# US 2025/0094829 A1 · Methods and mechanisms for trace-based transfer learning

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0094829 A1
- Document ID: US20250094829A1
- Status: Published application
- Kind: A1
- Application number: 18/368,311
- Filed: 2023-09-14
- Published: 2025-03-20
- Earliest claimed: 2023-09-14
- Assignees: Applied Materials, Inc.
- Applicants: Applied Materials, Inc.
- Inventors: Fei Li; Jimmy Iskandar; James Robert Moyne
- Web page: https://patentgrep.com/patent/US20250094829A1

## CPC

- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06N 20/00: Machine learning
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An electronic device manufacturing system configured identify a machine-learning model trained to generate analytic or predictive data for a first substrate processing domain associated with a type of substrate processing system. The system is further configured to obtain first trace data pertaining to the first domain used to train the machine-learning model. The system is further configured to a transfer model for a second substrate processing domain associated with the type of substrate processing system. The transfer model is generated based on the first trace data pertaining to the first substrate processing domain and second trace data pertaining to the second substrate processing domain. Using the transfer model, at least one of the machine-learning model or current trace data associated with the second substrate processing domain is modified to enable the machine-learning model to generate analytic or predictive data associated with the second substrate processing domain.

## Claims

### Claim 1

A method, comprising:
identifying a machine-learning model trained to generate analytic or predictive data for a first substrate processing domain associated with a type of substrate processing system;
obtaining first trace data pertaining to the first substrate processing domain, the first trace data used to train the machine-learning model;
generating a transfer model for a second substrate processing domain associated with the type of substrate processing system, wherein the transfer model is generated based on the first trace data pertaining to the first substrate processing domain and second trace data pertaining to the second substrate processing domain; and
modifying, using the transfer model, at least one of the machine-learning model or current trace data associated with the second substrate processing domain to enable the machine-learning model to generate analytic or predictive data associated with the second substrate processing domain.

### Claim 2 (depends on 1)

The method of claim 1, wherein the first substrate processing domain comprises a first process chamber and the second substrate processing domain comprises a second process chamber, wherein the first process chamber and the second process chamber are a same type of process chamber.

### Claim 3 (depends on 1)

The method of claim 1, wherein the first substrate processing domain comprises a first process recipe and the second substrate processing domain comprise a second process recipe.

### Claim 4 (depends on 1)

The method of claim 1, wherein the first trace data comprises a first set of traces associated with the first substrate processing domain and the second trace data comprises a second set of traces associated with the second substrate processing domain.

### Claim 5 (depends on 4)

The method of claim 4, further comprising:
generating, from the first set of traces, a first fundamental trace; and
generating, from the second set of traces, a second fundamental trace.

### Claim 6 (depends on 5)

The method of claim 5, further comprising:
generating, based on the first fundamental trace and the second fundamental trace, a transfer map reflecting a relationship between the first fundamental trace and the second fundamental trace.

### Claim 7 (depends on 6)

The method of claim 6, where the transfer map provides feature-based scaling in reflecting the relationship between the first fundamental trace and second fundamental trace.

### Claim 8 (depends on 6)

The method of claim 6, wherein the transfer map is used to generate the transfer model.

### Claim 9 (depends on 1)

The method of claim 1, further comprising:
providing, as input to the transfer model, current trace data pertaining to the second substrate processing domain;
obtaining one or more first output values of the transfer model;
providing, as input to the machine-learning model, the one or more first output values; and
obtaining one or more second output values of the machine learning model, the one or more second output values reflecting the analytic or predictive data associated with the second substrate processing domain.

### Claim 10 (depends on 9)

The method of claim 9, further comprising:
performing a corrective action based on the one or more second output values of the machine-learning model.

### Claim 11 (depends on 1)

The method of claim 1, further comprising:
retraining the machine-learning model using the transfer model;
providing, as input to the retrained machine-learning model, current trace data pertaining to the second substrate processing domain; and
obtaining one or more output values of the retrained machine-learning model reflecting the analytic or predictive data associated with the second substrate processing domain.

