San Francisco– Despite facing flak for leakage of personal data of millions of its users in recent times, Facebook is seeking to patent a software that could help it build profile of an user’s household – the number of people in the household, the interests that they share, nature of their relationships or even the devices that they use.
The software, which could be used to target ads, would analyse images posted to Facebook or Instagram, The Los Angeles Times reported on Friday.
An online system that predicts household features of a user — household size and demographic composition — provides improved and targeted content delivery to the user and the user’s household, according to the patent application.
To help determine whether people live in the same home, the software could look at how often people are tagged in pictures together and at the captions of the photos, it said.
“Without such knowledge of a user’s household features, most of content items that are sent to the user are poorly tailored to the user and are likely ignored,” said the patent application, which was filed last year and made public on Thursday.
Facebook could also incorporate “past posts, status updates, friendships, messaging history, past tagging history” and web browsing history to put together a profile of a household or family, the report added.
The proposed online system seeks to apply one or more models trained using deep learning techniques to generate the predictions.
“For example, a trained image analysis model identifies each individual depicted in the photos of the user; a trained text analysis model derive household member relationship information from the user’s profile data and tags associated with the photos,” stated the application.
Those profiles, in turn, could be made available to third parties that want to target “content” to users, it said.
Facebook told The Los Angeles Times that applying for the patent does not necessarily mean it will build or use the software.
Around 29 million Facebook accounts were hacked in September. (IANS)
