# US 2021/0174191 A1 · Automated fine-tuning of a pre-trained neural network for transfer learning

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

- Number: US 2021/0174191 A1
- Document ID: US20210174191A1
- Status: Published application
- Kind: A1
- Application number: 16/704,804
- Filed: 2019-12-05
- Published: 2021-06-10
- Earliest claimed: 2019-12-05
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Parijat Dube; Bishwaranjan Bhattacharjee; Patrick Watson; John Ronald Kender
- Web page: https://patentgrep.com/patent/US20210174191A1

## CPC

- G06N 3/08: Learning methods
- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/096: Transfer learning
- G06N 3/0985: Hyperparameter optimisation; Meta-learning; Learning-to-learn
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06N 3/045: Combinations of networks
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

In an embodiment, a method for fine-tuning a pre-trained neural network for transfer learning, the method comprising obtaining a first target feature vector from a first layer of a pre-trained neural network responsive to a first target data element of a target dataset passing therethrough, obtaining a first source feature vector associated with the first layer of the pre-trained neural network, calculating a first divergence value for the first layer of the pre-trained neural network based at least in part on the first target feature vector and the first source feature vector, and setting a learning rate for the first layer of the pre-trained neural network based at least in part on the first divergence value.

## Claims

### Claim 1

A computer implemented method comprising:
obtaining, by a processor, a first target feature vector from a first layer of a pre-trained neural network responsive to a first target data element of a target dataset passing therethrough;
obtaining, by the processor, a first source feature vector associated with the first layer of the pre-trained neural network;
calculating, by the processor, a first divergence value for the first layer of the pre-trained neural network based at least in part on the first target feature vector and the first source feature vector; and
setting, by the processor, a learning rate for the first layer of the pre-trained neural network based at least in part on the first divergence value.

### Claim 2 (depends on 1)

The computer implemented method of claim 1, further comprising obtaining, by the processor, a second target feature vector from the first layer of the pre-trained neural network responsive to a second target data element passing therethrough.

### Claim 3 (depends on 2)

The computer implemented method of claim 2, further comprising calculating, by the processor, a first target average feature vector based at least in part on the first and second target feature vectors.

### Claim 4 (depends on 2)

The computer implemented method of claim 2, further comprising obtaining, by the processor, a second source feature vector associated with the first layer of the pre-trained neural network.

### Claim 5 (depends on 4)

The computer implemented method of claim 4, further comprising:
calculating, by the processor, a first target average feature vector based at least in part on the first and second target feature vectors; and
calculating, by the processor, a first source average feature vector based at least in part on the first and second source feature vectors.

### Claim 6 (depends on 4)

The computer implemented method of claim 4, further comprising:
obtaining, by the processor, a third target feature vector from a second layer of the pre-trained neural network responsive to the first target data element passing therethrough; and
obtaining, by the processor, a fourth target feature vector from the second layer of the pre-trained neural network responsive to the second target data element passing therethrough.

### Claim 7 (depends on 6)

The computer implemented method of claim 6, further comprising:
calculating, by the processor, a first target average feature vector based at least in part on the first and second target feature vectors;
calculating, by the processor, a first source average feature vector based at least in part on the first and second source feature vectors; and
calculating, by the processor, a second target average feature vector based at least in part on the third and fourth target feature vectors.

### Claim 8 (depends on 6)

The computer implemented method of claim 6, further comprising:
obtaining, by the processor, a third source feature vector associated with the second layer of the pre-trained neural network; and
obtaining, by the processor, a fourth source feature vector associated with the second layer of the pre-trained neural network.

### Claim 9 (depends on 8)

The computer implemented method of claim 8, further comprising:
calculating, by the processor, a first target average feature vector based at least in part on the first and second target feature vectors;
calculating, by the processor, a first source average feature vector based at least in part on the first and second source feature vectors;
calculating, by the processor, a second target average feature vector based at least in part on the third and fourth target feature vectors; and
calculating, by the processor, a second source average feature vector based at least in part on the third and fourth source feature vectors.

### Claim 10 (depends on 9)

The computer implemented method of claim 9, further comprising:
calculating, by the processor, first and second normalized target average feature vectors; and
calculating, by the processor, first and second normalized source average feature vectors.

### Claim 11 (depends on 10)

The computer implemented method of claim 10, wherein the generating of the first divergence value comprises calculating a divergence between the first normalized target average feature vector and the first normalized source average feature vector.

### Claim 12 (depends on 11)

The computer implemented method of claim 11, wherein the setting of the learning rate for the first layer of the pre-trained neural network comprises mapping the first divergence value to the learning rate value based at least in part on divergence values associated with other layers of the pre-trained neural network.

### Claim 13 (depends on 12)

The computer implemented method of claim 12, wherein the obtaining of the first, second, third, and fourth source feature vector includes obtaining the first and second source feature vectors from the first layer of the pre-trained neural network responsive to respective first and second source data elements of a source dataset passing therethrough, and includes obtaining the third and fourth source feature vectors from the second layer of the pre-trained neural network responsive to the respective first and second source data elements of the source dataset passing therethrough.

### Claim 14 (depends on 12)

The computer implemented method of claim 12, wherein the obtaining of the first, second, third, and fourth source feature vectors includes obtaining the first, second, third, and fourth source feature vectors from memory.

### Claim 15

A computer program product for generating parameters of a neural network, the computer program product comprising one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable by controller circuitry to cause the controller circuitry to perform operations comprising:
obtaining a first target feature vector from a first layer of a pre-trained neural network responsive to a first target data element of a target dataset passing therethrough;
obtaining a first source feature vector associated with the first layer of the pre-trained neural network;
calculating a first divergence value for the first layer of the pre-trained neural network based at least in part on the first target feature vector and the first source feature vector; and
setting a learning rate for the first layer of the pre-trained neural network based at least in part on the first divergence value.

### Claim 16 (depends on 15)

The computer usable program product of claim 15, wherein the stored program instructions are stored in a computer readable storage medium in a data processing system, and wherein the stored program instructions are transferred over a network from a remote data processing system.

### Claim 17 (depends on 15)

The computer usable program product of claim 15, wherein the stored program instructions are stored in a computer readable storage medium in a server data processing system, and wherein the stored program instructions are downloaded over a network to a remote data processing system for use in a computer readable storage medium associated with the remote data processing system, further comprising:
program instructions to meter use of the computer usable code associated with the request; and
program instructions to generate an invoice based on the metered use.

### Claim 18 (depends on 15)

A computer usable program product of claim 15, wherein the obtaining of the first source feature vector includes obtaining the first source feature vector from the first layer of the pre-trained neural network responsive to the first source data element of a source dataset passing therethrough.

### Claim 19 (depends on 15)

A computer usable program product of claim 15, wherein the obtaining of the first source feature vector includes obtaining the first source feature vector from memory.

### Claim 20

A computer system comprising a processor and one or more computer readable storage media, and program instructions collectively stored on the one or more computer readable storage media, the program instructions executable by the processor to cause the processor to perform operations comprising:
obtaining a first target feature vector from a first layer of a pre-trained neural network responsive to a first target data element of a target dataset passing therethrough;
obtaining a first source feature vector associated with the first layer of the pre-trained neural network;
calculating a first divergence value for the first layer of the pre-trained neural network based at least in part on the first target feature vector and the first source feature vector; and
setting a learning rate for the first layer of the pre-trained neural network based at least in part on the first divergence value.
