Bisociation Networks for Creative Information Discovery

While frequently discussed in cognitive science, psychology and related literature, there does not seem to exist a serious attempt at trying to formalize and computerize the concept of bisociation. Discovering bisociative relationships requires sufficiently large and diverse underlying information bases, demanding dedicated data integration, storage, distribution and processing strategies.

We anticipate that the BISON system will provide truly creative solutions in an interactive environment that implements novel knowledge integration, network visualisation and machine learning methods to aid creative discovery. BISON builds on widely researched methodologies such as association rule learning, analogical, metaphoric and case-based reasoning.

The BISON implementation of bisociation will be a unique concept, driven by the need for bisociative discovery in innovative domains, by the effective and efficient integration of very large and highly heterogeneous information sources, and the confluence of human and machine capabilities.