Home » Machine Learning » Facebook Machine Learning Platform

Facebook Machine Learning Platform

For obvious reasons, Facebook uses machine learning. Otherwise, the 2.5 billion users would not benefit from features like “people you may know’ or suggested ads reach targeted users. It is okay to say Facebook is a global phenomenon rather than a social network. It is one of the biggest users of recommender systems.

How Facebook Uses Machine Learning?

UsageDescription
Facial RecognitionFacebook facial recognition helps users to know when their photo is uploaded to the platform, it helps tag suggestions to people with a similar facial fingerprint. It uses a facial recognition algorithm that analyzes the pixels of the face in the image. It can also help to inform people with visual impairments to know whenever their photo has been uploaded.
Textual AnalysisTexts just like photos are important on Facebook. Facebook uses Deeptext to interpret the message in the text and manage them. Deeptext is an engine in deep learning that has the capability of learning tons of posts within a second and can understand the posts in more than 20 languages using the most accurate manner. A good example is how Deeptext has to learn the meaning of the difference between apple the fruit and apple the company in a post.
Targeted AdvertisingFacebook uses deep neural networks to study user’s habits from previous sites like eCommerce platforms and make similar suggestions of products and services on their Facebook wall. Deep neural networks analyze gender, age, location, interests, page likes et cetera to profile their users into categories for targeted ads.
Language TranslationFacebook is a worldwide obsession. This means it is in all parts of the world including those countries that don’t speak English. A Facebook translator is used to translate more than 111 languages worldwide. Machine learning algorithms analyze the millions of already translated documents on the internet to find patterns and vocabulary; it then picks the most accurate answer to show translation.ML keeps updating to pick new slang and sayings.

Topics in Machine Learning

Hits: 13