What is machine learning?
Machine learning is the ability to teach a machine without programming it explicitly. Machine learning concerns everyone today. It has revolutionized the world of transport, especially with autonomous vehicles, bio-mechanics, the world of health which allows to diagnose cancers or to design new drugs, the world of finance where today 80% of banking transactions are made by AI. Nowadays, it is used a hundred times a day without being accountable.
Here are some applications using machine learning models:
- Every time you do a search in Google and you come across interesting results, it’s thanks to a machine learning algorithm that has learned how to find relevant results among the billions of possibilities that are offered.
- When you post a photo on Facebook, the algorithm recognizes the people in the photo so you can tag them later.
- When you access the Netflix service, the recommendation system helps you find the TV series or movies you might enjoy as easily as possible based on what you’ve watched.
There are three main types of machine learning:
Supervised learning :
The algorithm is provided with examples that are considered successful after having previously labeled them with expected results. Artificial intelligence then learns from each example by adjusting its parameters so as to reduce the gap between the results obtained and the expected results. The margin of error is thus reduced as the training progresses, with the aim of being able to generalize its learning to new cases.
Unsupervised learning :
In the case of unsupervised , the learning process is carried out completely independently by the machine. Data is then fed to the machine without providing examples of the expected output results. Unsupervised learning is mainly used in clustering, a method designed to group together a set of heterogeneous elements in the form of homogeneous sub-groups or linked by common characteristics.
Reinforcement learning :
It is a Machine Learning algorithm that allows software agents and machines to automatically determine the ideal behaviour within a specific context to maximize its performance. It does not have labelled dataset or results associated with data so the only way to perform a given task is to learn from experience. For every correct action or decision of algorithm, it is rewarded with positive reinforcement whereas, for every incorrect action, it is rewarded with negative reinforcement. In this way, it learns which actions are needed to perform and which are not.
In the rest of this course, we will go into more detail on the different types of supervised learning algorithms.