Method, apparatus, and non-temporary computer-readable medium that uses a vector neural network

A method of causing one or more processors to execute using a model learned in advance, in which the model is learned to reproduce correspondence between a plurality of first data elements included in a first data set and a pre-label corresponding to each of the plurality of first data elements, and the model uses an algorithm of a vector neural network type and has a plurality of neuron layers, each neuron layer of the plurality of neuron layers has one or more neuron groups, each of the one or more neuron groups has one or more neurons, and each of the one or more neurons outputs first intermediate data based on at least one of a first vector and a first activation, wherein a size of a partial range defined by multiplying a height, a width and a depth in each neuron layer of the plurality of neuron layers is different for some neuron layers;

• the method comprising:
• inputting the first data set into the learned model, and acquiring the first intermediate data output by the one or more neurons by being included with a first partial range which is a partial range on the plurality of first data elements and to which each neuron corresponds based on at least one of the first activation and the first vector included in each neuron of each of the plurality of neuron layers;
• inputting a second data element into the model, and acquiring second intermediate data based on at least one of a second vector and a second activation included in the one or more neurons, for each of the plurality of neuron layers;
• and
• calculating a similarity between the first intermediate data and the second intermediate data, for the plurality of neuron layers.

An apparatus comprising:

• one or more processors, wherein
• the one or more processors is configured to execute:
• performing learning of a model that is an algorithm of a vector neural network type to reproduce correspondence between a plurality of first data elements included in a first data set and a pre-label corresponding to each of the plurality of first data elements, in which the model has a plurality of neuron layers, each neuron layer of the plurality of neuron layers has one or more neuron groups, each of the one or more neuron groups has one or more neurons, and each of the one or more neurons outputs first intermediate data based on at least one of a first vector and a first activation, wherein a size of a partial range defined by multiplying a height, a width and a depth in each neuron layer of the plurality of neuron layers is different for some neuron layers;
• and
• inputting the first data set into the learned model and acquiring the first intermediate data output by the one or more neurons by being included with a first partial range which is a partial range on the plurality of first data elements and to which each neuron corresponds based on at least one of the first activation and the first vector included in each neuron of each of the plurality of neuron layers.

An apparatus comprising:

• a storage device that stores a model learned to reproduce correspondence between a plurality of first data elements included in a first data set and a pre-label corresponding to each of the plurality of first data elements, in which the model uses an algorithm of a vector neural network type and has a plurality of neuron layers, each neuron layer of the plurality of neuron layers has one or more neuron groups, each of the one or more neuron groups has one or more neurons, and each of the one or more neurons outputs first intermediate data based on at least one of a first vector and a first activation, wherein a size of a partial range defined by multiplying a height, a width and a depth in each neuron layer of the plurality of neuron layers is different for some neuron layers;
• and
• one or more processors, wherein
• the one or more processors is configured to execute:
• inputting the first data set into the learned model and acquiring the first intermediate data output by the one or more neurons by being included with a first partial range which is a partial range on the plurality of first data elements and to which each neuron corresponds based on at least one of the first activation and the first vector included in each neuron of each of the plurality of neuron layers;
• inputting a second data element into the model, and acquiring second intermediate data based on at least one of a second vector and a second activation included in the one or more neurons, for each of the plurality of neuron layers;
• and
• calculating a similarity between the first intermediate data and the second intermediate data, for the plurality of neuron layers.

A non-transitory computer-readable medium storing instructions for causing one or more processors to execute using a model learned in advance, wherein

• the model is learned to reproduce correspondence between a plurality of first data elements included in a first data set and a pre-label corresponding to each of the plurality of first data elements,
• the model uses an algorithm of a vector neural network type and has a plurality of neuron layers, each neuron layer of the plurality of neuron layers has one or more neuron groups, each of the one or more neuron groups has one or more neurons, and each of the one or more neurons outputs first intermediate data based on at least one of a first vector and a first activation, wherein a size of a partial range defined by multiplying a height, a width and a depth in each neuron layer of the plurality of neuron layers is different for some neuron layers;
• the instructions causes one or more processors to further execute:
• inputting the first data set into the learned model and acquiring the first intermediate data output by the one or more neurons by being included with a first partial range which is a partial range on the plurality of first data elements and to which each neuron corresponds based on at least one of the first activation and the first vector included in each neuron of each of the plurality of neuron layers;
• inputting a second data element into the model, and acquiring second intermediate data based on at least one of a second vector and a second activation included in the one or more neurons, for each of the plurality of neuron layers;
• and
• calculating a similarity between the first intermediate data and the second intermediate data, for the plurality of neuron layers.