# US 2021/0279423 A1 · Automatic summarization of content in electronic messages

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

- Number: US 2021/0279423 A1
- Document ID: US20210279423A1
- Status: Published application
- Kind: A1
- Application number: 17/328,572
- Filed: 2021-05-24
- Published: 2021-09-09
- Earliest claimed: 2019-01-21
- Assignees: Microsoft Technology Licensing, LLC
- Applicants: Microsoft Technology Licensing, LLC
- Inventors: Kausik Ghatak; Ganessh Kumar R P; Priyanka Goel; Neeraj Singh; Swathi Karri
- Web page: https://patentgrep.com/patent/US20210279423A1

## CPC

- G06F 40/30: Semantic analysis
- G06Q 10/107: Computer-aided management of electronic mailing [e-mailing]
- G06F 16/345: {Summarisation for human users}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- 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

Servers configured to perform automatic summarization of content in electronic messages are discloses herein. In one embodiment, upon receiving an email, an server determines whether the incoming email is a templated message. In response to determining that the incoming email is not a templated message, the server classifies one or more sentences in the email as a statement of decision, judgement, inference, or fact, cluster the classified statements into clusters, and select one or more of the clusters to automatically generate summaries of the incoming email. The server can then insert data representing the generated summaries into the email before transmitting the email to a destination via a computer network.

## Claims

### Claim 1

A method of automatic summarization of content in electronic messages, the method comprising:
upon receiving, at a server, an incoming electronic message having a message body with one or more sentences,
classifying the one or more sentences in the message body individually as a statement of decision, judgement, inference, or fact;
clustering the classified statements into one or more clusters according to a hierarchy of decision, judgement, inference, and fact;
selecting one or more of the clusters to automatically generate summaries of the incoming electronic message corresponding to each of the selected clusters; and
inserting data representing at least one of the generated summaries into the incoming electronic message before transmitting the incoming electronic message to a destination via a computer network.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
assigning, to each classified statement of decision, judgement, inference, or fact, a relevance score; and
calculating a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster.

### Claim 3 (depends on 1)

The method of claim 1, further comprising:
assigning, to each classified statement of decision, judgement, inference, or fact, a relevance score;
calculating a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster; and
wherein selecting one or more of the clusters includes selecting one or more of the clusters based on the calculated cluster scores.

### Claim 4 (depends on 1)

The method of claim 1, further comprising:
assigning, to each classified statement of decision, judgement, inference, or fact, a relevance score, wherein the assigned relevance scores are based on at least one of a sender identity of the incoming electronic message or recency of the corresponding statements;
calculating a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster; and
wherein selecting one or more of the clusters includes selecting one or more of the clusters based on the calculated cluster score.

### Claim 5 (depends on 1)

The method of claim 1 wherein clustering the classified statements includes:
determining whether the classified statements include at least one decision; and
in response to determining that the classified statements includes at least one decision, generating a cluster around the at least one decision.

### Claim 6 (depends on 1)

The method of claim 1 wherein clustering the classified statements includes:
determining whether the classified statements include at least one decision; and
in response to determining that the classified statements does not include at least one decision,
determining whether the classified statements include at least one judgement; and
in response to determining that the classified statements includes at least one judgement, generating a cluster around the at least one judgment.

### Claim 7 (depends on 1)

The method of claim 1 wherein clustering the classified statements includes:
determining whether the classified statements include at least one decision; and
in response to determining that the classified statements does not include at least one decision,
determining whether the classified statements include at least one judgement;
in response to determining that the classified statements does not include at least one judgement,
determining whether the classified statements include at least one inference; and
in response to determining that the classified statements includes at least one inference, generating a cluster around the at least one inference.

### Claim 8 (depends on 1)

The method of claim 1, further comprising:
determining whether the incoming electronic message is a templated message; and
in response to determining that the incoming electronic message is a templated message,
identifying a summary template corresponding to the incoming electronic message; and
generating a summary based on the identified summary template.

### Claim 9 (depends on 1)

The method of claim 1 wherein inserting data representing at least one of the generated summaries includes:
converting the at least one of the generated summaries from text to a voice message; and
inserting the converted voice message to the incoming electronic message for playback at a client device.

### Claim 10

A computing device for processing electronic messages, the computing device comprising:
a processor;
a memory containing instructions executable by the processor to cause the computing device to:
receive an incoming electronic message having a message body having multiple sentences,
determine whether the incoming electronic message is a tem plated message based on the multiple sentences in the message body; and
in response to determining that the incoming electronic message is not a tem plated message,
classify the one or more sentences in the message body individually as a statement of decision, judgement, inference, or fact utilizing a classification model;
generate multiple clusters of the classified statements according to a hierarchy of decision, judgement, inference, and fact;
select one of the clusters to automatically generate a summary of the incoming electronic message corresponding to the selected one or the clusters; and
outputting the automatically generated summary to a user as a voice or text message.

### Claim 11 (depends on 10)

The computing device of claim 10 wherein the memory includes additional instructions executable by the processor to cause the computing device to:
assign, to each classified statement of decision, judgement, inference, or fact, a relevance score; and
calculate a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster.

### Claim 12 (depends on 10)

The computing device of claim 10 wherein the memory includes additional instructions executable by the processor to cause the computing device to:
assign, to each classified statement of decision, judgement, inference, or fact, a relevance score;
calculate a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster; and
wherein to select one of the clusters includes to select one of the clusters having a highest calculated cluster score.

### Claim 13 (depends on 10)

The computing device of claim 10 wherein the memory includes additional instructions executable by the processor to cause the computing device to:
assign, to each classified statement of decision, judgement, inference, or fact, a relevance score, wherein the assigned relevance scores are based on at least one of a sender identity of the incoming electronic message or recency of the corresponding statements;
calculate a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster; and
wherein to select one of the clusters includes to select one of the clusters having a highest calculated cluster score.

### Claim 14 (depends on 10)

The computing device of claim 10 wherein to cluster the classified statements includes to:
determine whether the classified statements include at least one decision; and
in response to determining that the classified statements includes at least one decision, generate a cluster around the at least one decision.

### Claim 15 (depends on 10)

The computing device of claim 10 wherein to cluster the classified statements includes to:
determine whether the classified statements include at least one decision; and
in response to determining that the classified statements does not include at least one decision,
determine whether the classified statements include at least one judgement; and
in response to determining that the classified statements includes at least one judgement, generate a cluster around the at least one judgment.

### Claim 16 (depends on 10)

The computing device of claim 10 wherein to cluster the classified statements includes to:
determine whether the classified statements include at least one decision; and
in response to determining that the classified statements does not include at least one decision,
determine whether the classified statements include at least one judgement;
in response to determining that the classified statements does not include at least one judgement,
determine whether the classified statements include at least one inference; and
in response to determining that the classified statements includes at least one inference, generate a cluster around the at least one inference.

### Claim 17

A method of automatic summarization of content in electronic messages, the method comprising:
receiving, at an server, an incoming electronic message having an message body having one or more sentences; and
subsequent to receiving the incoming electronic message,
classifying the one or more sentences in the message body individually as a statement of decision, judgement, inference, or fact;
clustering the classified statements into one or more clusters according to a hierarchy of decision, judgement, inference, and fact;
automatically generating one or more summaries of the incoming electronic message based on one or more of the clusters of classified statements; and
inserting data representing one of the generated summaries into the incoming electronic message before transmitting the incoming electronic message to a destination via a computer network.

### Claim 18 (depends on 17)

The method of claim 17, further comprising:
assigning, to each classified statement of decision, judgement, inference, or fact, a relevance score; and
calculating a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster.

### Claim 19 (depends on 17)

The method of claim 17, further comprising:
assigning, to each classified statement of decision, judgement, inference, or fact, a relevance score;
calculating a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster; and
wherein selecting one or more of the clusters includes selecting one or more of the clusters based on the calculated cluster scores.

### Claim 20 (depends on 17)

The method of claim 17, further comprising:
assigning, to each classified statement of decision, judgement, inference, or fact, a relevance score, wherein the assigned relevance scores are based on at least one of a sender identity of the incoming electronic message or recency of the corresponding statements;
calculating a cluster score for the individual clusters based on the assigned relevance scores of the statements in each cluster; and
wherein selecting one or more of the clusters includes selecting one or more of the clusters based on the calculated cluster score.
