Facebook is working to improve its search and ad-ranking algorithms, as evident of several descriptions on its job site that provide insight into the type of algorithms the company is building.
While today its ranking algorithms for search and ads pull in signals from queries, searchers, and content, the company says it is investing in nearly every area that supports the ability for its engineers to better understand how natural-language processing determines the ranking of Facebook's graph, content and user engagement.
The project for search rankings focuses on trending topics; artificial intelligence; detection and ranking; indexing for people, entities, posts, photos, and videos; data analysis; and natural-language processing to better understand the query and documents.
The goal of the search-ranking team is to "assist the users to complete their intents, and to provide the most relevant and personalized set of results for these intents." Collectively, Facebook members search about 2 billion times daily for more than 1.5 billion people and trillion pieces of shared content such as posts, photos, videos, and links, according to the company.
Facebook's emphasis on ad ranking is similar to search ranking with a focus on image and text recognition, user and ad modeling, conversion modeling, and modeling new optimization objectives.
In late July, Facebook CEO Mark Zuckerberg called out three stages of the company's search strategy, saying that the third stage will focus on advertising the company will make from the billions of searches.