Date: 12/10/2016

AI algorithm automates deep-learning

Mitsubishi's new artificial intelligence algorithm utilises the learning data and high-level inferences to design deep learning automatically without expert's inputs.

Mitsubishi Electric is presenting this new system at the International Conference on Neural Information Processing going to be held during October 16 through 21 at Kyoto University and will publish a paper in Lecture Notes in Computer Science.

The algorithm adapts to specific purpose and also expected to support the effective structuring of networks and reduce the trial and error of design.

The benefit is in saving of time to development, where Mitsubishi claims the new algorithms reduces development time and also cost. This tool can develop deep learning structures within hours compared to days or weeks taken by experts.

The deep learning algorithm designs the structures and initial parameters of deep learning by extracting the most characteristic data from learning data without overlaps and without involving experts, who are required to make high-level inferences about specific environments. The developers Mitsubishi based this algorithm on "Neocognitron," a image processing algorithm considered based on visual cortex.

The new system allows use of artificial intelligence for diverse business
applications.