A non-transitory computer-readable recording medium has stored therein a machine learning program causing a computer to execute a process including, estimating a first intensity estimation value that indicates an intensity of a data change in data after change, performing a first update, by using an intensity adjustment learning device, on a learning device such that the data change of the intensity indicated by the first intensity estimation value is added to data before change, generating post-change data by adding the data change to the data before change by using the learning device that has been subjected to the first update, estimating a second intensity estimation value that indicates the intensity of the data change in the post-change data by using the intensity estimation AT, and, causing the intensity adjustment learning device to perform first learning such that the second intensity estimation value approaches the first intensity estimation value.
Full Text
What is claimed is: