The Future of Humanity is Actually Machines : Part One
Ever wonder what decides whether your emails are primary, updates or spam? Or what exactly controls the customized adds you receive on almost all your social medias? Well, that would all be Machine Learning (ML).
ML is a method of achieving Artificial Intelligence (AI), a term used for human-like intelligence shown in computers. ML allows the computer to “learn” on its own, by taking in data and using it to predict future results. It’s similar to teaching someone the skills to add, so that they can solve any addition problem you give them. A more accurate and relevant example is giving a computer data on cancer patients, and then having it predict the chances of someone having cancer based on the information inputted.
Machine Learning uses algorithms, which are a mathematical rules to follow that help solve a problem. Algorithms can be characterized into different groups, Supervised Learning, Unsupervised Learning and Reinforcement Learning. Each of these sections helps achieve different type of tasks. Examples of these tasks are Estimating Life Expectancy, Gaming, and Targeting Audiences which all use different algorithm structures to function.
In Supervised Learning, machines are given inputs and know what the output will be. The computer takes this data and tries to find the patterns within it.
If you compare it to a human, it’s similar to giving someone the equations, 2² = 4 and 3² = 9. Then, having the person find the pattern between the numbers, to better understand it.
This can be classified as Regression or Classification.
Regression is when we are trying to find the predicted value of a set of continuous data (ex. finding the height of a child using age as a input)
Classification is when we are trying to classify data into groups (ex. find out whether something is a cat or a dog using the features as a input)
In Unsupervised Learning, the machine is given inputs, and has to find the output the given data.
This type is what excites most AI developers, the ability for a machine to learn completely on its own. It copies how a infant learns, when you give them different shapes and they divide them in groups with their prior knowledge.
This can be done by Clustering and Dimensionality Reduction.
Clustering is to group data that share similar characteristics, this helps sort and find patterns with the data.
Dimensionality Reduction refers to different techniques that help eliminate data that isn’t needed.
Reinforcement Learning is when the computer is programmed to complete a certain task, and learns to do it by its self using trial and error. So it keeps trying to complete the task until it gets it right.
This relates to Unsupervised Learning as the machine is learning itself. Think of a child trying to fit shapes in the proper molds, and keeps trying until it fits.
In just a few minutes you’ve learnt the basics of Machine Learning. The uses for this technology are endless and will change the way our world works. It can be applied to healthcare, research, customer service, and entertainment, to name a few. Now, I urge you to try. What do you think you can use Machine Learning for?
Leave a clap if this was helpful! Stay tuned for Part Two of this series, where we’ll be discussing the next step to understanding Machine Learning!