Grannies in tech

Nowadays, technology and telecommunications have been very well received by the general public, without exception, and it is of great importance that day by day we put our knowledge into context, just as the technology around us innovates day by day. Today we will have an introduction to machine learning. What is it, how does it work, and who uses it, will be some of the questions in which we will be going a little deeper.

This blog is specially addressed to our grandmothers, those beautiful beings who are like our second mother, who when we were little always wanted to show us the world through their eyes, and today we are the ones who teach from the most basic to the most complex things in the world of technology, Although personally, my grandmother and even my great-grandmother are pro in terms of social networks, they handle chats perfectly, receiving and sending photos and videos, and making video calls with their relatives from other countries, I really feel very proud of the easy adaptation they have to new technologies.

Birth of Machine Learning

Machine learning was born in the 1960s, being an ambitious idea and a sub-discipline of artificial intelligence, a product of computer science and neuroscience. Its study was oriented to data recognition and learning by computers. One of its reasons for its existence was the need for scientists to find a way in which computers could learn only based on data, deepening in pattern recognition (in scientific processes, mathematics, computing, among others).

What is Machine Learning?

Machine Learning is a scientific discipline in the field of Artificial Intelligence that creates systems that learn automatically. In this context, learning is not about memorizing and collecting data, it is about identifying complex patterns from the information provided in order to generate conclusions on copies never seen before. This is achieved through algorithms that review the data and are able to predict future behavior. Automatically, also in this context, it implies that these systems improve autonomously with time, without the human being having to write instructions or codes for this. The main purpose of machine learning is that people and machines work hand in hand.

Many years ago Alan Turing said: “when a computer makes you believe that you are talking or interacting with a human, you could say that this computer is artificially intelligent”.

What is an algorithm?

A sequence or series of instructions, representing the solution to a given problem. For example, what are the instructions or steps to prepare a fruit salad, a possible solution could be the following:

• Go to the supermarket and buy the fruits you want to include.
• Wash and peel the fruits.
• Cut the fruits in squares or portions according to taste.
• Add all fruits in a bowl and stir.
• Serve the desired portion.

In the previous example we have a clear example that is an algorithm in a situation of daily living.

What is Artificial Intelligence?

In 1956 John McCarthy defined it as: “The science and ingenuity of making intelligent machines”.

There are four types of artificial intelligence and these are:

• Systems that act like a human being.
• Systems that think like a human being.
• Systems that think rationally, that is, that imitate people’s logical reasoning.
• Systems that act rationally.

At the same time, artificial intelligence is divided into two schools, the first is dedicated to the study of human behavior and the second is computer-based interactive learning.

Some examples of the application of artificial intelligence are: robots that drive cars, computers that control space travel, systems that report the latest advances in medicine, economics or other sciences, among others.

“Artificial intelligence is the attempt to make a device or an application equal to or more intelligent than a human. Machine learning is a series of algorithms that make your device or application artificially intelligent.” Haydé Martínez — Platzi conf.

What is Deep Learning?

Deep Learning carries out the Machine Learning process using an artificial neural network composed of a number of hierarchical levels. At the initial level of the hierarchy the network learns something simple and then sends this information to the next level. The next level takes this simple information, combines it, composes a somewhat more complex information, and passes it to the third level, and so on.

What is Big Data?

Big data is a term that describes the large volume of structured and unstructured data that floods a company every day. But it’s not the amount of data that matters. What matters is what organizations do with the data. Big data can be analyzed for insights that lead to better strategic business decisions and actions.

Types of Machine Learning

Supervised Learning:

Depends on pre-tagged data, e.g. traffic light labels, street labels, lights, platforms, cars, buses, bicycles, people, animals and other objects that can be seen on the streets. These labels are placed and supervised by human beings assuring the quality of the data. They are already solved problems and the idea is that the computers can feed back from these data and in the future it will be the ones that can differentiate each element.

The most commonly used algorithms in Supervised Learning are:

• k-Nearest Neighbors
• Linear Regression
• Logistic Regression
• Support Vector Machines
• Bayesian Classifiers
• Decision Trees and Random Forest
• Neural Networks
• Deep Learning

Unsupervised Learning

Unlike supervised learning, training data does not include labels and the computer must sort or attempt to differentiate the data itself. You are given a large amount of data with characteristics unique to each object so that it can decide what is according to what you have learned in each feed back.

