Learning classification device and learning classification method

Abstract

A learning classification device includes a data memory, a learning unit, and a classification unit. The data memory is configured to store training data for learning a decision tree. The learning unit is configured to read a plurality of feature quantities included in the training data from the data memory by single access and derive data of a node based on the plurality of feature quantities, to learn the decision tree. The classification unit is configured to determine where the training data read from the data memory is to be split to from the node, based on the data of the node derived by the learning unit.

Description

technical field [0001] The invention relates to a learning classification device and a learning classification method. Background technique [0002] Attempts have been made in various fields to replace human functions using machine learning, which is known to be related to artificial intelligence (AI), based on large amounts of data. This field is growing day by day, but there are still some problems in the current situation. Typical problems are limitations in accuracy, including generalization performance for extracting general knowledge from data, and limitations in processing speed due to large computational costs. Deep Learning (DL), especially Convolutional Neural Networks (CNN), where the input vector is bounded by a neighborhood, is a well-known algorithm for high-performance machine learning. Compared with these techniques, in the present case, it is known that gradient boosting decision tree (GBDT) is less accurate for input data such as images, sounds, and langu...

AI Technical Summary

Problems solved by technology

If there are many features, a single access cannot read all of them

Therefore, the technique cannot accelerate the learning process

Images