The United States Patent and Trademark Office (USPTO) recently published a patent application, owned by Facebook, which could predict the future geographical locations of users. Titled “Offline Trajectories,” the application gathers a user’s previous location data as well as that of others in order to make predictions as to where the user may be traveling next. For example, if the user typically goes to the gym after work and the application determines that the user is currently at their office, it could then infer that the user is likely to go to the gym at a certain time.
More particularly, the patent application discloses a method of determining a current location of a user based on location data and then calculating the probability that the user will transition to each of a number of “candidate geographic locations” within a predetermined time window. This probability is calculated based on previously logged location data associated with not only that user but other users who were at the current location. In other words, the application determines how long individuals usually remain at a given location to predict how long a particular user is likely to remain there. To accomplish this, the computer model implements machine learning and metadata.
The application may also calculate an offline probability representing a likelihood that the user will not have wireless connectivity at one of the candidate locations. Again, the application may do this by comparing data of other users who were previously at a candidate location in addition to the user’s known mobile carrier to determine the probability that the user will lack wireless connectivity. In turn, data may be provided to the user’s device so as to account for this offline probability. As an example, the data provided to the device may be a number of newsfeed items, such as Facebook friends status updates or photo albums, corresponding to the amount of time until the device has wireless connectivity restored. Stated another way, the application would preload Facebook content to keep the user occupied during the time in which connectivity is lacking.
Interestingly, Facebook’s patent application for “Offline Trajectories” received a notice of allowance without any office actions. This is a notable occurrence given the fact that it is highly unusual for a patent application to not be rejected and receive at least one office action from the patent examiner. In fact, typically, inventors receive multiple office actions. In a study tracking over 2 million utility patent applications filed from 1996 to 2013, it was revealed that only about 11% of patent applications issued to allowance without the patent examiner issuing at least one office action requiring any amendment. Overall, during this time, only 55.8% of applications emerged as patents without the use of continuation procedures, which include prosecution procedures that take place after an applicant has received the maximum allotted office actions. This is why it is imperative that inventors engage in a cost-benefit analysis balancing the cost of patent prosecution and the potential benefit if the patent ultimately issues prior to filing an application.
While Facebook has not revealed why this data would be useful to their company, it seems, in all likelihood, the social networking company will use this data to display targeted advertisements specific to a user’s predicted destination. That being said, as with any patent application, there is an equal likelihood that the invention may never come to fruition. Indeed, a Facebook spokesperson confirmed that the company often seeks patents for technology that is never implemented and that its patent applications “should not be taken as an indication of future plans.”