Case Study Of Machine Learning

Introduction Of Machine Learning

Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. Machine learning focuses on the development of computer programs that can access data and use it learn for themselves.

The process of learning begins with observations or data, such as examples, direct experience, or instruction, in order to look for patterns in data and make better decisions in the future based on the examples that we provide. The primary aim is to allow the computers learn automatically without human intervention or assistance and adjust actions accordingly.

Types of machine learning Algorithms

• Supervised machine learning algorithms can apply what has been learned in the past to new data using labeled examples to predict future events. Starting from the analysis of a known training dataset, the learning algorithm produces an inferred function to make predictions about the output values. The system is able to provide targets for any new input after sufficient training. The learning algorithm can also compare its output with the correct, intended output and find errors in order to modify the model accordingly.
• Unsupervised machine learning algorithms are used when the information used to train is neither classified nor labeled. Unsupervised learning studies how systems can infer a function to describe a hidden structure from unlabeled data. The system doesn’t figure out the right output, but it explores the data and can draw inferences from datasets to describe hidden structures from unlabeled data.
• Semi-supervised machine learning algorithms fall somewhere in between supervised and unsupervised learning, since they use both labeled and unlabeled data for training – typically a small amount of labeled data and a large amount of unlabeled data. The systems that use this method are able to considerably improve learning accuracy. Usually, semi-supervised learning is chosen when the acquired labeled data requires skilled and relevant resources in order to train it / learn from it. Otherwise, acquiring unlabeled data generally doesn’t require additional resources.
• Reinforcement machine learning algorithms is a learning method that interacts with its environment by producing actions and discovers errors or rewards. Trial and error search and delayed reward are the most relevant characteristics of reinforcement learning. This method allows machines and software agents to automatically determine the ideal behavior within a specific context in order to maximize its performance. Simple reward feedback is required for the agent to learn which action is best; this is known as the reinforcement signal.

Ways Facebook Uses Machine Learning

What comes first to mind when you think about social networking? It’s Facebook of course! There is even a movie called The Social Network that proves this statement! And with 2.41 Billion Monthly Active User in the second quarter of 2019, it’s safe to say that Facebook is actually not even a social network but a global phenomenon. And obviously, Machine Learning is a vital aspect of Facebook. It would not even be possible to handle 2.4 billion users while providing them the best service without using Machine Learning!

Let’s take an example. It is mind-boggling how Facebook can guess the people you might be familiar with in real life using “People You May Know”. And they are right most of the time!!! Well, this magical effect is achieved by using Machine Learning algorithms that analyze your profile, your interests, your current friends and also their friends and various other factors to calculate the people you might potentially know. And that’s only one aspect in which Facebook uses Machine Learning! Other aspects are the Facebook News Feed, Facial Recognition system, Targeted Advertising on your page, etc.

1. Facial Recognition

Facial Recognition is among the many wonders of Machine Learning on Facebook. It might be trivial for you to recognize your friends on social media (even under that thick layer of makeup!!!) but how does Facebook manage it? Well, if you have your “tag suggestions” or “face recognition” turned on in Facebook (this means you have provided permission for Facial Recognition), then the Machine Learning System analyses the pixels of the face in the image and creates a template which is basically a string of numbers. But this template is unique for every face (sort of a facial fingerprint!) and can be used to detect that face again in another face and suggest a tag.

So now the question is, What is the use of enabling Facial Recognition on Facebook? Well, in case any newly uploaded photo or video on Facebook includes your face but you haven’t been tagged, the Facial Recognition algorithm can recognize your template and send you a notification. Also, if another user tries to upload your picture as their Facebook profile picture (maybe to get more popular!), then you can be notified immediately. Facial Recognition in conjugation with other accessibility options can also inform people with visual impairments if they are in a photo or video.

2. Textual Analysis

While you may believe photos are the most important on Facebook (especially your photos!), the text is equally as important. And there is a lot of text on Facebook!!! To understand and manage this text in the correct manner, Facebook uses DeepText which is a text engine based on deep learning that can understand thousands of posts in a second in more than 20 languages with as much accuracy as you can!

But understanding a language-based text is not that easy as you think! In order to truly understand the text, DeepText has to understand many things like grammar, idioms, slang words, context, etc. For example: If there is a sentence “I love Apple” in a post, then does the writer mean the fruit or the company? Most probably it is the company (Except for Android users!) but it really depends on the context and DeepText has to learn this. Because of these complexities, and that too in multiple languages, DeepText uses Deep Learning and therefore it handles labeled data much more efficiently than traditional Natural Language Processing models.

3. Language Translation

Facebook is less a social networking site and more a worldwide obsession! There are people all over the world that use Facebook but many of them also don’t know English. So what should you do if you want to use Facebook but you only know Hindi? Never fear! Facebook has an in-house translator that simply converts the text from one language to another by clicking the “See Translation” button. And in case you wonder how it translates more or less accurately, well Facebook Translator uses Machine Learning of course!

The first click on the “See Translation” button for some text (Suppose it’s Beyonce’s posts) sends a translation request to the server and then that translation is cached by the server for other users (Who also require translation for Beyonce’s posts in this example). The Facebook translator accomplishes this by analyzing millions of documents that are already translated from one language to another and then looking for the common patterns and basic vocabulary of the language. After that, it picks the most accurate translation possible based on educated guesses that mostly turn out to be correct. For now, all languages are updated monthly so that the ML system is up to date on new slangs and sayings!

4. News Feed

The Facebook News Feed was one addition that everybody hated initially but now everybody loves!!! And if you are wondering why some stories show up higher in your Facebook News Feed and some are not even displayed, well here is how it works! Different photos, videos, articles, links or updates from your friends, family or businesses you like show up in your personal Facebook News Feed according to a complex system of ranking that is managed by a Machine Learning algorithm.

The rank of anything that appears in your News Feed is decided on three factors. Your friends, family, public figures or businesses that you interact with a lot are given top priority. Your feed is also customized according to the type of content you like (Movies, Books, Fashion, Video games, etc.) Also, posts that are quite popular on Facebook with lots of likes, comments and shares have a higher chance of appearing on your Facebook News Feed.

5. Facebook use it to inform ad delivery

To find the estimated action rate, machine learning models predict a particular person’s likelihood of taking the advertiser’s desired action, based on the business objective the advertiser selects for their ad, like increasing visits to their website or driving purchases. To do this, our models consider that person's behavior on and off Facebook, as well as other factors, such as the content of the ad, the time of day, and interactions between people and ads.

• Examples of behaviors on Facebook that the models consider include things a person does while using Facebook apps, like clicking on an ad or liking a post.
• Examples of behaviors off Facebook that the models consider include things a person does outside of Facebook that businesses share with us via our Business Tools, like visiting a website, purchasing a product or installing an app.

To generate an ad’s quality score, our machine learning models consider the feedback of people viewing or hiding the ad, as well as assessments of low-quality attributes (like too much text in the ad's image, sensationalized language or engagement bait).

The advertiser’s bid, the estimated action rate and the ad quality score are combined to

Looking to the future

Deep Learning is likely to continue to play a key part in the future development of Facebook. Although it is tight-lipped about potential new applications at the moment, ideas which have been suggested include automatically generating audio descriptions of pictures to assist the visually-impaired, and to predict where greater coverage is required in its mission to roll out internet access to poorly served parts of the world. In the long-run work of their well-resourced AI and Deep Learning labs is likely to provide benefit to countless other organizations too, either directly through use of their services, or indirectly thanks to their support of open source principles.