Facebook Builds Search-Related Patent Portfolio
The U.S. Patent and Trademark Office recently granted Facebook a patent for incorporating variable values into textual content. In addition to comments in text, it also includes video, articles, images, music files and locations. One patent claim describes the ability to identify context in one language and translate it into another, based on location.
The social network's patent portfolio continues to gain more search- and ad-serving-related patents that rely on metadata, as well as news feeds.
Social sites are paying close attention to metadata in a variety of media, as comments, content and context become more important in ranking search queries in Graph Search and matching ads to demographics and psychographics. Patrick Giblin, co-founder at 451degrees, said the U.S. Patent and Trademark Office recently granted his startup two patents that would support better matching for tools like Graph Search, as well as services like Google+ and Twitter.
451degrees' patent technology relies on metadata from text in comments and articles, music, videos and other content to generate search query results and match ads to member actions. It aggregates comments and content posted, building on the number of times the community searches for specific words in context.
Aside from text, it aggregates metadata in video, image, articles and music files, processing it through algorithms to improve ad targeting. The technology relies on linguistics to read the entire sentence in the comment, making sense of the meaning, along with sentiment to optimize keywords.
The technology can translate the metadata in real time to match the correct search terms, as well as target ads. The ranking of content and ads change as a breaking news story moves from a car chase to murder by gunshot wounds.
An application within video, for example, allows for real-time user-generated meta-tagging in clips. The meta-tag becomes searchable via a "spider network" that runs invisible behind Web pages of the site. These tagged words are broadcast on delivery networks, such as live feeds from SMS, MMS, news feeds within community Web sites, and video sharing Web sites dropped in real time to identify users and friends.