# US 10,769,524 B1 · Non-binary gender filter

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

- Number: US 10,769,524 B1
- Document ID: US10769524B1
- Status: Granted
- Kind: B1
- Application number: 15/695,936
- Filed: 2017-09-05
- Granted: 2020-09-08
- Earliest claimed: 2017-09-05
- Assignees: A9.com, Inc.
- Applicants: A9.com, Inc.
- Inventors: Aishwarya Natesh
- Web page: https://patentgrep.com/patent/US10769524B1

## CPC

- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06F 16/5838: {using colour}
- G06F 16/90335: {Query processing}
- G06F 18/24: Classification techniques
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/0475: Generative networks
- G06N 3/0495: Quantised networks; Sparse networks; Compressed networks
- G06N 3/08: Learning methods
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06N 3/09: Supervised learning
- G06N 3/094: Adversarial learning
- G06V 40/10: Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06Q 30/00: Commerce
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

Various embodiments utilize a machine learning-based approach to filter items, such as apparel items, based on non-binary gender styles. For example, an electronic catalog of apparel items can be assigned gender scores on a gender scale by a neural network trained to determine a gender score of an apparel item based on an image representation of the apparel item. The neural network may be trained on training data that includes images of various apparel items with gender designations. The apparel items in the electronic catalog are assigned a gender score attribute that reflects how masculine or feminine the apparel item may be. As such, the apparel items can be organized (e.g., sorted, filtered, ranked) based on a non-binary gender score in addition to other attributes, such as item type, size, color, brand, etc. Thus, a user can include non-binary gender style as a search or filtering criteria.

## Claims

### Claim 1

A computing system, comprising:
at least one computing device processor;
a memory device including instructions that, when executed by the at least one computing device processor, cause the computing system to:
train a neural network using training data comprising images of a plurality of apparel items with respective gender designations to generate a trained neural network, the trained neural network operable to assign a gender score to an apparel item, the gender score corresponding to a position on a gender scale;
assign, using the neural network, respective gender scores to a collection of apparel items offered for consumption in an electronic marketplace;
receive an apparel search query;
receive a gender input value, the gender input value selected from a non-binary range of values and corresponding to one or more acceptable gender scores;
execute the apparel search query and the acceptable gender scores against the collection of apparel items; and
determine, from the collection of apparel items, a set of apparel items responsive to the apparel search query and having gender scores included in the one or more acceptable gender scores.

### Claim 2 (depends on 1)

The computing system of claim 1, wherein the instructions when executed further cause the computing system to:
extract feature vectors associated with the training data from a layer of the neural network;
map the feature vectors onto a 2-dimensional scale; and
generate the gender scale based on the 2-dimensional scale.

### Claim 3 (depends on 1)

The computing system of claim 1, wherein the instructions when executed further cause the computing system to:
extract feature vectors associated with the training data from a layer of the neural network;
identify a plurality of clusters based on the feature vectors; and
determine a plurality of classification bins based at least in part on the plurality of clusters, the gender scale defined by the plurality of classification bins and the assigned gender score corresponding to one of the plurality of classification bins.

### Claim 4 (depends on 1)

The computing system of claim 1, wherein the gender scale is defined between a masculine end and a feminine end, and the gender score reflects how masculine or feminine an apparel item is as determined by the trained neural network.

### Claim 5

A computer-implemented method, comprising:
receiving a request for a content item;
receiving a gender input value associated with the request, the gender input value selected from a non-binary range of input values;
executing the request against a collection of content items associated with respective gender scores on a gender scale, the respective gender scores assigned by a trained model, and the gender input value corresponding to one or more acceptable gender scores; and
determining a set of content items, from the collection of content items, responsive to the request and having gender scores included in the one or more acceptable gender scores.

### Claim 6 (depends on 5)

The computer-implemented method of claim 5, further comprising:
training the model using training data comprising a plurality of content items with respective gender designations to generate the trained model.

### Claim 7 (depends on 6)

The computer-implemented method of claim 6, wherein the respective gender designations of the training data include male and female, and the gender scale includes a plurality of values between male and female.

### Claim 8 (depends on 6)

The computer-implemented method of claim 6, further comprising:
extracting feature vectors associated with the training data; and
generating the gender scale based on the extracted feature vectors.

### Claim 9 (depends on 5)

The computer-implemented method of claim 5, further comprising:
using the trained model to extract feature vectors associated with the collection of content items;
mapping the feature vectors associated with the collection of content items onto the gender scale; and
assigning the respective gender scores to the collection of content items based at least in part on the mapping of the feature vectors to the gender scale.

### Claim 10 (depends on 5)

The computer-implemented method of claim 5, further comprising:
determine a plurality of classification bins, the gender scale defined by the plurality of classification bins and the assigned gender score corresponding to one of the plurality of classification bins.

### Claim 11 (depends on 5)

The computer-implemented method of claim 5, further comprising:
presenting a slider having a plurality of discrete or continuous positions, the continuous positions corresponding to the non-binary range of input values; and
receiving a selected position of the slider, the selected position corresponding to the gender input value.

### Claim 12 (depends on 5)

The computer-implemented method of claim 5, further comprising:
presenting a plurality of selectable images corresponding to the non-binary range of input values, the plurality of selectable images illustrating apparel items exemplifying the one or more acceptable gender scores; and
receiving an image selection, the image selection corresponding to the gender input value.

### Claim 13 (depends on 5)

The computer-implemented method of claim 5, further comprising:
displaying the set of content items;
receiving interaction data with a selected content item of the set of content items; and
associating the selected content item with the gender input value and the interaction data as feedback data.

### Claim 14 (depends on 5)

The computer-implemented method of claim 5, further comprising:
associating the gender input value with a user account; and
automatically applying the gender input value to a future request associated with the user account.

### Claim 15 (depends on 5)

The computer-implemented method of claim 5, further comprising:
identifying a plurality of clusters of gender scores on the gender scale and the images associated with each of the plurality of clusters;
defining a plurality of classes based at least in part on the plurality of clusters; and
training a second model, using the images annotated with the respective class of the plurality of classes, to classify query images into one of the plurality of classes.

### Claim 16 (depends on 15)

The computer-implemented method of claim 15, further comprising:
receiving the gender input value from the content provider; and
transmitting the set of content items to the content provider.

### Claim 17

A computer-implemented method, comprising:
train a neural network using training data comprising images of a plurality of apparel items with respective gender designations to generate a trained neural network, the trained neural network operable to assign a gender score to an apparel item, the gender score corresponding to a position on a gender scale;
receiving content items from a content provider;
analyzing the content items using the trained neural network;
determining gender scores for the content items using the neural network;
associating the content items with the respective gender scores; and
transmitting data representing the content items with respective gender scores to the content provider.

### Claim 18 (depends on 17)

The computing system of claim 17, further comprising:
receiving the images of the plurality of apparel items with respective gender designations of the training data from the content provider.

### Claim 19 (depends on 18)

The computing system of claim 18, further comprising:
receiving the gender designations of the plurality of apparel items from the content provider.

### Claim 20 (depends on 17)

The computing system of claim 17, further comprising:
classifying the content items into a plurality of classifications, the gender scores representing the classifications of the content items; and
transmitting the classifications to the content provider.