### Claim 12 (depends on 11)

The method of claim 11, further comprising:
performing a corrective action based on the one or more output values of the machine-learning model.

### Claim 13

A system, comprising:
a memory device; and
a processing device, operatively coupled to the memory device, to perform operations comprising:
identifying a machine-learning model trained to generate analytic or predictive data for a first substrate processing domain associated with a type of substrate processing system;
obtaining first trace data pertaining to the first substrate processing domain, the first trace data used to train the machine-learning model;
generating a transfer model for a second substrate processing domain associated with the type of substrate processing system, wherein the transfer model is generated based on the first trace data pertaining to the first substrate processing domain and second trace data pertaining to the second substrate processing domain; and
modifying, using the transfer model, at least one of the machine-learning model or current trace data associated with the second substrate processing domain to enable the machine-learning model to generate analytic or predictive data associated with the second substrate processing domain.

### Claim 14 (depends on 13)

The system of claim 13, wherein the first trace data comprises a first set of traces associated with the first substrate processing domain and the second trace data comprises a second set of traces associated with the second substrate processing domain.

### Claim 15 (depends on 14)

The system of claim 14, wherein the operations further comprise:
generating, from the first set of traces, a first fundamental trace; and
generating, from the second set of traces, a second fundamental trace.

### Claim 16 (depends on 15)

The system of claim 15, wherein the operations further comprise:
generating, based on the first fundamental trace and the second fundamental trace, a transfer map reflecting a relationship between the first fundamental trace and the second fundamental trace.

### Claim 17 (depends on 13)

The system of claim 13, wherein the operations further comprise:
providing, as input to the transfer model, current trace data pertaining to the second substrate processing domain;
obtaining one or more first output values of the transfer model;
providing, as input to the machine-learning model, the one or more first output values; and
obtaining one or more second output values of the machine learning model, the one or more second output values reflecting the analytic or predictive data associated with the second substrate processing domain.

### Claim 18 (depends on 11)

The system of claim 11, wherein the operations further comprise:
retraining the machine-learning model using the transfer model;
providing, as input to the retrained machine-learning model, current trace data pertaining to the second substrate processing domain; and
obtaining one or more output values of the retrained machine-learning model reflecting the analytic or predictive data associated with the second substrate processing domain.

### Claim 19 (depends on 18)

The system of claim 18, wherein the operations further comprise:
performing a corrective action based on the one or more output values.

### Claim 20

A method, comprising:
providing, as input to a transfer model, current trace data associated with a target substrate processing domain, wherein the transfer model is generated based on historical trace data associated with the target substrate processing domain and historical trace data associated with a source substrate processing domain, wherein the source substrate processing domain and the target substrate processing domain are both associated with a type of substrate processing system;
obtaining one or more first output values of the transfer model reflective of the current trace data modified by a set of offset values;
providing, as input to a machine-learning model trained to generate analytic or predictive data for the source substrate processing domain, the one or more first output values from the transfer model; and
obtaining one or more second output values of the machine learning model, the one or more second output values representing analytic or predictive data associated with the target substrate processing domain.

### Claim 21 (depends on 20)

The method of claim 20, further comprising:
performing a corrective action based on the one or more second output values of the machine-learning model.

### Claim 22

A method, comprising:
retraining a machine-learning model using a transfer model, wherein the transfer model is generated based on historical trace data associated with a target substrate processing domain and historical trace data associated with a source substrate processing domain, wherein the source substrate processing domain and the target substrate processing domain are associated with a type of substrate processing system, wherein the machine-learning model is trained to generate analytic or predictive data for the source substrate processing domain;
providing, as input to the retrained machine-learning model, current trace data pertaining to the target substrate processing domain; and
obtaining one or more output values of the retrained machine-learning model reflecting analytic or predictive data associated with the target substrate processing domain.

### Claim 23 (depends on 22)

The method of claim 22, further comprising:
performing a corrective action based on the one or more second output values of the modified machine-learning model.
