Managing contextual artificial neural networks with a service-based mediator
This addresses integration challenges for researchers and practitioners using varied AI systems, though it appears incremental as it builds on existing service-based approaches.
The paper tackles the problem of disparate AI systems operating on different platforms with specific goals by introducing a mediator framework concept, proposing several architectures to combine signals from these networks for formalized high-level logic and signal processing.
Today, a wide variety of probabilistic and expert AI systems used to analyze real world inputs such as unstructured text, sounds, images, and statistical data. However, all these systems exist on different platforms, with different implementations, and with very different, often very specific goals in mind. This paper introduces a concept for a mediator framework for such systems and seeks to show several architectures which would support it, potential benefits in combining the signals of disparate networks for formalized, high level logic and signal processing, and its possible academic and industrial uses.