Mind-Blowing Ways Facebook Uses Machine Learning

Mind-Blowing Ways Facebook Uses Machine Learning
Mind-Blowing Ways Facebook Uses Machine Learning

What comes in your mind when you hear the word Facebook? Friends, Entertainment, Chats, that’s all we commonly think of, when we use Facebook. Facebook is one big hub of people from different age groups and country.

According to a report, in India, we have a total of 324,000,000 Facebook users in January 2020, which is 23.2% of its total active users. People aged 25 to 35 were the largest user group. Facebook is still growing day by day and the one big technology Facebook use is Machine Learning.

Machine Learning has become the most important aspect not only in Facebook but in many social networking applications that we use and see around us every day. These applications had made most of the work very easy now we can connect to new people, know about what is going on overseas and can share our talent with everyone and help them to grow. We can even start a new business and also let other people join, all these things are possible just because of Machine Learning.

Facebook uses Machine Learning in various ways like Facebook News Feed, Facial Recognition system, Targeted Advertising on your page, recommendations, etc.

Recommendation

Recommendation system of Facebook is one big part where Machine Learning is used, “People you may know” is what we see when we make a new account on Facebook this is where Machine Learning algorithms do their work to suggest new users how to find people that they may know to follow or to make friends.

Machine Learning algorithms take the profile of a user as input and starts matching the profile with the user that have same bio and when few or more things match then it starts recommending those matched profiles to the user

Example: Users-1 have an account and went to certain school for his/her studies and Facebook algorithm takes it as an input and check the profile of all the user who went to the same school for their studies and start showing all the profiles to the User-1, reducing the overhead of the User-1 to find the people that he/she may know.

Face Recognition

Face Recognition is the part of many applications, Google uses Face Recognition in many of its applications but how Facebook use Face Recognition? Whenever someone uploads their photo on Facebook the photo gets converted into strings of number which is different for every face that way Facebook is able to identify the person via his/her photos.

It uses deep learning algorithms that take a photo of the user and scan the photo convert the photo into pixel format from those pixels format that photo get convert into a template which is basically a string of numbers. This template is unique for every face and can be used to detect the face whenever someone uploads the photo with the same template.

blog banner 1

That also allows the user to tag that person without searching out from the list of friends that the user have. Face Recognition plays an important role when it comes to the security of the user, now the question is How Face Recognition is helpful in maintaining the security of a user?

The answer is simple, whenever someone upload photo or video that includes your face, Facebook takes that photo as input to try to understand each face and match them with a template if your face matches with the template then it will send you the notification allowing you to see whether that uploaded photo is appropriate for you or not.

Language Translation

People all over the world use Facebook and sometimes they struggle to communicate because not all the user know how to speak English, to make the Facebook more accessible to all its user from different part of the world Facebook use language translator. Language translation is being used in the Facebook from so long in order to translate the language the user needs to click “See translation” Facebook also provides the option of changing the language all of this is done using Machine Learning.

Facebook translation system uses Statistical machine translation (SMT). SMT analyses the complete document go through each word and produce a dictionary and match it with the already existing pattern and pick the one that makes the most sense out of the document.

Whenever someone clicks the “see translation” button of a certain post then that post content is taken as the input for the translation all those words first get a filter and tokenize, once this process is done then the occurrence of each word is counted and makes the finite dictionary which will be referred to make the meaningful sentences for the user in its native language.

Advertising

You must have noticed that these days whenever you open up your feed and scroll on Facebook many advertisements related to the stuff that you have liked on other applications, for example, you want to buy a smartphone and you checked few phones on Flipkart so now when you open your Facebook feed you may see some phone advertisements related to those some phones you viewed on Flipkart, Facebook is growing in such a rapid rate and now with such a growth many new business and start-ups are also coming on Facebook for their advertisements and promotions all these brands get paid for their work by Facebook.

This advertisement part is done using Machine learning more accurately using deep neural network Facebook collect data of user profile like age, gender, locations where the user has visited, post the user has liked, friends user interact the most, all the applications (partner with Facebook) the user has visited and things user have liked or saved their all this data worked as an input for the deep learning algorithm of Facebook and finally the similar content is shown on the feed of a Facebook user. It has a partnership with many applications and brands it keeps on referring to that current and fresh data for providing the user with the best experience.

News Update

Facebook News Feed is now the part of the life of many Facebook user and really living up to its heights many people gets their regular news updates from Facebook only, according to a survey there are more phone users than those who watch TV for their entertainment and news feeds, Newsfeeds follow the trend that is the post or the news which is viewed the most or have higher likes or comment tend to occur higher on our feed and some of them are not even displayed on our feed all this happen because of the machine learning algorithm that runs behind to provide the most recent news feed.

News post appear on your feed based on major factors which are your interaction with the post that you like all those posts that we see on our feed are the major outcome of machine learning algorithm which keep on reading and learning from our interaction with other user and the posts that we go through on daily bases all these things affect the organisation of feed of certain user and the way they see their post for example: if a user likes to see news of certain place let’s say Delhi then all the other news related to Delhi will be visible to this user feed and the user may be able to get all updates on time.

This is how Facebook is impacting the life of its users. To explore more articles, click here.

By Vikas Upadhyay