How Facebook uses Machine Learning?

Facebook is much bigger than you think. Where the Android Operating System runs on 12-15 million lines of code, alone Facebook runs on 62 million lines of code, where backend code is not even included. So, it is tough for a normal human to manage Facebook and machine learning helps it in doing so.
In the following article, I am going to tell that how Facebook uses Machine Learning. Let's go deeper.

Machine learning is a transformative way by the use of which a program learns to work in a more better way by its previous mistakes. Facebook uses Machine learning everywhere. Either you are scrolling the news feed, browsing the images or videos or playing games on Facebook you have been a part of an optimization technique, and that is Machine Learning.

The Power of Machine Learning by Facebook

For instance, if you upload an image at Facebook with your friends or simply with pretty background, then you might get a suggestion whether you want to tag other person or not with this image. Here you don’t need to seek that person over Facebook, Facebook will automatically do it for you. Facebook can also tell about a person is happy or sad . Either he or she is standing, sitting or even any other type of activity can easily be distinguished with the help of AI or ML. To accomplish the task, Facebook trained a deep neural network on tens of millions of photos. Facebook’s fortunate in this respect because it is already hosting billions of captioned images.
Theoretically, this type of learning is called Supervised Learning, which is a part of Machine Learning.

Facebook uses Machine Learning to sort News Feed

Facebook also use AI, in sorting the content of your feed automatically on basis of priority.
Here is the explanation.
Have you ever talk to someone seamlessly or go through his/her news feed regularly? I f your answer is yes then you must sure seen or noted something very usual on Facebook. The person would always give the high priority in every aspect i.e., that person would always be on the top, even top on your tag list. Facebook thought that you two are special to each other. Isn't that interesting?

Facebook uses Machine Learning to analyze Text

Facebook is spammed with a lot of text, and with billions of textual information posted on Facebook every day. In this case, something has to be out there supervising what is being written, after all Facebook is a social-platform and it is necessary to stop the spread of Fake News and controversial topics. For this, Facebook developed a textual analysis tool called Deep Text to deconstruct words in a text and discern the relationships between words. This model analyzes words contextually, determining the meanings of words from how they are used and then Facebook allows them in its platform.

Facebook uses Machine Learning for Translation

Facebook has over 2 billion users, and everyone don’t speak English. So to stay true to its goal of breaking down conventional barriers of social interaction, Facebook built a machine learning system that translates text to your preferred language. For this they use Facebook deep learning translation tool.

The End Notes

Facebook is a platform for connecting people in different parts of the world using images, videos, and text. With more than 2.41 billion unique visitors to the platform every month. So, from suggesting an article to a video, everything is done by using Machine Learning.
That's all from my side. If you know something else, please do comment below.

Post a Comment