Machine Learning for Baby boomers: Basic Concepts
A few years ago, remote communication was complicated because you had to go to specific places to make international calls or you had to send a letter and wait for it to reach the recipient and for him or her to reply, which could take days or even weeks. All of the above has changed thanks to the internet era along with technological evolution that has occurred during the last 20 years, without a doubt, has meant changes in people’s lifestyles, for example, now communicating with a person at a distance is just a click away, you can make a video call and see the person in real time.
Taking into account the above, as technology was advancing man realized that he could teach machines how to think, that is, automate the processes and decision making according to certain instructions that the human could provide to the machine, this is known last is given through algorithms. Years later this teaching process was given the name of machine learning. But let’s not focus on technical terms, let’s understand these terms through clear and real examples of how all this works.
Let’s suppose you are watching videos on Facebook about how to cook, specifically how to make cakes, pastries, and lasagna. At the end of a video, you will be able to see videos similar to the ones you have already seen, and how you like to watch cooking videos you keep clicking on. Then later when you enter the application again and see pictures of your acquaintances you will see cooking videos at some point. You will think… “Oh Facebook you know well, shows me the videos that I really like”, but do you know why Facebook recommends those videos? Basically it is because of what was mentioned before, the human managed to teach machines to think and consequently to make decisions, this is machine learning.
Imagine that the algorithms are like children, initially they don’t know anything but as they grow up they learn new things and are able to relate one thing to another. For example, if the child sees that the sky is dark, it is very likely that it will rain, or on the other hand, imagine that your grandson knows that you like chocolate cake. Probably because he has heard that you like this kind of cake or because you simply told him, so your grandson learned that information.
Likewise, the algorithms need to be educated. And how is this done? By providing them with data, and from them they make decisions, this is known as supervised learning. So, every time you watch a video and you like it or make a comment, you are teaching the algorithm what your preferences are and in this way you are learning from it. However, there is something curious that is important to mention, Facebook is composed of users and you are one of them. Just like you, there are users with a taste for cooking and like to watch this kind of videos, so Facebook makes recommendations based on the videos that other people with the same preferences as you have watched. In other words, Facebook is in charge of recommending what you really like. In this video you will see an explanation of how Netflix and Amazon use this methodology to recommend movies but the principle behind it is the same, machine learning. Going deeper into the above, take a look at the following image to further understand how the process works graphically.
As you can see there are two types of filters, on the one hand collaborative filtering and on the other hand content-based filtering. Social networks such as Facebook use these types of recommendations based on user behaviour and behind it there is a whole machine learning process.
With this in mind and going back to the example, suppose now that Christmas is approaching and you want to start making crafts, then look for videos about this. What are you really doing? Teaching that child called an algorithm to have more data, so he can predict what you like. Now she’ll not only recommend cooking videos but also watch craft videos for Christmas. Surprised? This is the true power of data and it is all thanks to the advancement that has been made with the machines.
Meanwhile, when you browse the Facebook application and find photos of family members you can see that a white box is drawn on the person’s face (SEE PHOTO 1) and you can see the name of the person. Why, what’s really going on there? In the field of technology it is known as facial recognition. This is nothing more than a model of trained algorithms allowing the identification of a person by analyzing the biometric characteristics of their face, things like the distance between the eyes or the size of the head, this way, the Facebook algorithm for facial recognition can be the name of the person according to the photo you are looking at.
Just like this simple example on Facebook there are many machine learning applications. Another case is probably that you have heard it and it is voice recognition, maybe you have done a search with the famous “Ok Google”, Siri or Alexa. Asking about any medication or weather conditions in the next few days. The main point is that you can also educate the machines through voice, as they can receive instructions and respond immediately. In this video you will see how an electronic device in the form of a speaker uses voice recognition to perform instructions that you tell it, it is a whole process of automation, that is, it can give you the indication to turn on the lights, remind you about the time you should take your medicine, even play a specific song, among other things.
I would like to conclude that there are many machine learning applications using approaches such as supervised and unsupervised learning that learn through data that humans can provide, so the more data you learn the more predictive you are able to be. In this article, basic concepts were addressed through real-life examples of machine learning applications in different contexts to understand the potential of machine education.
REFERENCES
[1] “How Recommender Systems Work (Netflix/Amazon)”, Youtube.com, 2020. [Online]. Available: https://www.youtube.com/watch?v=n3RKsY2H-NE.
[2] “What’s Going On With Facial Recognition? | Untangled”, Youtube.com, 2019. [Online]. Available: https://www.youtube.com/watch?v=BqQT4sIOYA0.
[3] “Amazon Echo Plus”, Youtube.com, 2017. [Online]. Available: https://www.youtube.com/watch?v=9qkW75JsY3U.
[4] A. Mello, “How do Netflix and Amazon know what I want?”, Medium, 2020. [Online]. Available: https://towardsdatascience.com/how-do-netflix-and-amazon-know-what-i-want-852c480b67ac.
