How Machine Learning is used by Famous Companies
Machine Learning has become the technology of the future! Some people believe this technology will end the world. Others think it can make our lives easier. It is not surprising that nearly all companies use this technology to attract customers and provide personalized customer experiences. There has been a 270% increase in companies that have adopted ML in the past four years.
It is easier for large tech companies to invest in Machine Learning or Artificial Intelligence. This tutorial will focus on the fascinating ways ML is used in companies such as Google and Pintin. Let's take a look at these companies and the different methods they use in Machine Learning.
Instead of asking, "Which Google apps use ML?" we should ask, "Do any Google Applications not use ML?" The answer is probably no! Google has a lot of money in Machine Learning Research and plans eventually to integrate it into all its products. Google's flagship products, Google Search and Google Translate use ML currently.
Google Search uses RankBrain, which is a deep neural net that assists in providing relevant search results. RankBrain uses intelligent guesses to determine if our search is for "Tim Cook" and if there are unique words or phrases in Google Search. However, Google Translate analyses millions of documents and is able to identify the most common patterns and vocabulary. Google Photos uses image recognition.
Google Photos uses image recognition. Deep Learning is used by Google Photos to sort millions upon millions of images online in order to classify them better. Google Assistant uses Image Recognition and Natural Language Processing, which allows it to be multi-talented and answer our questions.
Facebook is where we should go if we want to see our friends, watch celebrities, or look at cat photos. Facebook has 2.41 billion Monthly Active Users! Machine Learning is the only way to achieve this level of popularity. Facebook uses Machine Learning in all aspects of its News Feed, including Targeted Advertising.
Facebook utilizes Facial recognition to recognize our friends and suggests their names. A Machine Learning System analyses the pixels of an image to generate unique templates for every face. The facial fingerprint can be used to identify the face and suggest tags.
Targeted advertising on Facebook uses a deep neural network to analyses our location, age, gender, page likes, and interests, to identify users and show them ads targeted at these groups. Facebook now uses chatbots to provide human-like customer service interactions. These chatbots interact with users using ML and NLP and look almost human-like.
Twitter is the best place to find interesting tweets, intelligent debates, and more! Twitter is the best place to find out about current politics, global warming dangers, and smart comments from celebrities. Guess how all those tweets are managed? Machine Learning is the answer!
Twitter uses an ML algorithm for organizing our tweets. Tweets based on what we like and tweets from family and friends will be given a higher priority and appear higher in our feed. Tweets that receive a large number of retweets or likes will have a higher chance of getting noticed. These tweets can be located in the "In case you missed it" category. The tweets were previously arranged in reverse chronological order. This is what some people want back. Twitter currently uses the Natural Language Process capabilities by IBM Watson to find and delete abusive tweets.
Twitter uses deep learning to determine what's happening in the live feed. This is achieved by training the neural network using tags to recognize images in videos. Let's say we add the tags "Puppy", "Animal", "Poodle", "Husky" etc. The algorithm will identify a dog in our video and use that information to identify other dogs in our videos.
Baidu Google for China! While this may not be the case, Baidu is the Chinese Search Engine most often in comparison to Google. Like Google, it utilizes Machine Learning in many of Baidu's applications, such as Baidu Search, as well as DuerOS, Baidu's assistant for voice. The Xiaoyu Zaikia home robot, similar to Alexa, is also used.
Service. Baidu's Search Engine is the main focus, as 75% of Chinese use it. Machine Learning Algorithms (HMLA) are used to recognize images and voice recognition. This allows for the best possible (and also smarter!) service. Baidu also made significant investments in natural language processing. This is evident in DuerOS.
DuerOS Baidu's Voice Assistant makes use of natural language processing, image, and voice recognition to build an intelligent system that is able to have an entire conversation while sounding human. The voice assistant makes use of ML to understand the complexity of human speech and duplicates it in a flawless manner. Baidu's NLP expertise is also applied to the Little Fish home robot, similar to Alexa but different. It can turn its head in order to "listen" to the voice coming from the other direction and then respond accordingly.
The users might have heard of Pinterest, whether they are a regular pinner or a beginner. Pinterest allows us to pin images, videos, and GIFs we are interested in. Since this app relies on images being saved from the internet, it makes sense that its most important feature is to identify images.
Machine Learning is the answer! Pinterest uses Image Recognition algorithms for identifying patterns in images we pin so similar images can be displayed when we search them. Imagine we pin a green shirt. We will be able to view images of similar green shirts by Image Recognition. Pinterest can't guarantee that these green shirts will be fashionable!
Pinterest offers more personalized recommendations based on our Pinning history. This is in contrast to ML algorithms used for social networking apps that also consider our age, gender, and friends.
Next TopicIntroduction to Transfer Learning in ML