The U.S. Patent and Trademark office recently published an Apple invention indicating the method in which an iPhone could set off an alert or lock the device automatically depending on detected changes in the user behaviour.
As per its patent application for Generating notification based on user behaviour, an iPhone or even other mobile can employ behaviour recognition techniques wherein they can determine if the current user is the owner of the device or not. When the usage pattern does not match up with the owner, an alert, notification or system action gets triggered.
The system depends on pattern recognition and learning to identify unusual behaviour and any form of input or interaction with the device is compiled for behaviour data acquisition and system data such as location, input gesture patterns and motion sensor data. Besides this, other data like vocabulary, grammar as well as keyboard orientation preference could be stored for later analysis and recall.
Sensor Laden Device – Generating Information
With Apple’s new invention, there is a continuous monitoring of user behaviour which is compared to historical data known to be associated with the owner of the device. For instance, an iPhone’s accelerometer generated data on the user’s gait comparing the information with the data that is stored on-board or on a remote server and analysis could take place offsite on a pattern learning server.
A sensor laden device is in constant contact with the user most of the time, generating a lot of information. All details from common misspelling of words, routine device interactions to frequently used phrases and all personal information are gathered by the device.
The behaviour learning server, due to privacy, can be programmed by the user to overlook certain types of behaviour such as location preferences, in which case, generalizations and relative positioning replace granular data thereby protecting the individuals for intrusive analysis.
To overcome computational stress, the pattern learning server can manage group behaviour, classifications and data in clusters rather than analysing and comparing individual action separately. The clusters of behaviour can be compared with one another and data should fall within predefined boundaries in order to proceed with normal operation.
Recognition Server Transmits Signal
If the boundaries are crossed, it indicates unusual behaviour has been detected and the recognition server then transmits a signal to remotely lock the device till the user verifies through password or any other form of trusted identity like Touch ID credentials. The server can also send notification to third party device or any other mobile phone owned by a different party.
It is not known if Apple will be adopting behaviour learning features in future iOS versions but the latest iOS 8 beta builds does not currently support these functions Due to concern over the growing data privacy, users would be likely to manually opt in to the feature if the technology were to roll out one day and the breadth of measured behaviours would tend to narrow down.
Behaviour learning patent application of Apple was first filed last January and creditsas its inventors, go to Gregory T. Lydon and Sylvain Rene Yves.
As per its patent application for Generating notification based on user behaviour, an iPhone or even other mobile can employ behaviour recognition techniques wherein they can determine if the current user is the owner of the device or not. When the usage pattern does not match up with the owner, an alert, notification or system action gets triggered.
The system depends on pattern recognition and learning to identify unusual behaviour and any form of input or interaction with the device is compiled for behaviour data acquisition and system data such as location, input gesture patterns and motion sensor data. Besides this, other data like vocabulary, grammar as well as keyboard orientation preference could be stored for later analysis and recall.
Sensor Laden Device – Generating Information
With Apple’s new invention, there is a continuous monitoring of user behaviour which is compared to historical data known to be associated with the owner of the device. For instance, an iPhone’s accelerometer generated data on the user’s gait comparing the information with the data that is stored on-board or on a remote server and analysis could take place offsite on a pattern learning server.
A sensor laden device is in constant contact with the user most of the time, generating a lot of information. All details from common misspelling of words, routine device interactions to frequently used phrases and all personal information are gathered by the device.
The behaviour learning server, due to privacy, can be programmed by the user to overlook certain types of behaviour such as location preferences, in which case, generalizations and relative positioning replace granular data thereby protecting the individuals for intrusive analysis.
To overcome computational stress, the pattern learning server can manage group behaviour, classifications and data in clusters rather than analysing and comparing individual action separately. The clusters of behaviour can be compared with one another and data should fall within predefined boundaries in order to proceed with normal operation.
Recognition Server Transmits Signal
If the boundaries are crossed, it indicates unusual behaviour has been detected and the recognition server then transmits a signal to remotely lock the device till the user verifies through password or any other form of trusted identity like Touch ID credentials. The server can also send notification to third party device or any other mobile phone owned by a different party.
It is not known if Apple will be adopting behaviour learning features in future iOS versions but the latest iOS 8 beta builds does not currently support these functions Due to concern over the growing data privacy, users would be likely to manually opt in to the feature if the technology were to roll out one day and the breadth of measured behaviours would tend to narrow down.
Behaviour learning patent application of Apple was first filed last January and creditsas its inventors, go to Gregory T. Lydon and Sylvain Rene Yves.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.