Using Nonlinear Dynamical Attention Allocation to Focus Probabilistic Logical Inference Upon Relevant Information

In a previous article, the authors described a series of experiments combining probabilisitic logical inference with an artificial economics based attention allocation system. In those experiments, the authors compared their results with those from two standard examples chosen from the Markov Logic Networks literature. The examples were insufficient to determine the full usefulness of the integrated system, as the information provided in the examples was precisely the information required for inference. Due to limitations of the test suite, any attention allocation system would be unable to provide additional direction. In this current followup article, the authors describe a new series of experiments and tests intended to demonstrate the effective utilization of attention allocation for inference control. The authors conclude by demonstrating the success, on the first of these experiments, of the cognitive synergy provided via an integration of attention control and probabilisitic inference mechanisms.