Machine learning method and computing system
An information processing apparatus calculates a value based on the lengths of wires included in inner layers selecting the inner layers other than outermost layers and layers adjacent to the outermost layers from among a plurality of layers included in circuit data. The information processing apparatus generates training data including first layer data corresponding to the patterns of the outermost layers, second layer data corresponding to the patterns of the layers adjacent to the outermost layers, and the value based on the lengths of the wires. The information processing apparatus trains a machine learning model by using the training data.
1. A non-transitory computer-readable storage medium storing a machine learning program executable by one or more computers, the machine learning program comprising:
an instruction for calculating a length of one or more wires included in one or more inner layers by selecting the one or more inner layers other than outermost layers and layers adjacent to the outermost layers from among a plurality of layers included in circuit data;
an instruction for generating training data including first layer data corresponding to patterns of the outermost layers, second layer data corresponding to patterns of the layers adjacent to the outermost layers, and the length; and
an instruction for training a machine learning model by using the training data, the machine learning model being for estimating electromagnetic waves radiated from an electronic circuit.
2. The non-transitory computer-readable storage medium according to claim 1 , wherein the training data includes a tensor including first and second channels corresponding to the outermost layers, third and fourth channels corresponding to the layers adjacent to the outermost layers, and a fifth channel indicating the length.
3. The non-transitory computer-readable storage medium according to claim 1 , wherein a size of an input to the machine learning model is fixed without respect to a number of layers included in the one or more inner layers.
4. The non-transitory computer-readable storage medium according to claim 1 , wherein
the outermost layers include a wire, and
each of the layers adjacent to the outermost layers is either a power supply layer or a ground layer.
5. A computer-implemented machine learning method comprising:
calculating a length of one or more wires included in one or more inner layers by selecting the one or more inner layers other than outermost layers and layers adjacent to the outermost layers from among a plurality of layers included in circuit data;
generating training data including first layer data corresponding to patterns of the outermost layers, second layer data corresponding to patterns of the layers adjacent to the outermost layers, and the length; and
training a machine learning model by using the training data, the machine learning model being for estimating electromagnetic waves radiated from an electronic circuit.
6. The computer-implemented machine learning method according to claim 5 , wherein the training data includes a tensor including first and second channels corresponding to the outermost layers, third and fourth channels corresponding to the layers adjacent to the outermost layers, and a fifth channel indicating the length.
7. The computer-implemented machine learning method according to claim 5 , wherein a size of an input to the machine learning model is fixed without respect to a number of layers included in the one or more inner layers.
8. The computer-implemented machine learning method according to claim 5 , wherein
the outermost layers include a wire, and
each of the layers adjacent to the outermost layers is either a power supply layer or a ground layer.
9. A computing system comprising:
a memory that stores therein circuit data including a plurality of layers; and
a processor coupled to the memory, the processor being configured to
calculate a length of one or more wires included in one or more inner layers by selecting the one or more inner layers other than outermost layers and layers adjacent to the outermost layers from among the plurality of layers,
generate training data including first layer data corresponding to patterns of the outermost layers, second layer data corresponding to patterns of the layers adjacent to the outermost layers, and the length, and
train a machine learning model by using the training data, the machine learning model being for estimating electromagnetic waves radiated from an electronic circuit.