What is AI ? (Artificial Intelligence)

I would like to start with my definition of intelligence, since from there it will develop into deeper concepts, intelligence is what allows us, humans, animals and now even machines, to solve problems using the information we have learned in some field, intelligence varies between individuals and is diverse in different areas such as mathematics, social skills or others, now in the case of machines, it is called artificial intelligence. This is the capacity that machines (computers) have in being able to learn from data and be able to carry out a process that allows them to solve or optimize a task for which it is programmed, just as we humans solve things. There are a wide variety of fields in which AI can be applied to, such as computer vision, natural language processing, analyzing data, generative text, among many more applications.

Let's say that artificial intelligence is a bubble that allows machines to understand the world and for that, it requires data to digest the information, which we call machine learning, it is found within the bubble of artificial intelligence and it allows machines to learn.

There are certain important fields in machine learning, I am going to discuss 4 concepts. Let's say you want to predict the weather, how do you know that next winter will be cold? Precisely as it has been in recent years, you already have a history in which you know the temperature in winter, this is how supervised learning works, through the seasons, the machine will learn through “cold” and “hot” labels, associating that in winter it will be “cold” and in summer “hot”, thanks to historical information. Supervised learning, then, is a branch of machine learning that allows us to associate labels and predict future values that do not yet have a label.

Certain supervised learning models are:

- Regression

- Classifiers

- Random forests

- Support Vector Machines

Another important concept of machine learning is unsupervised learning. When you put together a LEGO, do you separate the pieces by color? This is precisely what unsupervised learning does, it groups the pieces by colors and shapes, allowing the assembly and categorization of the pieces to be much simpler.

Certain unsupervised learning models are:

-?Principal component analysis

- Decomposition of singular values

- Clustering methods (Clustering, k-means)

On the other hand, there are more concepts of artificial intelligence, a subcategory of machine learning is deep learning, which combines layered processes, that is, continuing the example, it first groups the LEGOS by color and then by size, probably then starts to put it together. Let's say that each process is called a node, deep learning consists of what are called neural networks, it is not specifically a brain, however it seeks to simulate the way one works.

Since different concepts of machine learning and its subcategory, deep learning, have been touched upon, it is important to know that there are levels at which artificial intelligence can be classified, at its most basic level narrow (weak) artificial intelligence, which is where you've probably interacted with Netflix, Spotify or Amazon, the recommendation system.

This type of intelligence is based solely on one type of process to be carried out. While there are systems or applications that can combine different artificial intelligence tools in their process, one cannot combine processes. An example where types of artificial intelligence are combined would be if Netflix had a Chatbot in which it recommended movies.

From here, they are hypothetical concepts, that is, currently 2024 they do not exist, rather, strong artificial intelligence (general) is artificial intelligence that has knowledge about several domains and is even equal to or more intelligent than a person, as well as artificial superintelligence which can solve complex problems much faster than strong artificial intelligence and humans. Although they are not currently real, it is considered interesting to take them into account since in some future there may be models that are similar.

We have already reviewed what artificial intelligence, machine learning, supervised and unsupervised learning and types of artificial intelligence are, however,

How does this provide a benefit to companies?

It is important to take into account that artificial intelligence is not an adversary towards the worker, rather it is a tool that will allow tasks to be carried out faster, allowing the worker to focus on more important tasks.

Next, I will list 4 benefits with examples of supervised and unsupervised models:

? The efficiency of solving repetitive problems:

The benefit is to leave repetitive tasks to a machine so you can focus on other projects and not harass the worker by arduously reviewing the same problem.

Supervised: Classify emails using text tag classification.

Unsupervised: Grouping similar documents into categories using clustering.

? Solve complex problems:

Thanks to machine learning, it is possible to analyze large-scale data more quickly, identifying anomalies and patterns for decision making.

Supervised: Demand forecast based on historical behavior of sales, consumption, purchases, etc...

Unsupervised: Detection of anomalies in team performance to support the HR team in the company.

? Improved decision making:

Due to the large scale analysis, it allows you to project the data and analyze different types of data.

Supervised: Inventory predictions.

Unsupervised: Text-based sentiment analysis in social networks.

? Security and fraud detection:

Supervised: Detection of credit card fraud through classification models.

Unsupervised: Detect patterns of anomalies to prevent suspicious activities.

As can be seen, it is beneficial for a company to incorporate artificial intelligence, since it facilitates processes and positively impacts its clients as employees.

It is important to consider that artificial intelligence is expanding today and little by little there will be a business process where at least one artificial intelligence system is involved.

In conclusion, artificial intelligence, with its subcategories such as machine learning and deep learning, represents a technological revolution in which both companies and people benefit from the ability to address complex, repetitive problems and by providing a better experience.