Deep Learning

 

Deep learning, also known as deep learning, leverages the enormous amount of data available to train computers to function similarly to our brain. By doing so, machines can tackle problems and achieve desired results without explicit instructions on how to do it. For example, if we wanted to teach what a fork is, instead of describing it in detail, we would just need to provide an extensive set of images for it to learn to identify it by itself. This learning model is considered supervised because the images we show to our program have already been filtered and labeled as “forks,” so our model analyzes the common denominator of these images to learn what a fork is.

 

On the other hand, we can find unsupervised learning models, where the data has not been previously labeled. An example of this would be if we have many images of kitchen utensils without the corresponding label for each one. In this case, our unsupervised learning algorithm will analyze the characteristics of the items in the images, such as their shapes, colors, sizes, and textures, and group the ones it finds similar according to these and other patterns, but it will not have assigned labels like “forks” or “spoons” to them since that is not the purpose of these models.

 

Finally, we find reinforcement learning models, where our intelligent agent proceeds to learn to make decisions according to a system of rewards and punishments, aiming to obtain the highest amount of cumulative rewards. An example of reinforcement learning could be a model dedicated to learning to play chess. Our AI may not know that it is playing chess, what chess is, or what its rules are, but it will interact with its environment by moving its pieces, and over several games, it will be rewarded or penalized as it wins or loses each game. Thus, little by little, it will begin to perfect the actions that lead it to receive more rewards systematically.

 

This combination of deep learning and the vast amount of data available today has resulted in remarkable achievements recently. After all, we must not forget that when we walk or move using or not using some type of vehicle, we always carry a cell phone with GPS in our pockets, which collects information and sends it to the companies behind the development of each application we have installed. The same happens when we use our web browsers and click on different buttons on a web page. If we spend more time focused on a particular area of a website or if we quickly move away from it. All this helps to form those enormous databases that feed the algorithms that make our lives easier and that sometimes suggest products that we might not have even known we wanted or existed.

 

We now need to understand what Artificial Intelligence really means along with the exponential increase in technological power for the labor market. With this new ability of machines to learn, it is possible that not many of the jobs we know today are protected, and we might even be creating jobs that can already be automated.

 

Click here to read the next chapter 👉 
 


Click here to return to the Index 🔍