# US 2026/0065131 A1 · Ensuring fairness in a generative ai model via model pruning

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

- Number: US 2026/0065131 A1
- Document ID: US20260065131A1
- Status: Published application
- Kind: A1
- Application number: 18/821,151
- Filed: 2024-08-30
- Published: 2026-03-05
- Earliest claimed: 2024-08-30
- Assignees: Cisco Technology, Inc.
- Applicants: Cisco Technology, Inc.
- Inventors: Yuguang Yao; Akshay Jajoo; Gaowen Liu; Yihua Zhang; Ramana Rao V.R. Kompella; Charles Fleming; Myungjin Lee
- Web page: https://patentgrep.com/patent/US20260065131A1

## CPC

- G06N 20/00: Machine learning
- G06N 3/045: Combinations of networks
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06T 11/00: Two-dimensional [2D] image generation
- G06T 2211/441: AI-based methods, deep learning or artificial neural networks
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

In one implementation, a device obtains one or more terms of interest. The device also obtains one or more bias terms. The device selects a generative model configured to generate an output given a textual prompt. The device generates a debiased model by pruning neuron connections in a text encoder of the generative model associated with the one or more terms of interest and the one or more bias terms.

## Claims

### Claim 1

A method, comprising:
obtaining, by a device, one or more terms of interest;
obtaining, by the device, one or more bias terms;
selecting, by the device, a generative model configured to generate an output given a textual prompt; and
generating, by the device, a debiased model by pruning neuron connections in a text encoder of the generative model associated with the one or more terms of interest and the one or more bias terms.

### Claim 2 (depends on 1)

The method as in claim 1, wherein the device obtains the one or more terms of interest via a user interface.

### Claim 3 (depends on 1)

The method as in claim 1, wherein the device obtains the one or more bias terms via a user interface.

### Claim 4 (depends on 1)

The method as in claim 1, wherein the device selects the generative model based on a selection of the generative model by a user via a user interface.

### Claim 5 (depends on 1)

The method as in claim 1, wherein the generative model is a text-to-image diffusion model.

### Claim 6 (depends on 1)

The method as in claim 1, wherein the output comprises an image depicting a person.

### Claim 7 (depends on 1)

The method as in claim 1, further comprising:
obtaining, by the device, a sparsity ratio, wherein the device prunes the generative model by applying a binary mask to its text encoder based on the sparsity ratio.

### Claim 8 (depends on 1)

The method as in claim 1, wherein the one or more terms of interest correspond to one or more types of people.

### Claim 9 (depends on 1)

The method as in claim 1, wherein the one or more bias terms correspond to at least one of: a race, an ethnicity, or a gender.

### Claim 10 (depends on 1)

The method as in claim 1, further comprising:
deploying the debiased model in replacement for the generative model.

### Claim 11

An apparatus, comprising:
one or more network interfaces;
a processor coupled to the one or more network interfaces and configured to execute one or more processes; and
a memory configured to store a process that is executable by the processor, the process when executed configured to:
obtain one or more terms of interest;
obtain one or more bias terms;
select a generative model configured to generate an output given a textual prompt; and
generate a debiased model by pruning neuron connections in a text encoder of the generative model associated with the one or more terms of interest and the one or more bias terms.

### Claim 12 (depends on 11)

The apparatus as in claim 11, wherein the apparatus obtains the one or more terms of interest via a user interface.

### Claim 13 (depends on 11)

The apparatus as in claim 11, wherein the apparatus obtains the one or more bias terms via a user interface.

### Claim 14 (depends on 11)

The apparatus as in claim 11, wherein the apparatus selects the generative model based on a selection of the generative model by a user via a user interface.

### Claim 15 (depends on 11)

The apparatus as in claim 11, wherein the generative model is a text-to-image diffusion model.

### Claim 16 (depends on 11)

The apparatus as in claim 11, wherein the output comprises an image depicting a person.

### Claim 17 (depends on 11)

The apparatus as in claim 11, wherein the process when executed is further configured to:
obtain a sparsity ratio, wherein the apparatus prunes the generative model by applying a binary mask to its text encoder based on the sparsity ratio.

### Claim 18 (depends on 11)

The apparatus as in claim 11, wherein the one or more terms of interest correspond to one or more types of people.

### Claim 19 (depends on 11)

The apparatus as in claim 11, wherein the one or more bias terms correspond to at least one of: a race, an ethnicity, or a gender.

### Claim 20

A tangible, non-transitory, computer-readable medium storing program instructions that cause a device to execute a process comprising:
obtaining, by the device, one or more terms of interest;
obtaining, by the device, one or more bias terms;
selecting, by the device, a generative model configured to generate an output given a textual prompt; and
generating, by the device, a debiased model by pruning neuron connections in a text encoder of the generative model associated with the one or more terms of interest and the one or more bias terms.
