# US 2022/0035994 A1 · Computer-implemented presentation of synonyms based on syntactic dependency

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

- Number: US 2022/0035994 A1
- Document ID: US20220035994A1
- Status: Published application
- Kind: A1
- Application number: 17/160,193
- Filed: 2021-01-27
- Published: 2022-02-03
- Earliest claimed: 2020-07-31
- Assignees: Grammarly, Inc.
- Applicants: Grammarly, Inc.
- Inventors: Roman Guliak; Stanislav Levental; Yuriy Tanskiy
- Web page: https://patentgrep.com/patent/US20220035994A1

## CPC

- G06F 40/166: Editing, e.g. inserting or deleting
- G06F 3/0481: based on specific properties of the displayed interaction object or a metaphor-based environment, e.g. interaction with desktop elements like windows or icons, or assisted by a cursor's changing behaviour or appearance
- G06F 3/04886: by partitioning the display area of the touch-screen or the surface of the digitising tablet into independently controllable areas, e.g. virtual keyboards or menus
- G06F 16/374: {Thesaurus}
- G06F 40/211: Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
- G06F 40/247: Thesauruses; Synonyms
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F 40/30: Semantic analysis
- G06N 3/0495: Quantised networks; Sparse networks; Compressed networks
- G06N 3/09: Supervised learning
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

In an embodiment, the disclosed technologies are capable of identifying a target word within a text sequence; displaying a subset of candidate synonyms for the target word, determining a synonym selected from the subset of candidate synonyms, and replacing the target word with the selected synonym, where the subset of candidate synonyms has been created using syntactic dependency data for the target word.

## Claims

### Claim 1

A mobile computing device, comprising:
a touchscreen;
a processor coupled to the touchscreen;
memory coupled to the processor and storing one or more sequences of instructions which, when executed by the processor, cause the processor to perform:
displaying, on the touchscreen, a graphical user interface that includes an input area and an output area;
receiving a text sequence input including a target word;
displaying, in the output area, a subset of candidate synonyms for the target word, the subset of candidate synonyms being created by
inputting the text sequence including the target word into a dependency parser computer program stored on the mobile computing device and receiving, from the dependency parser computer program, output comprising syntactic dependency data for the target word including data indicating a syntactic dependency relationship between the target word and at least one other word of the text sequence;
inputting the syntactic dependency data into a neural network-based model that has been trained to cause the neural network-based model to output a dependency context vector indicating a semantic meaning of the syntactic dependency data;
retrieving a set of candidate synonyms for the target word from a search engine;
for each candidate synonym of the set of candidate synonyms, using a target word vector indicating a semantic meaning of the target word, a candidate word vector indicating a semantic meaning of the candidate synonym, and the dependency context vector, computing a similarity score indicating a semantic similarity of the candidate synonym to the target word; and
when the similarity score for a candidate synonym satisfies a criterion, adding the candidate synonym to the subset of candidate synonyms.

### Claim 2 (depends on 1)

The mobile computing device of claim 1, wherein receiving a text sequence input comprises a voice interface receiving audio signals containing speech and converting the audio signals to words.

### Claim 3 (depends on 2)

The mobile computing device of claim 2, wherein receiving a text sequence input comprises transcribing words spoken in the vicinity of a microphone.

### Claim 4 (depends on 1)

The mobile computing device of claim 1, wherein receiving a text sequence input comprises extracting text from a digital content item.

### Claim 5 (depends on 4)

The mobile computing device of claim 4, wherein extracting the text from a digital content item comprises extracting the text from a software application.

### Claim 6 (depends on 5)

The mobile computing device of claim 5, wherein the software application comprises one of an email program, an instant messaging program or a social media program.

### Claim 7 (depends on 4)

The mobile computing device of claim 4, wherein extracting the text from a digital content item comprises extracting the text from a document, a message, a social media posting, a list of search results, or a web page.

