# US 2021/0344691 A1 · Detection of abnormal entities based on syntactic grouping of strings

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

- Number: US 2021/0344691 A1
- Document ID: US20210344691A1
- Status: Published application
- Kind: A1
- Application number: 16/866,411
- Filed: 2020-05-04
- Published: 2021-11-04
- Earliest claimed: 2020-05-04
- Assignees: Microsoft Technology Licensing, LLC
- Applicants: Microsoft Technology Licensing, LLC
- Inventors: Andrey KARPOVSKY; Tomer ROTSTEIN; Fady NASERELDEEN; Naama KRAUS; Roy LEVIN; Yotam LIVNY
- Web page: https://patentgrep.com/patent/US20210344691A1

## CPC

- H04L 63/1416: {Event detection, e.g. attack signature detection}
- H04L 63/1433: {Vulnerability analysis}
- G06N 5/04: Inference or reasoning models
- G06N 20/00: Machine learning
- H04L 63/1441: {Countermeasures against malicious traffic (countermeasures against attacks on cryptographic mechanisms H04L9/002)}
- G06F 16/90344: {by using string matching techniques}
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Unauthorized use of user credentials in a network is detected. Data indicative of text strings being used to access resources in the network is accessed. Regex models are determined for the text strings. Groupings of the regex models are determined based on an optimization of a cumulative weighted function. A regex model having a cumulative weighted function that exceeds a predetermined threshold is identified. An alert is generated when the cumulative weighted function for the identified regex model exceeds the predetermined threshold.

## Claims

### Claim 1

A method for detecting unauthorized use of user credentials in a network, the method comprising:
accessing data indicative of usernames being used to access resources in the network;
determining regex models for the usernames;
determining groupings of the regex models based on an optimization of a cumulative weighted function comprising entropy and coverage;
based on the determined groupings, identifying a regex model having a cumulative weighted function that exceeds a predetermined threshold; and
generating an alert when the cumulative weighted function for the identified regex model exceeds the predetermined threshold.

### Claim 2 (depends on 1)

The method of claim 1, wherein the entropy is determined based on a log of a size of a character pool raised to a power of a length of a string.

### Claim 3 (depends on 1)

The method of claim 1, wherein the coverage comprises a number of usernames covered by the regex model.

### Claim 4 (depends on 1)

The method of claim 1, wherein the optimization is determined using an iterative tree-based approach.

### Claim 5 (depends on 1)

The method of claim 1, further comprising using a machine learning model to identify regex models that exceed the predetermined threshold.

### Claim 6 (depends on 1)

The method of claim 1, wherein the groupings of the regex models are determined for a time window before generating alerts for regex models that exceed the predetermined threshold.

### Claim 7

A computing device configured to detect outlying strings in a computing environment, the computing device comprising:
a processor;
a storage device coupled to the processor;
an application stored in the storage device, wherein execution of the application by the processor configures the computing device to perform acts comprising:
accessing data indicative of text strings being used to access resources in a network;
determining regex models for the text strings;
determining groupings of the regex models based on an optimization of a cumulative weighted function;
based on the determined groupings, identifying a regex model having a cumulative weighted function that exceeds a predetermined threshold; and
generating an alert when the cumulative weighted function for the identified regex model exceeds the predetermined threshold.

### Claim 8 (depends on 7)

The computing device of claim 7, wherein the cumulative weighted function comprises entropy and coverage.

### Claim 9 (depends on 7)

The computing device of claim 7, wherein the text strings comprises one or more of usernames, process identifiers, and application names.

### Claim 10 (depends on 8)

The computing device of claim 8, wherein the entropy is determined based on a log of a size of a character pool raised to a power of a length of a string.

### Claim 11 (depends on 8)

The computing device of claim 8, wherein the coverage comprises a number of usernames covered by the regex model.

### Claim 12 (depends on 7)

The computing device of claim 7, wherein the optimization is determined using an iterative tree-based approach.

### Claim 13 (depends on 7)

The computing device of claim 7, further comprising using a machine learning model to identify regex models that exceed the predetermined threshold.

### Claim 14 (depends on 7)

The computing device of claim 7, wherein the groupings of the regex models are determined for a time window before generating alerts for regex models that exceed the predetermined threshold.

### Claim 15

A computer-readable medium having stored thereon a plurality of sequences of instructions which, when executed by a processor, cause the processor to perform a method comprising:
accessing data indicative of text strings being used to access resources in a network;
determining regex models for the text strings;
determining groupings of the regex models based on an optimization of a cumulative weighted function;
based on the determined groupings, identifying a regex model having a cumulative weighted function that exceeds a predetermined threshold; and
generating an alert when the cumulative weighted function for the identified regex model exceeds the predetermined threshold.

### Claim 16 (depends on 15)

The computer-readable medium of claim 15, further comprising instructions which, when executed by the processor, cause the processor to use a machine learning model to identify regex models that exceed the predetermined threshold.

### Claim 17 (depends on 15)

The computer-readable medium of claim 15, wherein the cumulative weighted function comprises entropy and coverage.

### Claim 18 (depends on 17)

The computer-readable medium of claim 17, wherein the entropy is determined based on a log of a size of a character pool raised to a power of a length of a string.

### Claim 19 (depends on 17)

The computer-readable medium of claim 17, wherein the coverage comprises a number of usernames covered by the regex model.

### Claim 20 (depends on 15)

The computer-readable medium of claim 15, wherein the optimization is determined using an iterative tree-based approach.
