A brief story: Machine Learning for newbies

We all have read sci-fi stories about robots taking control over our jobs, people and life around us, we have seen movies about robots with human figure behaving like one (Machina is an example of this), but most of us don’t truly know what’s going on with them, how are they created neither the process they go through to become robots, some people may believe this is not true but our grandparents grew up believing this imagining a future world manage by technology and they might be angry about it even if they don’t comprehend it.

Robots might be the most representative of Artificial Intelligence but they are not all Artificial Intelligence (AI), in fact, AI is everywhere around us, the streets, at work and even in our homes, it doesn’t have to be something tangible it can be within the internet, specially on the social media, like facebook, youtube, Netflix, Spotify and so on; but there is something beyond that goes farther into the intelligence used in these “systems”, managed not by the human but the “system” itself, which is Machine Learning. Machine Learning is part of the AI, is a sub-field, an area, an application; a technical definition taken from Forbes

“ Machine Learning is a sub-set of artificial intelligence where computer algorithms are used to autonomously learn from data and information. In machine learning computers don’t have to be explicitly programmed but can change and improve their algorithms by themselves ”

In much more specific words it means a team program a machine or a “system” to learn by itself, taking everything that can be found around them or in a cloud of information and “form ideas out of it”; they also have the chance to work along with humans to connect with them and learn from that “connection” but that depends in their specified used. Companies are the ones who need and use Machine Learning the most, specially if they have a platform where users are their main thing.

The process of Machine Learning can be quickly and a little bit though, think of it as a child, first they try to walk, then to communicate, and they do so while been in an environment that can allow this; to expand the grip of knowledge they have to be expose to certain situation, as they grow, they develop skills and talents and the level of communication is bigger than years before, that continues until they die, a similar process happens with machines, but the human interaction is a little distant at some point. Each step that machine goes through has a classification and goals they have to commit in order to make bigger steps, a significant progress.

Gathering, is the first step, can be seen as baby finding its way to communicate with the world; the machine collects varied data from different sources.

Data preparation, the baby goes to school; is extracting data, analyzing and evaluating it, to filter it.

Training, the baby goes into adolescence; the filter data is split into two parts, one part is for data training, which creates the model and the other one is used as referenced data.

Evaluation, the not so baby goes to college to find the career; as its name says, it test out all the data trained to check for its accuracy and veracity.

Feedback and reevaluation, the not so baby has riched the adulthood, having all data processed, it starts to “play around”, thinking and rethinking which way is the best way, and changing.

All these steps are given in each of the following ways of learning, since it can be different depending on the purpose of the research or what the company wants.

Supervised learning

An example can be a teacher that is always there given you the knowledge and making sure you get it the right way, in this case, it is given labelled data to the machine to proceed with its training so it can make tasks based on the knowledge acquired.

Unsupervised learning

Theres no one that teaches to the machine, so it takes all the information and data found and label it on its own, and filtering as it think it’s convenience and spends a lot of time practicing with this, trying to produce similar outputs to the found results in the environment.

Semi-supervised learning

Is a mix of both supervised and unsupervised learning, it takes all data found and the labelled data given to learn, and its expected to do predictions based on the collection of the information filtered

Reinforcement learning

“There is no correct answer known to the system”, based more on experience, since the machine doesn’t know what’s right and wrong it takes decisions and make predictions based on the experience it has on the situation.

As said before, we can find Machine Learning everywhere around us, since the technology is evolving in big steps and becoming necessary for our daily basis, and kind of a helpful hand.

The simplest and easiest one is facial recognition, there are some applications that requiere the user to take a live photo in order to register, and China is trying to use surveillance cameras with this feature; which means, the camera would be able to recognize your face at first sight and show the description and information of your persona.

The closest one is on your cellphone, Siri, Alexa are some of the most popular ones, you have the option to personalize some features they have but the looking for an phone number or going into the internet to look up for something you need to know about is part of machine learning because with each search they learn and practice so they can become a better tool for the user.

We use them and although we despite them we enjoy them, social media; in Youtube you search for music or tutorials or videos in general and the more you consume it the more it recommends you similar videos, something similar happens with instagram, facebook, Spotify and so on.

Machine Learning, when well used, can serve to help the environment, our surrenders, our jobs, our studies, and it’s constantly upgrading itself to be more precise with it own process, to be more accurate in the predictions it gives, or to be enjoyable. While being kept to the serve of the human and its evolution, Machine Learning it a tool of the future.

Sources