Explaining Machine Learning to a Grandfather
Keys: Machine Learning, AI, Statistic, Supervised learning, Unsupervised learning, Deep learning, Neural network, Data.
Explain Machine Learning to a grandfather could be a difficult task, especially that the technology industry has grown fast in the least years. But Why is important explain this to a grandfather? Why a grandfather should know about machine learning?
Throughout history, humans have tried to automate the processes they have designed to improve their quality of life, to increase their productivity. For example, the design of hydraulic systems, the design of engines, the design of controllers for industrial processes like PID, PLC, so on. The idea with automatization is control the processes designed. The problem is automatization does not have into account the changes through the time because automatization is designed with a specific set of rules, but Machine Learning is here to change the rules.
What is Machine Learning?
First, we must define what machine learning is. A brief definition is:
Machine Learning is the iterative process a computer follows to learn from data to improve, describe data, and predict outcomes.
For example, suppose that you feel ill and describe perfectly the symptoms you present, based on this data and comparing it with medical records of patients who meet the same criteria, a diagnosis can be made. This is the work that a doctor performs daily, and machine learning can also perform it.
What we need for Machine Learning?
Some skills are required to apply machine learning:
1. Computer science fundamentals and programming
2. Probability and statistics
3. Math’s: Calculus, Linear algebra and so on
Computer science fundamentals and programming are required to understand algorithms and data structures (Algorithms are the steps you make to find a solution to a problem and data structures are the different ways that data is organized when we are working with it), their complexity and so on.
Probability and statistics are required because, you know, we need the criteria to decide when to choose an option. For example, suppose that you are driving, and the car fails for any reason, based on the observation you can deduce what part of the car made that the car fail.
Math’s are required because we need to take the data and try to model the behavior of this data through mathematical models.
These are the basic skills that we need but maybe, more are necessary.
How Machine Learning works?
Machine learning is classified in different categories based on how learning is received or how feedback on the learning is given to the system developed.
Two off the most widely adopted machine learning methods are supervised learning and unsupervised learning.
Supervised learning
Supervised learning is based on understanding how a data set is classified and is intended to find patterns in this data set to apply analytical processes to classified data that are not classified.
For example, with supervised learning, an algorithm may be fed data with images of sharks labeled as fish and images of oceans labeled as water. Training this algorithm on this data, it should be able to identify unlabeled shark images as fish and unlabeled ocean images as water.
Unsupervised learning
On the other hand, unsupervised learning is best suited when the problem requires a data set that is not classified (labeled) because unlabeled data are more abundant than labeled data.
For example, in healthcare, collecting huge amounts of data about a specific disease can help practitioners gain insights into the patterns of symptoms and relate those to outcomes from patients
Reinforcement learning
Reinforcement learning is another machine learning method. It differs from other types because it is not trained with a data set, it learns from trial and error, reinforcing itself and given better outputs on time.
One of the most common applications of reinforcement learning is in robotics or game playing. Take the example of the need to train a robot to navigate a set of stairs. The robot changes its approach to navigating the terrain based on the outcome of its actions. When the robot falls, the data is recalibrated, so the steps are navigated differently until the robot is trained by trial and error to understand how to climb stairs. In other words, the robot learns based on a successful sequence of actions. It is like when a baby is learning to walk, he learns to make better decisions with each step he takes.
Neural networks and Deep learning
Deep learning is a specific method that incorporates neural networks to learn from data in an iterative manner. The idea with this method is emulate how the human brain works so computers can be trained to deal with abstractions and problems that are poorly defined.
A neural network consists of three or more layers: an input layer, one or many hidden layers, and an output layer. Data is supplied through the input layer. Then the data is modified in the hidden layers and the output layers based on the weights applied to these nodes. Then, the term deep learning is used when there are multiple hidden layers within a neural network.
The typical neural network may consist of thousands or even millions of simple processing nodes that are densely interconnected, every node is a mathematical function that collects and classifies information based on statistical methods. Using an iterative approach, a neural network continuously adjusts and makes inferences until a specific stopping point is reached.
We have talked a lot about data in machine learning, but What is exactly data in machine learning?
What is data in Machine Learning?
Data can be any unprocessed fact, value, text, sound, or picture that is no being interpreted and analyzed.
Data can represent valuable information once it has been interpreted and becomes knowledge when we combine this information with experiences in our life.
How data is split in Machine Learning?
Data could be divided into three categories:
- Training Data: Data that we use to train our model.
- Validation Data: Data that we use to evaluate our model comparing this data with the output of our model. This data plays its part when the model is training.
- Testing Data: Data that we use to evaluate our model once our model is completely trained.
So, I hope grandfather that you had enjoyed this little explanation about what machine learning is. We saw what machine learning is, what we need for machine learning, and how machine learning works. I hope that you do not forget anything and that you have aroused a little interest in studying machine learning. Oh… I almost forget it… remember that machine learning is a subfield of AI, but we can have this discussion in another time.
Machine Learning is a powerful technique that will change our life in every sense.
