Recording medium, machine learning method, and machine learning device
A machine learning process including identifying a first axis including elements represented by one-hot vectors and a second axis including elements not represented by one-hot vectors, among a plurality of axes included in tensor-form data, when calculating a core tensor from the tensor-form data via a plurality of intermediate tensors, calculating the core tensor from the tensor-form data by a first process of concatenating elements included in a first element matrix corresponding to the first axis to elements included in a first intermediate tensor among the plurality of intermediate tensors and a second process of calculating a mode product between a second intermediate tensor among the plurality of intermediate tensors and a second element matrix corresponding to the second axis, and performing machine learning of a machine learning model with the core tensor as an input.
1 . A computer-implemented machine learning method comprising:
identifying a first axis including elements represented by one-hot vectors and a second axis including elements not represented by one-hot vectors, among a plurality of axes included in tensor-form data, using a processor;
calculating a core tensor from the tensor-form data via a plurality of intermediate tensors by a first process and a second process, the first process concatenating elements included in a first element matrix corresponding to the first axis to elements included in a first intermediate tensor among the plurality of intermediate tensors, the second process calculating a mode product between a second intermediate tensor among the plurality of intermediate tensors and a second element matrix corresponding to the second axis; and
performing machine learning of a machine learning model with the core tensor as an input.
2 . A machine learning device comprising:
a memory; and
a processor coupled to the memory, the processor being configured to:
identify a first axis including elements represented by one-hot vectors and a second axis including elements not represented by one-hot vectors, among a plurality of axes included in tensor-form data,
calculate a core tensor from the tensor-form data via a plurality of intermediate tensors by a first process and a second process, the first process concatenating elements included in a first element matrix corresponding to the first axis to elements included in a first intermediate tensor among the plurality of intermediate tensors, the second process calculating a mode product between a second intermediate tensor among the plurality of intermediate tensors and a second element matrix corresponding to the second axis, and
perform machine learning of a machine learning model with the core tensor as an input.
3 . A non-transitory computer-readable recording medium storing therein a machine learning program that causes a computer to execute a process, the process comprising:
identifying a first axis including elements represented by one-hot vectors and a second axis including elements not represented by one-hot vectors, among a plurality of axes included in tensor-form data;
calculating a core tensor from the tensor-form data via a plurality of intermediate tensors by a first process and a second process, the first process concatenating elements included in a first element matrix corresponding to the first axis to elements included in a first intermediate tensor among the plurality of intermediate tensors, the second process calculating a mode product between a second intermediate tensor among the plurality of intermediate tensors and a second element matrix corresponding to the second axis; and
performing machine learning of a machine learning model with the core tensor as an input.
4 . The non-transitory computer-readable recording medium according to claim 3 , wherein the machine learning program causes the computer to further perform determining an execution order of the first process and the second process, and
the calculating of the core tensor includes calculating the core tensor by performing the first process
and the second process in the order determined by the process of determining.
5 . The non-transitory computer-readable recording medium according to claim 4 , wherein the determining includes a process of determining that the first process is performed before the second process when the elements of the first axis depend on the second axis and determining that the first process is performed after the second process when the elements of the first axis do not depend on the second axis.
6 . The non-transitory computer-readable recording medium according to claim 3 , wherein the identifying includes a process of identifying an axis of a label mode as the first axis and identifying an axis of a topology mode as the second axis.