Select Page

Machine Learning: an overview


If we are talking about robots that speak like humans, algorithms that predict our behavior or cars that drive alone, it may seem that we are referring to our own technology of the future. Behind these examples are the Artificial Intelligence and Machine Learning, disciplines that have been with us much longer than we imagine.

We would have to go back to the middle of the last century to talk about the origin of Artificial Intelligence. With the 'Turing test'This new discipline began to be considered and, therefore, many people consider Alan Turing as the creator of Artificial Intelligence. This Test consisted of the creation of a computer that should be able to impersonate a human in a chat conversation with another.

This concept that Turing raised has evolved by leaps and bounds thanks to the advancement of technology and, both Machine Learning and Artificial Intelligence, are disciplines that are very present in any aspect of our daily lives. These algorithms are used, for example, to recognize our face when unlocking the mobile or to create intelligent assistants such as'Crab' or 'Alexa'.

Now, what makes Machine Learning such an important discipline today? If we want to answer this question, we must first ask an even simpler one: What is Machine Learning?


Definition of Machine Learning

Next, we are going to introduce ourselves to the concept of Machine Learning, but for this, it is very important to understand its differences with other disciplines that we will also define, such as Artificial Intelligence and Deep Learning.

Machine Learning, or machine learning, is a scientific discipline that comes from Artificial Intelligence, whose objective is make machines or computers learn autonomously thanks to the use of algorithms.


These algorithms are responsible for identifying patterns in data sets and make predictions. In this way, algorithms improve over time without the need for human intervention as they learn from the information they process.

On the other hand, we could define the Artificial Intelligence (AI) as the ability of machines to imitate, reason or act like people.


Artificial Intelligence and Machine Learning are concepts that we should not confuse. Therefore, Machine Learning was born from the search for Artificial Intelligence, but with the purpose of making the machines themselves learn.


Another concept to keep in mind when we talk about machine learning is that of Deep Learning. Deep Learning is, in turn, a branch of Machine Learning whose objective is make machines work similar to the human brain using artificial neural networks. These neural networks are made up of layers of neurons and their operation is based on that of a biological neural network.

In future posts we will learn more about the concepts of Deep Learning y Neural networks and its importance in the development of machine learning.


Types of Machine Learning

Now that we know the concept of Machine Learning, we are going to classify the types of machine learning according to the way we 'teach' our algorithms:

  • Supervised learning: is the technique in which we provide the model with a set of data that have been previously labeled or classified and that include the expected result. the objective of these algorithms is to determine the desired output based on the input received.

We are going to see an example to understand it in a simpler way. Imagine that we want to train our model to learn to classify photos of cars and motorcycles. To do this, we will offer a set of images with the correct result in each case. In this way, the algorithm will be able to differentiate the image of a car from that of a motorcycle thanks to its previous training.

  • Unsupervised learning: It is the type of learning in which data is not entered with labels and we do not provide the expected output as we did in the previous type. With this technique, the algorithm will look for patterns in the input data to produce knowledge.

Let's see with the previous example how we would now train our algorithm. In this case, we will introduce the photos of cars and motorcycles into the model without giving any information. The algorithm will try to find patterns and group the photos according to its own criteria, since we have not "helped" it with the expected result in each case.

  • Reinforcement learning: it is a method similar to supervised learning in which, instead of correcting the algorithm with the appropriate answer, we will limit ourselves to indicating whether the system has been right or wrong in its response.

Imagine that now we want to classify images of cars, airplanes and ships. When the algorithm processes the photos and determines your answer, we will only tell you if you have succeeded or failed. For this example, if the model has mistaken a car for a boat, we will not give it the appropriate solution, as we would in the supervised one, but our goal is for it to learn with as little information as possible.


To learn more about the different Machine Learning systems and their classification, you can read the article: «Learning about Machine Learning systems », Antonio Mendez.


Machine Learning Applications

The application of these machine learning algorithms has become fashionable in all sectors of society and, as we mentioned in the introduction, more and more uses are being made of this technology.

It is surprising how easily we can find examples in any field. For example, if you came to this post through a search engine, you should know that companies like Google or Yahoo use powerful algorithms to learn from users and their search intentions in order to always offer the best results. You have probably also used the quick response suggestions when you receive an email, or the word and sentence recommendations as you write a new email.

These systems can also be applied for analysis of large volumes of customer data. The greatest advantage of these algorithms is that they can be used for different purposes within the same industry and company.

Machine Learning can be used to detect patterns in customer behavior when making a purchase and also to improve your shopping experiencea analyzing your tour of the store or to improve customer service with the use of chatbots that learn natural language.

As you have seen, it is fascinating to discover the immense number of examples in which these algorithms are present.


What are the advantages of applying Machine Learning in the company?

This branch of Artificial Intelligence is gaining a niche in any company both for the process automation as well as to improve decision making. Thanks to Machine Learning, there are many companies that are already taking advantage of these algorithms to obtain competitive advantages, but do you think it is too late to start using this technology in your company?

Without a doubt, the answer is: 'No.' Despite experiencing a true boom in this discipline, the opportunities it can offer are still immense. If Machine Learning systems are not yet being used in your sector, implementing them can be a huge competitive advantage. Is he recent Tesla case, who is making huge strides towards the goal of 'total autonomous driving', using machine learning to train your neural networks.

In short, the application of these algorithms allows improve existing processes in any company in order to optimize them. In Panel we have a team of experts who will accompany you in the implementation of this type of techniques that will optimize and strengthen your business processes, helping you in the use of new tools and skills that will increase your analytical capacity in real time, enhancing the taking of decisions.

If you want to know more in detail about our team and our project, we encourage you to know what we do in Panel at Data & Analytics Services and Solutions.


Daniel Checa

Daniel Checa

Daniel Checa He is a collaborator in the Panel's Data & Analytics area. You can contact him via email here

Leave us your comment


Send a comment

Your email address will not be published. Required fields are marked with *

Share This