Intive Blog

How to Use Machine Learning when You Have Zero Knowledge on it

Machine Learning is one of the tech trends considered worth investing in by experts to add value to a company’s services portfolio. Some other time we will discuss its benefits and how to get started in this world. In this article, we will focus on the services that can help us apply it in a simple way, in order to leverage what this technology has to offer for our products.

First of all, we should bear in mind that cloud providers like AWS, Google Cloud and Azure have a services offering for different levels to apply Machine Learning. This is known as AI democratization, and it aims to help clients adopt the technology faster. In that way, everyone can have access and build from their own infrastructure, using previously designed services. The image below shows AWS three-level offering for implementing a learning architecture.

Let us dig deeper into them:

  • For the lower level, you need a team with expertise in Data Science able to build infrastructure from scratch. Deep knowledge is required, and great efforts are needed to achieve the goal.
  • The intermediate level is intended for experienced teams that want to transfer part of the infrastructure to the provider and focus solely on implementation. It is very demanding but the cloud allows to simplify many configuration-related processes.
  • The higher level is perfect to start learning about and using Machine Learning in our projects. These are high-level services that have been previously implemented and tested.

The Top Level

The higher level is the easiest one to integrate in our applications, and little knowledge on Machine Learning is required. We just have to access through the API provided by the cloud. These services can be split into three categories:

They analyze images and video, from facial expressions (happiness, anger) to objects. In my opinion, these services are great to create amazing presentations for your audience.

They search for words in images and video. Services that obtain results from text or voice can also be included in this category. We may be familiar with this kind of service, but thanks to the potential of Machine Learning, they can deliver more accurate results, with a better conversion and analysis.

This kind of service is pretty known. I believe most of us has used Google Translate sometime.  Since always, science-fiction has been promising a service able to translate to any language in real time —something like the Babel fish (in case you miss the reference, ask Google what the answer to life, the universe and everything else is).

Once we have chosen which high-level service is appropriate for us, we must define the information to be collected. Quick tip: it’s best to focus on what we really want and can do, from the point of view of ability and lawfulness. Then, it’s only a matter of integrating services into our application and compile the information. The information obtained will allow us to add value to our business.

In this way, we can quickly leverage a few of the benefits of Machine Learning. This can be our card to adopt this trend and, ultimately, something that can make a difference for our products.

Rodolfo Cordero

Rodolfo Cordero has been a developer at intive since June 2016. He is a graduate in Software Development from the Universidad Latina de Costa Rica, his country of origin. A regular reader and music lover, he took courses in cocktailing and to become a barista, skills that delight the staff of intive in the after parties organized by the company.

Add comment