The most important Unsupervised Learning algorithms are:

• Clustering K-Means
• Principal Component Analysis
• Anomaly Detection

Reinforcement Learning

In this particular case, the basis of learning is reinforcement. The computer is able to learn from trial and error in a number of different situations. Although it knows the results from the beginning, it does not know what the best decisions are in order to obtain them. What happens is that the algorithm progressively associates the patterns of success, repeating them again and again until they are perfected and become infallible.

Uses of Machine Learning

Although in the beginning its functions were basic and were limited to filtering emails, today it can do things as complex as traffic predictions at busy intersections, detect cancer, map sites to generate construction projects in real time, and even define the compatibility between two people. The practical field of application depends on the imagination and the data available in the company. Here are a few more examples:

• Detect fraud in transactions.
• Predict technological equipment failures.
• Predict which employees will be more profitable next year.
• Select potential customers based on behaviors in social networks, web interactions.
• Know when is the best time to publish tweets, Facebook updates or send newsletters.
• Make medical pre-diagnoses based on patient symptoms.
• Change the behavior of a mobile app to adapt to the customs and needs of each user.
• Detect intrusions in a data communications network.
• Decide what is the best time to call a client.

Applications of daily use that count on Artificial Intelligence

Every time in our day to day is more common to find this term that a few years ago we could only associate with science fiction films, in which robots were completely autonomous, and we might think that perhaps this is the only way in which we can see applied artificial intelligence, but the reality is completely different, every day we are in constant use of these technologies without knowing that they are part of what we thought only belonged to fiction, this small list is an example of the various applications of daily use that incorporate this technology:

• Siri: is one of the most popular personal assistants and one of the best examples of speech recognition software. It is an Apple program that offers you its various computers, such as iPhone, iPad or Mac computers. The program automatically links to all your information (messages, calendar, music, reminders, mail, contacts, notes, etc.) and uses machine learning technology to learn and become “smarter.
• Gmail: As in the previous case, Google’s email platform uses machine learning to control spam in your inbox. The system understands and learns from past examples to make future decisions based on them.
• Tesla: one of the technological leaders of the automotive industry. Tesla not only manufactures cars with impeccable and attractive design, but they are also intelligent cars with predictive capabilities. One of the most outstanding features of their cars is their ability to drive on autopilot (without the need for a human behind the wheel) plus a variety of technological innovations that make them extremely attractive to those passionate about the latest technologies and cars. But that’s not all, the automotive company’s system is capable of continuously updating its models through the “cloud”; thus, cars are becoming increasingly intelligent.
• Amazon: the predictions and suggestions made by this website have allowed it to boost its sales in an extraordinary way, being the great giant of e-commerce. More than a third of its sales are due to the recommendations it makes. This technology can know and guess, with great precision, what are your tastes and buying behaviors.
• Google Now: an application that works as your personal assistant, which collects data from multiple Google services, such as: Gmail, Calendars, Google Maps, Google Search, YouTube, etc., to provide you with information that it considers may be useful for you.
• Netflix: taking advantage of predictive technologies (such as algorithms), Netflix is able to recommend movies, series, or documentaries, after analyzing hundreds of records and compiling material similar to what you have seen or rated positively. Also consider factors such as the time and day, to offer content more in line with your consumption habits.
• Google Translate: in order to do its translations, it makes a statistical analysis of the patterns that exist in millions of previously translated documents. By the way, this is possible due to the artificial intelligence it uses. Thanks to that, it is now even able to translate posters or signs that you approach with your smartphone and in a matter of a few years your results will be much more accurate.
• Facebook: the AI is the reason why Facebook presents you with attractive and relevant content in the News Feed, according to your preferences. It does this by analyzing your behavior and participation within the social network, and interprets your interactions (“likes”, “shares”, comments, etc.) as interests, showing you more content similar to what you already liked.
• Google Maps: this map and navigation application uses algorithms to suggest the most convenient routes and means of transport to your destination of choice.
• Spotify: this app uses the AI to identify with the user, generating playlists or mixes daily, according to what you have been listening to recently. He also suggests new artists and albums, which he considers might be to your liking. And since, on an annual basis, it generates statistical summaries of all the user’s activity, you can know for sure what your favorite genres, artists and languages have been, in order to offer music that is more in tune with the user.

https://es.wikipedia.org/wiki/Aprendizaje_autom%C3%A1tico

https://es.wikipedia.org/wiki/Inteligencia_artificial

https://www.elmundo.es/ciencia/2015/12/10/5669a7f122601d71068b464e.html

https://blog.adext.com/cosas-apps-programas-inteligencia-artificial/