Simple Notion about Machine Learning
At the beginning, man has sought ease in his daily tasks for what he has so far built and been the basis of modern society, research. Over time, he has developed tools that make work easier, making it less complex and more productive. Leading him to do more in a short time industrializing society with great technological advances. Starting with the creation of the wheel transformed over time in cars, airplanes etc.
The research carries with me not only technological discoveries but also in different areas, equipping us with more advanced and even automated tools.
To understand maching learning it is necessary to create simple analogies. Industry can be a clear example of technological advancement. The industry is made up of companies that offer goods and services and make exchanges between them and society and even compete with each other. To start it is necessary to know how they are formed and how they perform.
Companies are structured in areas, which work with the same purpose but each with a different approach. At the beginning the owner must distribute the time and effort to perform different activities, such as keeping accounting, buying raw materials, selling it, looking for clients, etc. as the owner grows, he will have so many things to do that he could lose the business if he continues to do all the same, so it is necessary to hire more people, with whom he will divide the work thus with the passage of time the successful company, must hire more person and with better preparation, so it will be necessary to be structured in distributed efforts, being now more efficient but always with the same objective.
However, in the same way that the amount of personnel grows, as well as the payment of salaries. People have basic needs that must be met and even weaknesses that the industry harms, such as fatigue, leisure time, and not everyone works at the same pace with different margins of error. So the industry uses technology to automate certain processes, making them more efficient and effective.
Said earlier, technology evolved by making the tasks easier to the point that the machines have come to perform tasks that people used to do. Thus reducing the hand of man in certain processes. By involving the machine the margin of error is minimized and the time is optimized. At present, even administrative processes are being automated. Every process within the company must be reported; So, long ago everything was done in writing, eventually, the typewriter arrived but still, each document had to be stored and classified in folders and according to the level of importance a place was delegated for the type of information and while another It was discarded.
With the arrival of the computer, each document began to be digitized and stored on discs, also bringing different forms of storage and a new form of communication. The sending of a letter, memorandums etc that was previously done by sending a letter by mail which took several days, came in a click.
Communication between people became closer and easier no matter where in the world you are and even more with the arrival of smartphones (a computer in your pocket) expediting the management of processes, such as negotiations, shipments of location documents for personnel control. Although not only companies also benefit people in general, anyone who owns a smartphone.
Said above, the industry is the simplest analogy to start addressing the theme of machine learning.
Given the introduction to the benefits within the industry, it is time to get more into the technological part. One of the industries in which automation is common through the use of robots is car assemblers.
“The assembly of cars travels stations where either a person or a robot is responsible for a specific process, this depends on the number of units to be manufactured” — car factory wikipedia — https://es.wikipedia.org/wiki / Manufacturing% C3% B3n_de_autom% C3% B3viles
These robots that at the beginning were only tools, began to change according to the foolishness of the company and for each one, there is a different one. Initially they were simple and mechanical, however, in the search, leaving aside the intervention of man in the processes and with the arrival of the microcontrollers (circuits) with which it is possible to give them written orders to do a specific job, which It is called an algorithm, something like programming on a sheet, what you plan to do during the day but more meticulous and in the language of the chip. Instructing you to do as you finish another. This principle is used in computing. It is the way in which robots are programmed, telling them what to do, step by step. For what each activity is already established, he will not do more than what is requested, otherwise, there is something wrong with the instructions. This is the same as telling a person what they should do, according to their position. Unlike robots, he will never go to the bathroom, he will not take rest, he will have no worries that the performance will come down. However, if he does something wrong he will not repeat the process without thinking if what he did is right or wrong, as it is an uninterrupted process, he will continue to the next process until he completes all the instructions. This is part of one of the disadvantages of the robots, if it is wrong he does not learn to later not repeat it again if there is something wrong in his programming, the instructions should be reviewed and see what part should be corrected and continue with the process.
From here we can start with the machine learning, and it is this, avoid mistakes, what is intended to be done is that the machine gets experience from the task performed and thus improve the task that is being done. Replicate the cognitive faculties of the human being. That through the same instructions given, that is, learn from mistakes and look for an optimal way to solve them without human intervention. However, maching learning is still in this process.
“Initially the functions were basic and were limited to filtering emails, although currently, it does more complex things such as traffic predictions, mapping sites, cancer detection, emulating construction projects in real-time even in defining compatibility between two people” — Gabriel Kent , CTO of Adext.
Gabriel Kent, CTO of Adext. He goes on to say, the learning process must be carried out by means of data insertion, capable of assimilating a wide range of information, which is also called big data.
These data are divided into three types: supervised learning, unsupervised learning, reinforcement learning.
Supervised learning is to teach the machine as a child, to relate images, behaviors, and even cases in which they will be given an answer.
In the unsupervised, he will process these data, as well as a child who, based on information, must relate them, such as explaining to a child the values that are good and bad, that are allowed and that are not.
Learning by reinforcement, the machine is able to learn by trial and error, posing various situations. Now it’s like leaving and trying to make a child make their own decisions and knowing the consequences.
The learning process should be similar to that of a child. What for this is necessary to understand learning theories.
As defined by Isabel García, learning is all that knowledge that is acquired from the things that happen to us in daily life, in this way knowledge, skills, etc. are acquired. This is achieved through three different methods, experience, instruction, and observation.
According to Patricia Duce, one of the things that greatly influence learning is the interaction with the environment, with the other individuals, these elements modify our experience, and therefore our way of analyzing and appropriating the information. Through learning, an individual can adapt to the environment and respond to the changes and actions that take place around him, changing if this is necessary to survive.
Why cite theories? It is important to be clear about the concept of learning to go as related to the instructions as it is important to know what should be indicated and how to do it, what it requires and how to acquire it.
The learning process of man is daily and requires his senses and the perspective of the environment.
So it requires a lot of processes and data analysis and therefore can be used in any area, such as:
- DNA classification
- Economic predictions and stock market fluctuations
- 3 D mapping and models Fraud detection
- Medical diagnoses
- Search engines
- Voice recognition systems
- Optimization and implementation of digital advertising campaigns
- Even in search of better audience or demographic groups for ads
