Graph Search to continue using natural-language processing


As Facebook gives a taste of the Graph Search to a few million from a whooping 1.06 billion active users who had signed up for the social networking website by the end of 2012. The Graph Search is a feature that will replace the search bar. It boasts of finding people, places and things for you. Offering more than the taskbar that Facebook is currently giving to most of its users, the Graph Search often comes up with more interesting and sometimes surprising content when you type something in it. Little thought is given to how many wheels might be turning behind the Graph Search that finds everything so easily.

To make input by the user more effortless, natural-language processing is used. This only works with English now but engineers are hard at work to make this product available in other languages as well. Maxime Boucher, a research scientist and Xiao Lee, an engineering manager on the natural-language processing team in Graph Search posted in Facebook’s engineering blog on Monday saying that the product uses natural-language processing to make an estimate of what the user wants.

Open Graph Natural Language Processing

How does the Graph Search Work?

Graph Search breaks down search strings into multiple components that serve as commands with which the system can access the database. For example, if a user is searching for a friend who lives in Egypt, they might type in the Graph Search ‘my friend who lives in Egypt’. the search will pull up the user, grab their list of friends, call on the filter for the people who have Egypt in that field. Graph Search considers that search query “intersect (friends (me), residents (12345)).” The query of the user is converted into something what is easy to digest for the Unicorn search engine.

The Graph Search analyzes the words that have been typed in the bar by the set to look for possible results across more than 20 categories such as cities, employers and schools. Based on probability and statistics, the Graph Search chooses options that are most likely what the user wants. For instance, if ‘San’ is followed by ‘Diego’, the user is most likely to be looking for.

Importance of Search terms

Some users tend to include such terms in their search that may be of no use to the Graph Search. Sometimes, keywords that are supposed to act as the guiding light for the Graph Search are missing. In this case, the graph search does tend to give you options in order to guess what you were looking for. Even though the Graph Search will be going to help many a Facebook spy, it is still not available to a majority of users.

Users may also type the keywords in the wrong sequence or order. Nonetheless, the graph search will provide the user with several options but the results may not be that accurate. This simple interface helps people to search for hundreds and millions of places, objects and other people. This will, however, require engineers to produce a number of systems and models to bring the Graph Search to the home page of all the users.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.