# US 2021/0004662 A1 · Systems and methods for providing content

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

- Number: US 2021/0004662 A1
- Document ID: US20210004662A1
- Status: Published application
- Kind: A1
- Application number: 17/030,157
- Filed: 2020-09-23
- Published: 2021-01-07
- Earliest claimed: 2016-12-30
- Assignees: Facebook, Inc.
- Applicants: Facebook, Inc.
- Inventors: Kai Li; Fei Yang; Balamanohar Paluri
- Web page: https://patentgrep.com/patent/US20210004662A1

## CPC

- G06F 16/783: using metadata automatically derived from the content
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/09: Supervised learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Systems, methods, and non-transitory computer-readable media can receive a first content item having a set of frames. A binary hash code that represents the first content item is generated using at least an aggregation model and an iterative quantization hash model, the binary hash code being determined based at least in part on the set of frames of the first content item. The binary hash code is stored, wherein a similarity between the first content item and a second content item is capable of being measured based at least in part on a comparison of the binary hash code of the first content item and a binary hash code of the second content item.

## Claims

### Claim 1

A computer-implemented method comprising:
generating, by a computing system, a first binary hash code that represents a first portion of a first content item;
generating, by the computing system, a second binary hash code that represents a second portion of the first content item; and
determining, by the computing system, a similarity between the first content item and at least one second content item, wherein the similarity is measured based at least in part on a comparison of the first binary hash code or the second binary hash code associated with the first content item and a binary hash code associated with the second content item.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, further comprising:
storing, by the computing system, (i) the first binary hash code that represents the first portion of the first content item and (ii) the second binary hash code that represents the second portion of the first content item in a search index, wherein the similarity between the first content item the at least one second content item is determined based at least in part on the search index.

### Claim 3 (depends on 2)

The computer-implemented method of claim 2, wherein the search index identifies the first portion as a first video clip associated with the first content item, and wherein the search index identifies the second portion as a second video clip associated with the first content item.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, wherein determining the similarity between the first content item and the at least one second content item further comprises:
performing, by the computing system, a bitwise operation based on the first binary hash code associated with the first content item and the binary hash code associated with the at least one second content item; or
performing, by the computing system, a bitwise operation based on the second binary hash code associated with the first content item and the binary hash code associated with the at least one second content item.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, wherein generating the first binary hash code that represents the first portion of the first content item further comprises:
generating, by the computing system, the first binary hash code based on a video aggregation model and an iterative quantization hash model, wherein the first binary hash code is determined based at least in part on a set of frames associated with the first portion of the first content item.

### Claim 6 (depends on 5)

The computer-implemented method of claim 5, wherein generating the first binary hash code based on the video aggregation model and the iterative quantization hash model further comprises:
transforming, by the computing system, the set of frames associated with the first portion of the first content item to a fixed-length feature representation based on the video aggregation model; and
quantizing, by the computing system, the fixed-length feature representation based on the iterative quantization hash model.

### Claim 7 (depends on 6)

The computer-implemented method of claim 6, wherein transforming the set of frames associated with the first portion of the first content item to a fixed-length feature representation further comprises:
decomposing, by the computing system, the set of frames associated with the first portion of the first content item into a set of components; and
concatenating, by the computing system, the set of components to generate the fixed-length feature representation of the set of frames associated with the first portion of the first content item.

### Claim 8 (depends on 7)

The computer-implemented method of claim 7, wherein the decomposition of the set of frames associated with the first portion of the first content item identifies a topic or concept associated with the first portion of the first content item.

### Claim 9 (depends on 1)

The computer-implemented method of claim 1, wherein generating the second binary hash code that represents the second portion of the first content item further comprises:
generating, by the computing system, the second binary hash code based on a video aggregation model and an iterative quantization hash model, wherein the second binary hash code is determined based at least in part on a different set of frames associated with the second portion of the first content item.

### Claim 10 (depends on 9)

The computer-implemented method of claim 9, wherein generating the second binary hash code based on the video aggregation model and the iterative quantization hash model further comprises:
transforming, by the computing system, the different set of frames associated with the second portion of the first content item to a different fixed-length feature representation based on the video aggregation model; and
quantizing, by the computing system, the different fixed-length feature representation based on the iterative quantization hash model.

### Claim 11

A system comprising:
at least one processor; and
a memory storing instructions that, when executed by the at least one processor, cause the system to perform:
generating a first binary hash code that represents a first portion of a first content item;
generating a second binary hash code that represents a second portion of the first content item; and
determining a similarity between the first content item and at least one second content item, wherein the similarity is measured based at least in part on a comparison of the first binary hash code or the second binary hash code associated with the first content item and a binary hash code associated with the second content item.

### Claim 12 (depends on 11)

The system of claim 11, wherein the instructions further cause the system to perform:
storing (i) the first binary hash code that represents the first portion of the first content item and (ii) the second binary hash code that represents the second portion of the first content item in a search index, wherein the similarity between the first content item the at least one second content item is determined based at least in part on the search index.

### Claim 13 (depends on 12)

The system of claim 12, wherein the search index identifies the first portion as a first video clip associated with the first content item, and wherein the search index identifies the second portion as a second video clip associated with the first content item.

### Claim 14 (depends on 11)

The system of claim 11, wherein determining the similarity between the first content item and the at least one second content item further comprises:
performing a bitwise operation based on the first binary hash code associated with the first content item and the binary hash code associated with the at least one second content item; or
performing a bitwise operation based on the second binary hash code associated with the first content item and the binary hash code associated with the at least one second content item.

### Claim 15 (depends on 11)

The system of claim 11, wherein generating the first binary hash code that represents the first portion of the first content item further comprises:
generating the first binary hash code based on a video aggregation model and an iterative quantization hash model, wherein the first binary hash code is determined based at least in part on a set of frames associated with the first portion of the first content item.

### Claim 16

A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:
generating a first binary hash code that represents a first portion of a first content item;
generating a second binary hash code that represents a second portion of the first content item; and
determining a similarity between the first content item and at least one second content item, wherein the similarity is measured based at least in part on a comparison of the first binary hash code or the second binary hash code associated with the first content item and a binary hash code associated with the second content item.

### Claim 17 (depends on 16)

The non-transitory computer-readable storage medium of claim 16, wherein the instructions further cause the computing system to perform:
storing (i) the first binary hash code that represents the first portion of the first content item and (ii) the second binary hash code that represents the second portion of the first content item in a search index, wherein the similarity between the first content item the at least one second content item is determined based at least in part on the search index.

### Claim 18 (depends on 17)

The non-transitory computer-readable storage medium of claim 17, wherein the search index identifies the first portion as a first video clip associated with the first content item, and wherein the search index identifies the second portion as a second video clip associated with the first content item.

### Claim 19 (depends on 16)

The non-transitory computer-readable storage medium of claim 16, wherein determining the similarity between the first content item and the at least one second content item further comprises:
performing a bitwise operation based on the first binary hash code associated with the first content item and the binary hash code associated with the at least one second content item; or
performing a bitwise operation based on the second binary hash code associated with the first content item and the binary hash code associated with the at least one second content item.

### Claim 20 (depends on 16)

The non-transitory computer-readable storage medium of claim 16, wherein generating the first binary hash code that represents the first portion of the first content item further comprises:
generating the first binary hash code based on a video aggregation model and an iterative quantization hash model, wherein the first binary hash code is determined based at least in part on a set of frames associated with the first portion of the first content item.
