Back to results
US 12,724,961 B2Granted
Plagiarism detection via revision analysis
AssigneeRumi Technologies, Inc.
InventorsMohammad Hossein Ghasemzadeh, MohammadJavad NoroozOliaee, Ghazaleh Sadooghi
Life of this patent
Estimating the term…
Abstract
To detect plagiarism, a computing system obtains, for media content, a revision history comprising one or more revision events. Each revision event describes a modification of the media content performed via user interaction with a content editor. The computing system detects plagiarism within the media content based on the revision events.
Claims18
A method of detecting plagiarism, implemented by a computing system, the method comprising:
- monitoring, by a revision handler executing on one or more processors of the computing system, a content editor for user interactions that result in modifications to an individual document, the user interactions comprising keystrokes, paste operations, and mouse events;
- detecting one or more of the user interactions and, in response, generating one or more revision events, each revision event describing a modification type and time spent performing a modification resulting from one or more of the detected user interactions;
- sending one or more inter-process communications comprising the one or more revision events from the revision handler to a plagiarism analysis engine executing on the one or more processors of the computing system;
- assigning different weights to the revision events based on the modification types;
- using, by the plagiarism analysis engine, a machine learning model to calculate an originality score based on the revision events, wherein the originality score reflects:
- increasing amounts of originality with greater numbers of revision events;
- increasing amounts of originality with greater amounts of time spent performing the modifications described by the revision events;
- and
- different amounts of originality depending upon the different weights assigned to the revision events;
- detecting, by the plagiarism analysis engine, plagiarism within the individual document based on a determination that the originality score reflects less than a threshold amount of originality;
- and
- learning, by the machine learning model, to detect further plagiarism from the one or more revision events.
9 more elements
A computing system for detecting plagiarism, the computing system comprising:
- processing circuitry and memory circuitry, the memory circuitry storing instructions executable by the processing circuitry whereby the computing system is configured to:
- monitor, by a revision handler, a content editor for user interactions that result in modifications to an individual document, the user interactions comprising keystrokes, paste operations, and mouse events;
- detect one or more of the user interactions and, in response, generate one or more revision events, each revision event describing a modification type and time spent performing a modification resulting from one or more of the detected user interactions;
- send one or more inter-process communications comprising the one or more revision events from the revision handler to a plagiarism analysis engine;
- assign different weights to the revision events based on the modification types;
- use, by the plagiarism analysis engine, a machine learning model to calculate an originality score based on the revision events, wherein the originality score reflects:
- increasing amounts of originality with greater numbers of revision events;
- increasing amounts of originality with greater amounts of time spent performing the modifications described by the revision events;
- and
- different amounts of originality depending upon the different weights assigned to the revision events;
- detect, by the plagiarism analysis engine, plagiarism within the individual document based on a determination that the originality score reflects less than a threshold amount of originality;
- and
- learn, by the machine learning model, to detect further plagiarism from the one or more revision events.
10 more elements
A non-transitory computer readable medium storing software instructions for controlling a computing system to detect plagiarism, wherein running the software instructions on processing circuitry of the computing system, causes the computing system to:
- monitor, by a revision handler, a content editor for user interactions that result in modifications to an individual document, the user interactions comprising keystrokes, paste operations, and mouse events;
- detect one or more of the user interactions and, in response, generate one or more revision events, each revision event describing a modification type and time spent performing a modification resulting from one or more of the detected user interactions;
- send one or more inter-process communications comprising the one or more revision events from the revision handler to a plagiarism analysis engine;
- assign different weights to the revision events based on the modification types;
- use, by the plagiarism analysis engine, a machine learning model to calculate an originality score based on the revision events, wherein the originality score reflects:
- increasing amounts of originality with greater numbers of revision events;
- increasing amounts of originality with greater amounts of time spent performing the modifications described by the revision events;
- and
- different amounts of originality depending upon the different weights assigned to the revision events;
- detect, by the plagiarism analysis engine, plagiarism within the individual document based on a determination that the originality score reflects less than a threshold amount of originality;
- and
- learn, by the machine learning model, to detect further plagiarism from the one or more revision events.
9 more elements
Dependencies are inferred from the claim text.