Machine learning is not a new term, but in the last few years it has been brought to the spotlight due to technological advances and bets made by major players like Google and Facebook.
We have to go back to 1959, and see Arthur Samuel define it for the first time as the “Field of study that gives computers the ability to learn without being explicitly programmed”, but since that it has evolved.
We use machine learning (ML) when we want a computer to process big amounts of data, putting it working for us, categorizing, making predictions, searching for trends etc…
Based on the tasks that you want to see executed, there are 3 major types of ML algorithms:
Classification: For when you want the algorithm to discriminate and categorize classes on the data available. For example we want it to select from a set of images, the ones who have a computer in it.
Regression: In this case it gives us a correlation of outcomes, based on the past experience. In a simple example, if a team always looses when it’s raining, the algorithm can show that correlation and the coach can see how can he solve that problem.
Clustering: This is one of the most complex. It analyses trends in large amounts of data. This means that a clustering algorithm could go through a client database, look for patterns that would indicate the profile of our client.
There are another types of tasks that machine learning can perform, but basically, if a Computer program can improve the way it does a specific task based on past experience, we can say that it has learn.
Why Everyone is talking about it?
The advances of technology has brought us to a point where you don’t need a super computer to run machine learning models, with their own personal computer more and more techies experimented, implemented and improved algorithms, openning up a new bunch of possibilities that were not available 50 years ago.
Big companies and innovation leaders are already using it to improve its services.
Sundar Pichai, Google’s CEO stated on the company blog that the future of Google is based on Machine Learning and Artificial Intelligence and it’s effectiveness in continuous making the search engine and it’s related services more and more powerful and user-friendly
Uber is using Machine learning to reduce it’s surge price (when Uber fare rates automatically increase, due to lake of drivers around a high demand) by predicting when and where demands will increase they will be able to send drivers to a specific place and hour.
In the long term they expect the algorithms to even overcome the need for the surge price to appear, balancing automatically the demand and the supply with efficiency.
On it’s latest F8 conference, Facebook announced that Artificial intelligence is one of their main focus for the future, and machine learning is how they are scaling it.
Facebook even created an entire division, the Applied Machine Learning team with the main objective of optimizing and apply machine learning to facebook developing innovations.
Know more in this Stacey Higginbotham Article on Fortune Magazine.
At Wizdee we are using machine learning to present the most relevant charts for each query, while it continuously adapts the results for each user.
Still in implementation, is a new system to discover patterns, predictions and new knowledge without requiring any technical know-how from the user.
How this will impact our lives?
If the above big players are betting on Machine Learning we know that this technology is here to stay.
It is already improving and personalizing our search results on Google, and even suggesting complementary products for something that we are buying.
By analysing our groceries, it will soon be able to send notifications with a shopping list, with all the products we are running out in our storeroom.
Machine learning is making us do smarter decisions, for example, if you use it on your banking data, it can show your expected incomes and spents in the months ahead, then you can decide if you can buy that sports car.
Guts and instinct are still important in business intelligence. With data analysis and machine learned outcomes, it should be easy to prove your guts, or avoid making wrong decisions, saving you big amounts of money.
Isn’t that what we all want?
We are now in a new generation, as Roger Waters said more than 40 years ago…