US Patent Application 18266876. DESIGN SPACE REDUCTION APPARATUS, CONTROL METHOD, AND COMPUTER-READABLE STORAGE MEDIUM simplified abstract

From WikiPatents
Jump to navigation Jump to search

DESIGN SPACE REDUCTION APPARATUS, CONTROL METHOD, AND COMPUTER-READABLE STORAGE MEDIUM

Organization Name

NEC Corporation

Inventor(s)

Salita Sombatsiri of Tokyo (JP)

DESIGN SPACE REDUCTION APPARATUS, CONTROL METHOD, AND COMPUTER-READABLE STORAGE MEDIUM - A simplified explanation of the abstract

This abstract first appeared for US patent application 18266876 titled 'DESIGN SPACE REDUCTION APPARATUS, CONTROL METHOD, AND COMPUTER-READABLE STORAGE MEDIUM

Simplified Explanation

- The patent application describes a design space reduction apparatus that modifies the design space of a neural network architecture based on characteristics of the dataset it will analyze. - The apparatus acquires original design space information that represents the initial design space of the neural network architecture. - It also acquires dataset characteristics information that represents the specific characteristics of the dataset to be analyzed. - Using this information, the apparatus generates customized design space information that narrows down the design space of the neural network architecture. - The customized design space is tailored to the specific dataset, allowing for more efficient analysis.


Original Abstract Submitted

A design space reduction apparatus () acquires original design space information () that represents an original design space of an architecture of a target neural network. The design space reduction apparatus () acquires dataset characteristics information () that represents characteristics of a target dataset. The target data set is a collection of data to be analyzed by the target neural network. The design space reduction apparatus () generates customized design space information () using the original design space information () and the dataset characteristics information (). The customized design space represents a customized design space of the architecture of the target neural network that is narrower than the original design space.