### Claim 8

A computer-implemented method comprising:
in response to receiving a text sequence input of two or more words:
determining and storing in memory of a mobile computing device, the text sequence and a target word of the text sequence;
inputting the text sequence including the target word into a dependency parser computer program stored on the mobile computing device and receiving, from the dependency parser computer program, output comprising syntactic dependency data for the target word including data indicating a syntactic dependency relationship between the target word and at least one other word of the text sequence;
retrieving a target word vector indicating a semantic meaning of the target word, a set of candidate synonyms for the target word and a candidate word vector for each candidate synonym, the candidate word vector indicating a semantic meaning of the candidate synonym;
inputting the syntactic dependency data into a neural network-based model that has been trained to cause the neural network-based model to output a dependency context vector indicating a semantic meaning of the syntactic dependency data;
computing a similarity score for each candidate synonym of the set of candidate synonyms using the target word vector, the candidate word vector, and the dependency context vector, the similarity score indicating a semantic similarity of the candidate synonym to the target word; and
displaying a subset of candidate synonyms on the graphical user interface, wherein a candidate synonym of the set of candidate synonyms is added to the subset of candidate synonyms when the similarity score for the candidate synonym satisfies a criterion.

### Claim 9 (depends on 8)

The method of claim 8, wherein receiving a text sequence input comprises receiving audio signals containing speech and converting the audio signals to words.

### Claim 10 (depends on 9)

The method of claim 9, wherein receiving a text sequence input comprises transcribing words spoken in the vicinity of a microphone.

### Claim 11 (depends on 8)

The method of claim 8, wherein receiving a text sequence input comprises extracting text from a digital content item.

### Claim 12 (depends on 11)

The method of claim 11, wherein extracting the text from a digital content item comprises extracting the text from a software application.

### Claim 13 (depends on 12)

The method of claim 12, wherein the software application comprises one of an email program, an instant messaging program or a social media program.

### Claim 14 (depends on 11)

The method of claim 11, wherein extracting the text from a digital content item comprises extracting the text from a document, a message, a social media posting, a list of search results, or a web page.

### Claim 15 (depends on 8)

The method of claim 8, wherein receiving a text sequence input including a target word comprises detecting a trigger signal and switching from displaying a virtual keyboard to displaying a thesaurus interface.

### Claim 16 (depends on 15)

The method of claim 15, wherein the trigger signal comprises one or more of a position of a cursor relative to a word in the text sequence input and a time elapsed from a last cursor movement.

### Claim 17

A computer-implemented method comprising:
receiving a text sequence input of two or more words:
in response to one or more of a position of a cursor relative to a word in the text sequence input and a time elapsed from a last cursor movement, determining and storing in memory of a mobile computing device, the text sequence and a target word of the text sequence;
inputting the text sequence including the target word into a dependency parser computer program stored on the mobile computing device and receiving, from the dependency parser computer program, output comprising syntactic dependency data for the target word including data indicating a syntactic dependency relationship between the target word and at least one other word of the text sequence;
retrieving a target word vector indicating a semantic meaning of the target word, a set of candidate synonyms for the target word and a candidate word vector for each candidate synonym, the candidate word vector indicating a semantic meaning of the candidate synonym;
inputting the syntactic dependency data into a neural network-based model that has been trained to cause the neural network-based model to output a dependency context vector indicating a semantic meaning of the syntactic dependency data;
computing a similarity score for each candidate synonym of the set of candidate synonyms using the target word vector, the candidate word vector, and the dependency context vector, the similarity score indicating a semantic similarity of the candidate synonym to the target word;
displaying a subset of candidate synonyms on the graphical user interface, wherein a candidate synonym of the set of candidate synonyms is added to the subset of candidate synonyms when the similarity score for the candidate synonym satisfies a criterion.

### Claim 18 (depends on 17)

The method of claim 17, wherein receiving a text sequence input comprises receiving audio signals containing speech and converting the audio signals to words.

### Claim 19 (depends on 17)

The method of claim 17, wherein receiving a text sequence input comprises extracting text from a digital content item.

### Claim 20 (depends on 19)

The method of claim 19, wherein extracting the text from a digital content item comprises extracting the text from a software application.
