Understanding Machine Learning in 5 Minutes

Understanding Machine Learning in 5 Minutes

“Machine learning” sounds mysterious for most people. Only a small fraction of professionals really know what it stands for. The reason for this, is this technology hype word is rather technical and difficult. I would like to bridge this gap and explain simply what machine learning (ML) is and how it is being used in our everyday life.

So what is Machine Learning?

In the most simple terms we could use –  Machine’s imitating and adapting human like behavior.

Machine Learning works with data and processes or sorts through it to find patterns that can be later used to analyse new information. Machine Learning is when a computer is programmed with the ability to self-teach and improve its performance on tasks we provide. In a nutshell, the computer analyses data, automatically takes information and uses it correctly to place into categories. If placed incorrectly, it will learn from its mistake to improve in the future.

An extremely common example of machine learning is Apples voice recognition friend, Siri! Siri uses machine learning to imitate human interaction. As we speak more with Siri, the program learns and betters itself based on our responses which differ from person to person.

Another example of machine learning that you may have not noticed is when Facebook suggests that you tag your friends in photos. But how does it know who that person is? Facebook’s systems are searching through hundreds of your photos to recognize familiar faces in your contact list to better match peoples faces to your photos.

You can further divide machine learning algorithms into 2 main groups based on their purpose and abilities:

Supervised learning

Unsupervised learning

The supervised approach is similar to a student learning under the supervision of a teacher. The teacher provides good examples for the student to memorise, and the student then forms general rules from these specific examples.

The systems are more likely to make judgments that humans can relate to because we have provided the basis for decisions and limitations. However, supervised learning systems generally have trouble dealing with new information. For example, if a system with categories for cars and boats is presented with a bicycle, it would have to be incorrectly selected in one category or the other.

Unsupervised learning resembles the methods we use to figure out that certain objects or events are from the same group, such as by observing the similarity between objects.

While an unsupervised learning system might figure out on its own how to sort cats from dogs, it might also add unwanted and unnecessary categories to deal with unusual breeds, creating clutter instead of order.

As Machine Learning in everyday devices are steadily becoming a daily occurrence it is clearly helping us make faster and more accurate decisions. It may even be feeding our instant gratification needs?

“Machine learning is going to change every single aspect of technology, but no machine will be able to mimic the creative ability of the human mind.” –  Shantanu Narayen, CEO, Adobe