# US 12,602,919 B2 · Generative data augmentation with task loss guided fine-tuning

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

- Number: US 12,602,919 B2
- Document ID: US12602919B2
- Status: Granted
- Kind: B2
- Application number: 18/506,018
- Filed: 2023-11-09
- Published: 2025-05-15
- Granted: 2026-04-14
- Earliest claimed: 2023-11-09
- Assignees: QUALCOMM Incorporated
- Applicants: QUALCOMM Incorporated
- Inventors: Davide Abati; Amirhossein Habibian; Auke Joris Wiggers; Jens Petersen
- Web page: https://patentgrep.com/patent/US12602919B2

## CPC

- G06N 3/045: Combinations of networks
- G06V 10/82: using neural networks
- G06N 3/0475: Generative networks
- G06V 10/774: Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A method includes generating a synthetic dataset with a generative model. The method also includes tuning the generative model based on feedback from a task network that receives the synthetic dataset as input. The task network may perform image recognition. The synthetic dataset may be generated based on a set of classes and labels of the classes. The method may iteratively generate the synthetic dataset and tune the generative model, based on feedback from the task network.

## Claims

### Claim 1

An apparatus, comprising:
at least one memory storing instructions; and
at least one processor coupled to the at least one memory and configured to instruct the apparatus to:
generate a synthetic dataset with a generative model;
obtain a prediction, with a task network, based on the synthetic dataset;
calculate a loss of the task network based on the prediction from the synthetic dataset;
back propagate the loss of the task network to the generative model instead of a task model;
tune the generative model by reducing the loss of the task network to obtain a tuned generative model; and
generate an updated synthetic dataset with the tuned generative model.

### Claim 2 (depends on 1)

The apparatus of claim 1, in which the task network performs image recognition.

### Claim 3 (depends on 1)

The apparatus of claim 1, in which the at least one processor is further configured to instruct the apparatus to iteratively generate the synthetic dataset and tune the generative model, based on feedback from the task network.

### Claim 4 (depends on 1)

The apparatus of claim 1, in which the synthetic dataset is generated based on a set of classes and labels of the classes.

### Claim 5 (depends on 1)

The apparatus of claim 1, in which the at least one processor is further configured to instruct the apparatus to retrain the task network based on the updated synthetic dataset.

### Claim 6

A method, comprising:
generating a synthetic dataset with a generative model based on a set of real images;
obtaining a first prediction, with a task network, based on the synthetic dataset;
calculating a first loss of the task network based on the first prediction from the synthetic dataset;
updating the task network based on the first loss of the task network to obtain an updated task network;
calculating a second loss of the updated task network based on a second prediction of the updated task network and the set of real images;
back propagating the second loss of the updated task network to the generative model;
tuning the generative model by reducing the loss of the updated task network to obtain a tuned generative model; and
generating an updated synthetic dataset with the tuned generative model.

### Claim 7 (depends on 6)

The method of claim 6, in which the task network performs image recognition.

### Claim 8 (depends on 6)

The method of claim 6, further comprising iteratively generating the synthetic dataset and tuning the generative model, based on feedback from the task network.

### Claim 9

A non-transitory computer-readable medium having program code recorded thereon, the program code executed by a processor and comprising:
program code to generate a synthetic dataset with a generative model;
program code to obtain a prediction, with a task network, based on the synthetic dataset;
program code to calculate a loss of the task network based on the prediction from the synthetic dataset;
program code to back propagate the loss of the task network to the generative model instead of a task model;
program code to tune the generative model by reducing the loss of the task network to obtain a tuned generative model; and
program code to generate an updated synthetic dataset with the tuned generative model.

### Claim 10 (depends on 9)

The non-transitory computer-readable medium of claim 9, in which the task network performs image recognition.

### Claim 11 (depends on 9)

The non-transitory computer-readable medium of claim 9, in which the program code comprises program code to iteratively generate the synthetic dataset and tune the generative model, based on feedback from the task network.

### Claim 12 (depends on 9)

The non-transitory computer-readable medium of claim 9, in which the synthetic dataset is generated based on a set of classes and labels of the classes.

### Claim 13 (depends on 9)

The non-transitory computer-readable medium of claim 9, in which the program code comprises program code to retrain the task network based on the updated synthetic dataset.

### Claim 14

An apparatus, comprising:
means for generating a synthetic dataset with a generative model based on a set of real images;
means for obtaining a first prediction, with the task network, based on the synthetic dataset;
means for calculating a first loss of the task network based on the first prediction from the synthetic dataset;
means for updating the task network based on the first loss of the task network to obtain an updated task network;
means for calculating a second loss of the updated task network based on a second prediction of the updated task network and the set of real images;
means for back propagating the second loss of the updated task network to the generative model;
means for tuning the generative model by reducing the loss of the updated task network to obtain a tuned generative model; and
means for generating an updated synthetic dataset with the tuned generative model.

### Claim 15 (depends on 14)

The apparatus of claim 14, in which the task network performs image recognition.

### Claim 16 (depends on 14)

The apparatus of claim 14, further comprising means for iteratively generating the synthetic dataset and means for tuning the generative model, based on feedback from the task network.