LEARNING SYSTEM, LEARNING SERVER APPARATUS, PROCESSING APPARATUS, LEARNING METHOD, AND PROGRAM
Provided are a learning system and the like that significantly shorten a search time of NAS and enable machine learning in a practical time. The learning system includes a learning server apparatus and n processing apparatuses i. The processing apparatus i includes a score calculation unit that calculates a score s ir when local data d i is applied to each of A neural networks r. The learning server apparatus includes an aggregation unit that aggregates A neural networks using A×n scores s ir and selects an optimal neural network. The first federated learning unit and the second federated learning units of the n processing apparatuses i cooperate to perform federated learning using the selected optimal neural network as a first global model. The score s ir includes an index with which a neural network having an excellent learning effect can be searched for.
1 . A learning system comprising:
a learning server apparatus; and
n processing apparatuses i,
wherein, when i=1, 2, . . . , n, and r=0, 1, . . . , A−1,
the processing apparatuses i each include:
second processing circuitry configured to:
execute a second federated learning processing; and
execute a score calculation processing in which the second processing circuitry calculates a score s ir when local data d i is applied to each of A neural networks r,
the learning server apparatus includes:
first processing circuitry configured to:
execute a first federated learning processing; and
execute an aggregation processing in which the first processing circuitry aggregates A neural networks using A×n scores s ir , and selects an optimal neural network,
in the first federated learning processing and the second federated learning processes, the first processing circuitry and the second processing circuitries of the n processing apparatuses i cooperate to perform federated learning using the selected optimal neural network as a first global model, and
the score s ir includes an index with which a neural network having an excellent learning effect can be searched for.
2 . The learning system according to claim 1 , wherein
the score s ir is a correlation score of a weight when the local data d i is applied to the neural network r, and
the aggregation processing in which the first processing circuitry calculates a variation of the score s ir for each neural network r, calculates a score S r for each neural network r in consideration of a number of pieces of data of the local data d i in a case where the variation is larger than a predetermined threshold value, and calculates the score S r for each neural network r without considering the number of pieces of data of the local data d i in a case where the variation is equal to or smaller than the predetermined threshold value.
3 . The learning system according to claim 1 , wherein
in the aggregation processing the first processing circuitry selects Q optimal neural network possibilities in a case where an optimal neural network cannot be selected, and
the first processing circuitry divides the n processing apparatuses i into Q groups, performs federated learning in cooperation with a second processing circuitry of a processing apparatus belonging to each group using the Q optimal neural network possibilities as a first global model, compares accuracies of the Q optimal neural network possibilities after the federated learning, and selects an optimal neural network with the highest accuracy as the optimal neural network.
4 . A learning server apparatus of the learning system according to claim 1 .
5 . A processing apparatus of the learning system according to claim 1 .
6 . A learning method using a learning server apparatus that includes first processing circuitry and n processing apparatuses i that include second processing circuitry, the learning method comprising:
when i=1, 2 . . . , n, and r=0, 1, . . . , A−1,
a score calculation step of calculating, by the second processing circuitry of the processing apparatus i, a score s ir when local data d i is applied to each of A neural networks r;
an aggregation step of aggregating, by the first processing circuitry of the learning server apparatus, A neural networks using A×n scores s ir , and selecting an optimal neural network; and
a federated learning step of cooperating, by the first processing circuitry of the learning server apparatus and the second processing circuitries of the n processing apparatuses i, to perform federated learning using the selected optimal neural network as a first global model,
wherein the score s ir includes an index with which a neural network having an excellent learning effect can be searched for.
7 . A non-transitory computer-readable recording medium having recorded thereon a program for causing a computer to function as the learning server apparatus according to claim 4 .
8 . A non-transitory computer-readable recording medium having recorded thereon a program for causing a computer to function as the processing apparatus according to claim 5 .