# US 2023/0056869 A1 · Method of generating deep learning model and computing device performing the same > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2023/0056869 A1 - Document ID: US20230056869A1 - Status: Published application - Kind: A1 - Application number: 17/689,115 - Filed: 2022-03-08 - Published: 2023-02-23 - Earliest claimed: 2021-08-19 - Assignees: Samsung Electronics Co., Ltd. - Applicants: Samsung Electronics Co., Ltd. - Inventors: Gwangnae GIL; Seyoung PARK; Sola WOO; Jonghyun LEE - Web page: https://patentgrep.com/patent/US20230056869A1 ## CPC - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06F 30/27: using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model - G06F 30/367: Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods - G06N 3/0464: Convolutional networks [CNN, ConvNet] - G06N 3/047: Probabilistic or stochastic networks - G06N 3/0475: Generative networks - G06N 3/048: Activation functions - G06N 3/063: using electronic means - G06F 2119/06: Power analysis or power optimisation - G06F 2119/08: Thermal analysis or thermal optimisation - G06N 3/082: modifying the architecture, e.g. adding, deleting or silencing nodes or connections - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) ## Abstract To generate a deep learning model, basic training data corresponding to a combination of device data and simulation result data is generated using a compact model that generates the simulation result data indicating characteristics of a semiconductor device corresponding to the device data by performing simulation based on the device data. A deep learning model is trained based on the basic training data such that the deep learning model outputs prediction data indicating the characteristics of the semiconductor device and uncertainty data indicating uncertainty of the prediction data. The deep learning model is retrained based on the uncertainty data. The deep learning model may precisely predict the characteristics of the semiconductor device by training the deep learning model to output the prediction data and the uncertainty data and retraining the deep learning model based on the uncertainty data. ## Claims ### Claim 1 A method of generating a deep learning model, the method being performed by executing program codes by at least one processor, the program codes being stored in computer readable media, the method comprising: generating basic training data corresponding to a combination of device data and simulation result data using a compact model, the compact model configured to generate the simulation result data by performing a simulation based on the device data, the simulation result data indicating characteristics of a semiconductor device corresponding to the device data; training the deep learning model based on the basic training data such that the deep learning model is configured to output prediction data and uncertainty data, the prediction data indicating the characteristics of the semiconductor device, and the uncertainty data indicating uncertainty of the prediction data; and retraining the deep learning model based on the uncertainty data. ### Claim 2 (depends on 1) The method of claim 1, wherein the uncertainty data include a model uncertainty value indicating the uncertainty of the prediction data caused by insufficiency of the basic training data. ### Claim 3 (depends on 2) The method of claim 2, wherein the retraining the deep learning model includes: comparing the model uncertainty value with a model reference value; generating addition training data using the compact model when the model uncertainty value is larger than the model reference value; and retrain the deep learning model based on the addition training data, wherein the addition training data is different from the basic training data. ### Claim 4 (depends on 3) The method of claim 3, wherein retraining the deep learning model further includes: determining an addition data range corresponding to a range of data such that the model uncertainty value is larger than the model reference value. ### Claim 5 (depends on 4) The method of claim 4, wherein the addition training data correspond to a combination of the device data included in the addition data range and the simulation result data. ### Claim 6 (depends on 3) The method of claim 3, wherein the deep learning model that has been trained based on the basic training data is further trained based on the addition training data. ### Claim 7 (depends on 1) The method of claim 1, wherein the uncertainty data include a data uncertainty value, the data uncertainty value indicating the uncertainty of the prediction data caused by noises in the basic training data. ### Claim 8 (depends on 7) The method of claim 7, wherein retraining the deep learning model includes: comparing the data uncertainty value with a data reference value; providing measurement data by measuring the characteristics of the semiconductor device, when the data uncertainty value is larger than the data reference value; correcting the compact model based on the measurement data; generating updated training data using the corrected compact model; and retraining the deep learning model based on the updated training data. ### Claim 9 (depends on 8) The method of claim 8, wherein retraining the deep learning model further includes: determining a measurement data range corresponding to a range of data such that the data uncertainty value is larger than the data reference value. ### Claim 10 (depends on 9) The method of claim 9, wherein the characteristics of the semiconductor device correspond to the device data included in the measurement data range. ### Claim 11 (depends on 8) The method of claim 8, wherein the deep learning model that has been trained based on the basic training data is initialized, and the initialized deep learning model is trained based on the measurement data. ### Claim 12 (depends on 1) The method of claim 1, wherein the uncertainty data include a model uncertainty value and a data uncertainty value, the model uncertainty value indicating the uncertainty of the prediction data caused by insufficiency of the basic training data, and the data uncertainty value indicating the uncertainty of the prediction data caused by noises of the basic training data. ### Claim 13 (depends on 12) The method of claim 12, wherein retraining the deep learning model includes at least one of: performing a first retraining such that the deep learning model that has been trained based on the basic training data is further trained based on the model uncertainty value; or performing a second retraining such that the deep learning model that has been trained based on the basic training data is initialized, and the initialized deep learning model is trained based on the data uncertainty value. ### Claim 14 (depends on 13) The method of claim 13, wherein whether to perform the first retraining is determined based on the model uncertainty value, and when it is determined that the first retraining is not performed, whether to perform the second retraining is determined based on the data uncertainty value. ### Claim 15 (depends on 13) The method of claim 13, wherein whether to perform the first retraining and whether to perform the second retraining are determined independently of each other. ### Claim 16 (depends on 1) The method of claim 1, wherein the deep learning model includes a Bayesian Neural Network (BNN). ### Claim 17 (depends on 1) The method of claim 1, wherein the device data indicate structure and operation condition of the semiconductor device, the simulation result data and the prediction data indicate electrical characteristics of the semiconductor device, and the device data is included in input data of the deep learning model. ### Claim 18 (depends on 17) The method of claim 17, wherein the input data of the deep learning model further includes process data indicating a condition of manufacturing process of the semiconductor device. ### Claim 19 A method of generating a deep learning model, the method being performed by executing program codes by at least one processor, the program codes being stored in computer readable media, the method comprising: generating basic training data corresponding to a combination of device data and simulation result data using a compact model, the compact model configured to generate the simulation result data by performing a simulation based on the device data, the simulation result data indicating characteristics of a semiconductor device corresponding to the device data; training the deep learning model based on the basic training data such that the deep learning model is configured to output prediction data, a model uncertainty value, and a data uncertainty value, the prediction data indicating the characteristics of the semiconductor device, the model uncertainty value indicating the uncertainty of the prediction data caused by insufficiency of the basic training data, and the data uncertainty value indicating the uncertainty of the prediction data caused by noises of the basic training data; performing a first retraining such that the deep learning model that has been trained based on the basic training data is further trained based on the model uncertainty value; and performing a second retraining such that the deep learning model that has been trained based on the basic training data is initialized, and the initialized deep learning model is trained based on the data uncertainty value. ### Claim 20 A computing device comprising: at least one processor; and a computer readable medium storing program codes and a compact model, the program codes being executed by the at least one processor to generate a deep learning model, the compact model generating simulation result data indicating characteristics of a semiconductor device corresponding to device data by performing simulation based on the device data, the at least one processor executing the program codes to: generate basic training data corresponding to a combination of the device data and the simulation result data using the compact model; train the deep learning model based on the basic training data such that the deep learning model is configured to output prediction data and uncertainty data, the prediction data indicating the characteristics of the semiconductor device and the uncertainty data indicating uncertainty of the prediction data; and retrain the deep learning model based on the uncertainty data.