摘要 |
PURPOSE:To provide the inference method by which inference processing can be executed at high speed. CONSTITUTION:AI inference is performed by an inference system computer based on inference information received from a general-purpose system computer for executing general-purpose processing (steps 101-104) and the inference data in that case are preserved (step 105) and used for the inference after the next time. At the same time, the transmission of inference information from the general-purpose system computer with respective times of inference after the next time is limited just for the changed part (steps 109-111) so as to efficiently perform inference. |