see that some of our own work on finding structurally similar pieces in large networks actually fit into that framework very well: Random walks, and related diffusion-based methods, can help find correlated nodes in bisociative networks. The concept of bisociative knowledge discovery formalizes an aspect of data mining that people have been aware of to some degree but were unable to formally pin down. The present volume serves as a great basis for future work in this direction.