# US 2024/0005082 A1 · Embedding Texts into High Dimensional Vectors in Natural Language Processing

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

- Number: US 2024/0005082 A1
- Document ID: US20240005082A1
- Status: Published application
- Kind: A1
- Application number: 17/825,311
- Filed: 2022-05-26
- Published: 2024-01-04
- Earliest claimed: 2022-05-26
- Assignees: AT&T Mobility II LLC
- Applicants: AT&T Mobility II LLC
- Inventors: Changchuan Yin; Shahzad Saeed
- Web page: https://patentgrep.com/patent/US20240005082A1

## CPC

- G06F 40/126: Character encoding
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F 40/20: Natural language analysis (semantic analysis of natural language G06F40/30)
- G06F 40/216: using statistical methods
- G06F 40/289: Phrasal analysis, e.g. finite state techniques or chunking
- G06F 40/30: Semantic analysis
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Concepts and technologies disclosed herein are directed to embedding texts to high dimensional vectors in natural language processing (“NLP”). According to one aspect, an NLP system can receive an input text that includes n number of words. The NLP system can encode the input text into a first matrix using a word embedding algorithm, such as Word2Vec algorithm. The NLP system can encode the input text into the Word2Vec by embedding each word in the n number of words of the input text into a k-dimensional Word2Vec vector using the Word2Vec algorithm. The NLP system also can decode the first matrix into a second matrix using a text embedding algorithm. In some embodiments, the second matrix is a congruence derivative matrix. The NLP system can then output the second matrix to a machine learning module that implements a machine learning technique such as short text classification.

## Claims

### Claim 1

A method comprising:
receiving, by a natural language processing system, an input text;
encoding, by the natural language processing system, the input text into a first matrix using a word embedding algorithm;
decoding, by the natural language processing system, the first matrix into a second matrix using a text embedding algorithm; and
outputting, by the natural language processing system, the second matrix.

### Claim 2 (depends on 1)

The method of claim 1, wherein the input text comprises n number of words.

### Claim 3 (depends on 2)

The method of claim 2, wherein the word embedding algorithm comprises Word2Vec algorithm, and wherein the first matrix comprises a Word2Vec matrix.

### Claim 4 (depends on 3)

The method of claim 3, wherein encoding, by the natural language processing system, the input text into the Word2Vec matrix using the Word2Vec algorithm comprises embedding each word in the n number of words of the input text into a k-dimensional Word2Vec vector using the Word2Vec algorithm.

### Claim 5 (depends on 4)

The method of claim 4, wherein decoding, by the natural language processing system, the first matrix into the second matrix using the text embedding algorithm comprises decoding, by the natural language processing system, the Word2Vec matrix into a congruence derivative matrix using a congruence derivative vector representation.

### Claim 6 (depends on 5)

The method of claim 5, wherein outputting, by the natural language processing system, the second matrix comprises outputting, by the natural language processing system, the congruence derivative matrix to a machine learning module.

### Claim 7 (depends on 6)

The method of claim 6, wherein the machine learning module implements short text classification.

### Claim 8

A natural language processing system comprising
a processor; and
a memory comprising instructions that, when executed by the processor, cause the processor to perform operations comprising
receiving an input text,
encoding the input text into a first matrix using a word embedding algorithm,
decoding the first matrix into a second matrix using a text embedding algorithm, and
outputting the second matrix.

### Claim 9 (depends on 8)

The natural language processing system of claim 8, wherein the input text comprises n number of words.

### Claim 10 (depends on 9)

The natural language processing system of claim 9, wherein the word embedding algorithm comprises Word2Vec algorithm, and wherein the first matrix comprises a Word2Vec matrix.

### Claim 11 (depends on 10)

The natural language processing system of claim 10, wherein encoding the input text into the Word2Vec matrix using the Word2Vec algorithm comprises embedding each word in the n number of words of the input text into a k-dimensional Word2Vec vector using the Word2Vec algorithm.

### Claim 12 (depends on 11)

The natural language processing system of claim 11, wherein decoding the first matrix into the second matrix using the text embedding algorithm comprises decoding the Word2Vec matrix into a congruence derivative matrix using a congruence derivative vector representation.

### Claim 13 (depends on 12)

The natural language processing system of claim 12, wherein outputting the second matrix comprises outputting the congruence derivative matrix to a machine learning module.

### Claim 14 (depends on 13)

The natural language processing system of claim 13, wherein the machine learning module implements short text classification.

### Claim 15

A computer-readable storage medium comprising computer-executable instructions that, when executed by a processor, cause the processor to perform operations comprising:
receiving an input text;
encoding the input text into a first matrix using a word embedding algorithm;
decoding the first matrix into a second matrix using a text embedding algorithm; and
outputting the second matrix.

### Claim 16 (depends on 15)

The computer-readable storage medium of claim 15, wherein the input text comprises n number of words.

### Claim 17 (depends on 16)

The computer-readable storage medium of claim 16, wherein the word embedding algorithm comprises Word2Vec algorithm, and wherein the first matrix comprises a Word2Vec matrix.

### Claim 18 (depends on 17)

The computer-readable storage medium of claim 17, wherein encoding the input text into the Word2Vec matrix using the Word2Vec algorithm comprises embedding each word in the n number of words of the input text into a k-dimensional Word2Vec vector using the Word2Vec algorithm.

### Claim 19 (depends on 18)

The computer-readable storage medium of claim 18, wherein decoding the first matrix into the second matrix using the text embedding algorithm comprises decoding the Word2Vec matrix into a congruence derivative matrix using a congruence derivative vector representation.

### Claim 20 (depends on 19)

The computer-readable storage medium of claim 19, wherein outputting the second matrix comprises outputting the congruence derivative matrix to a machine learning module, and wherein the machine learning module implements short text classification.
