岡島 義憲’s scientific contributions

Interactive computing architectures have been similarly proposed in various computer science fields, such as the interactive proof systems, evolutionary computation, and Marin Minsky's Quasi-Separate Models. So, an interactive structure of two AI circuit parts will be preliminary studied to make an AGI system provide a self-improving function; as that the AGI system is modeled as being assumed to be consisted of an intellectual master circuit and a set of cascade-connected narrow-AI circuits and the narrow-AI circuit to be composed of a few types of similar module circuits.

Introducing spiking signals to neural network operations is an important and fundamental feature of the neuromorphic architectures, because of reducing energy-efficiency dramatically, and for emulating information operations of human cortex. Though, in terms of application accuracy, the Spiking Neural Network, SNN, has not been outperforming the Artificial Neural Networks, the reason of which is regarded as that the SNN is following ideas of existing Deep-Neural-Network topologies and their learning methodologies. The outline of this situation is reviewed and studied for seeing into next deployments of the SNN.