System and method for substrate processing using models at multiple layers

1 . A model management system comprising:

a first management device configured to manage a model at a predetermined layer among layers, the layers being three or more layers;

one or more second management devices configured to manage one or more models at a layer one level lower than the predetermined layer; and

one or more executing devices,

wherein the first management device includes:

a processor; and

a memory storing program instructions that cause the processor to:

when one or more model parameters of the one or more models managed by the one or more second management devices are updated, calculate a new model parameter based on each of the updated one or more model parameters, and

perform control so that the new model parameter is set in a plurality of models respectively managed by a plurality of management devices at a lowest layer, the plurality of management devices belonging to the first management device,

transmit the new model parameter to a management device that manages a model at a layer one level higher than the model managed by the first management device, and

wherein the model management system is configured to separately manage models applied to a substrate manufacturing process in the layers,

wherein the one or more executing devices collect data measured while a substrate is processed by a substrate processing device and execute a model among the plurality of models based on the collected data to output an inference result, and

wherein the one or more executing devices transmit, to the substrate processing device, an instruction based on the output inference result, the instruction causing the substrate processing device to perform the substrate manufacturing process on the substrate.

2 . The model management system as claimed in claim 1 , wherein the layers are separated according to a device model of the substrate processing device that performs the substrate manufacturing process, a chamber type of the substrate processing device that performs the substrate manufacturing process, or a process group of the substrate processing device that performs the substrate manufacturing process.

3 . The model management system as claimed in claim 1 , wherein the models include a model that outputs an inference result of fault detection of the substrate processing device that performs the substrate manufacturing process.

4 . The model management system as claimed in claim 1 , wherein the processor calculates the new model parameter when a predetermined condition is satisfied, and the processor performs the control so that the new model parameter is set in the plurality of models at the lowest layer at predetermined timings.

5 . The model management system as claimed in claim 1 , wherein the processor performs the control so that, when a new device that performs the substrate manufacturing process is added, the new model parameter is set in a model corresponding to the added device.

6 . The model management system as claimed in claim 1 , wherein the one or more model parameters of the one or more models managed by the one or more second management devices are updated based on data collected in the substrate processing device.

7 . The model management system as claimed in claim 1 , wherein the inference result includes an inference result of fault detection.

8 . The model management system as claimed in claim 7 , wherein the inference result of the object obtained as the result of the process treatment includes an inference of a film formation rate, an inference of an etching rate, an inference of a shape of the object.

9 . A model management method of a model management system, comprising:

managing, by a first management device, a model at a predetermined layer among layers, the layers being three or more layers;

managing, by one or more second management devices, one or more models at a layer one level lower than the predetermined layer,

wherein the managing of the model at the predetermined layer includes:

when one or more model parameters of the one or more models are updated, calculating a new model parameter based on each of the updated one or more model parameters, and

performing control so that the new model parameter is set in a plurality of models respectively managed by a plurality of management devices at a lowest layer, the plurality of management devices belonging to the first management device,

transmitting the new model parameter to a management device that manages a model at a layer one level higher than the model managed by the first management device, and

wherein the model management system is configured to separately manage models applied to a substrate manufacturing process in the layers,

wherein the model management method further comprises:

collecting, by one or more executing devices, data measured while a substrate is processed by a substrate processing device and executing, by the one or more executing devices, a model among the plurality of models based on the collected data to output an inference result; and

transmitting, by the one or more executing devices, to the substrate processing device, an instruction based on the output inference result, the instruction causing the substrate processing device to perform the substrate manufacturing process on the substrate.

10 . A non-transitory computer-readable recording medium having stored therein a model management program causing a computer of a model management system to perform a process comprising:

managing, by a first management device, a model at a predetermined layer among layers, the layers being three or more layers;

managing, by one or more second management devices, one or more models at a layer one level lower than the predetermined layer,

wherein the managing of the model at the predetermined layer includes:

when one or more model parameters of the one or more models are updated, calculating a new model parameter based on each of the updated one or more model parameters, and

performing control so that the new model parameter is set in a plurality of models respectively managed by a plurality of management devices at a lowest layer, the plurality of management devices belonging to the first management device,

transmitting the new model parameter to a management device that manages a model at a layer one level higher than the model managed by the first management device, and

wherein the model management system is configured to separately manage models applied to a substrate manufacturing process in the layers,

wherein the process further comprises:

collecting, by one or more executing devices, data measured while a substrate is processed by a substrate processing device and executing, by the one or more executing devices, a model among the plurality of models based on the collected data to output an inference result; and

transmitting, by the one or more executing devices, to the substrate processing device, an instruction based on the output inference result, the instruction causing the substrate processing device to perform the substrate manufacturing process on the substrate.