Artificial intelligence, Deep Learning, and Machine Learning are the buzzwords of the moment. A simple search for machine learning on Google Trends paints a clear picture. The interest in the subjects has increased dramatically over the last couple of years.
So what exactly does the term ‘Machine Learning’ mean?
Even among Machine Learning practitioners there isn’t a well accepted definition of what is and what isn’t Machine Learning, but the most general idea is given as –
“ Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.”
Trending AI Articles:
The world is filled with data. Lots and lots of data. Everything from pictures, music, words, spreadsheets, videos and more. It doesn’t look like it’s going to to slow down anytime soon. Machine learning brings the promise of deriving meaning from all of that data.
At first ML may seem like magic, but once you dive in, you’ll see that it’s a set of tools to derive meaning from data.
Data all around us
Traditionally, humans have analyzed data and adapted systems to the changes in data patterns. However, as the volume of data surpasses the ability for humans to make sense of it and manually write rules, we will turn increasingly to automated systems that can learn from the data, and, importantly, changes in data, to adapt to a shifting data landscape.
Machine Learning is already everywhere
We see machine learning all around us in the products we use today, but it isn’t always apparent to us that machine learning is behind it all.
- Online recommendations. Machine learning allows retailers to offer you personalized recommendations based on your previous purchases or activity.
- Ride sharing Apps Like Uber and Lyft. How do they determine the price of your ride? How do they minimize the wait time once you hail a car? How do these services optimally match you with other passengers to minimize detours? The answer to all these questions is ML.
- Spam Filters. Your email inbox seems like an unlikely place for ML, but the technology is largely powering one of its most important features: the spam filter.
- Facebook’s News Feed. The News Feed uses machine learning to personalize each member’s feed. If a member frequently stops scrolling to read or like a particular friend’s posts, the News Feed will start to show more of that friend’s activity earlier in the feed.
How Machine Learning works?
Machine learning is divided into three main categories.
- Supervised Learning: Here, the system is trained using past data, which includes input (also known as features) and output (also known as labels), and is able to take decisions or make predictions, when new data is encountered.
- Unsupervised Learning: The system is able to recognize patterns, similarities and anomalies, taking into consideration only the input data, i.e. using only the features.
- Reinforcement Learning: Decisions are made by the system on the basis of the reward/ punishment it received for the last action it performed. There is no past data available in this method. The system takes dynamic decision as the environment, in which it is present, changes.
These categories are further divided as:
Here the data needs to be divided into a number of different categories based on training using past data. An example of a classification problem, would be how we are able to sort emails are spam or otherwise using previously received emails that have been already identified. A famous algorithm that can be used to solve classification problems, is the Naive Bayes theorem.
We predict a value for an input based on previously received information. Although this sounds similar to classification, considering that they both use past data to make predictions, their similarity ends there. In the case of regression, you’re trying to estimate a value and not just a class of an observation.
Now, let’s consider you want to predict the value of a house you are trying to sell. This problem doesn’t come under classification because we are not predicting whether the will be sold or not. Rather we are predicting what the price of the house will be when it will be sold.
This uses a method where we assign a set of observations into subsets. These subsets are known as clusters. And the observations inside these clusters are similar to one another, based on some parameter or other. Hence, all the data is divided into clusters.
An example of when clustering is used, when a telecom provider wants to set up a network in a region by setting up towers there, they use the clustering algorithm, taking into consideration areas that would provide optimum connectivity to all users and the maximum range a cell tower would have, to divide the entire region into clusters. K-Means is a prominently used method to cluster data in k-clusters based on some similarity measures.
In an association problem, we identify patterns of associations between different variables or items.
Its concepts are applied to e-commerce websites, where they’re able to suggest other items for you to buy, based on the prior purchases that you’ve done.
And the future of Machine Learning?
The possibilities are endless. Here’s some of them:
Remember those robot helpers you saw in I, Robot? Imagine those in our day-to-day lives. Helping clean up our homes and generally making life even easier.
Traffic annoying you? How about you relaxed in the air conditioning of your car, and it took care of taking you to your destination? On its own? All thanks to autonomous cars.
Or how about as soon as you entered your doctor’s office, they have access to all your relevant medical details. Enabling them to provide you with a more personalized diagnosis?
Machine learning could bring about all of this and more!
Don’t forget to give us your 👏 !
What exactly is Machine Learning? was originally published in Becoming Human: Artificial Intelligence Magazine on Medium, where people are continuing the conversation by highlighting and responding to this story.