Learning apparatus, inferring apparatus, computer program, computer system, inferring method, and learning method

A learning apparatus that learns a set value in a machine learning model based on learning data in which a value with respect to a one-dimensional or two-or-more dimensional space in an image is represented as a discrete value at a sample point in the one-dimensional or two-or-more dimensional space in the image, the learning apparatus comprising:

• an input unit that receives an input of learning data;
• a predictable area determining unit that determines whether learning data input to the input unit has a predictable area in which there is low necessity of learning the set value;
• a machine learning model to which data is input from the input unit;
• a learning process executing unit that performs a process of learning a set value in a machine learning model to be learned based on learning data, by using the machine learning model to be learned;
• and
• a learning result storage unit that stores therein the set value, wherein
• the learning process executing unit outputs data required for inferring an area, which is not a predictable area, to the machine learning model, and does not output at least a part of data required for inferring the predictable area to the machine learning model.

An inferring method for causing an inferring apparatus or a computer program to function to perform a predetermined inferring process with respect to data, in which a value with respect to a one-dimensional or two-or-more dimensional space in an image is represented as a discrete value at a sample point in the one-dimensional or two-or-more dimensional space in the image, by using a machine learning model, the inferring method comprising:

• an input step of receiving an input of target data;
• a predictable area determining step of determining whether the input of the target data at the input step has a predictable area in which an inferring result of the predetermined inferring process can be easily predicted;
• an inferring process executing step of performing a predetermined inferring process with respect to the target data by using the machine learning model to which data is input;
• and
• a predicted data generating step of performing predetermined data processing with respect to a data area, which is determined as the predictable area, wherein
• at the inferring process executing step, data required for inferring an area, which is not a predictable area, is output to the machine learning model, and
• at the predicted data generating step, the data processing is performed in an average time shorter than that of the inferring process.

A learning method for causing a learning apparatus or a computer program to function to learn a set value in a machine learning model based on learning data, in which a value with respect to a one-dimensional or two-or-more dimensional space in an image is represented as a discrete value at a sample point in the one-dimensional or two-or-more dimensional space in the image, the learning method comprising:

• an input step of receiving an input of learning data;
• a predictable area determining step of determining whether the input of the learning data at the input step has a predictable area in which there is low necessity of learning the set value;
• a learning process executing step of performing a process of learning a set value in a machine learning model as a target of learning based on the learning data, by using the machine learning model as the target of learning;
• and
• a learning result storing step of storing the set value, wherein
• at the learning process executing step, data required for inferring an area, which is not a predictable area, is output to the machine learning model, and at least a part of data required for inferring the predictable area is not output to the machine learning model.