What is machine learning?

5 mins read - Updated on Aug 26, 2022

Machine Learning is a subfield within Artificial Intelligence that builds algorithms that allow computers to learn to perform tasks from data instead of being explicitly programmed. Machine learning algorithms are able to detect patterns in data and learn from them, in order to make their own predictions. The pretty technical definition, isn't it?In short, machine learning algorithms and models learn through experience. Wow, that sounds interesting.

In traditional programming, a computer engineer writes a series of directions that instruct a computer how to transform input data into a desired output. Machine learning, on the other hand, is an automated process that enables machines to solve problems with little or no human input, and take actions based on past observations.

To explain this definition through a more understandable example, it is possible to compare the machine learning process to the process of a student learning a lesson and solving a test. At school, the subject of the lesson is explained to the student, and then the student works on the subject himself and starts to solve the test. However, it is not possible to teach everything about a subject to any human being (or even to the computer with the highest processor despite today's advanced technology). After the school has "taught" up to a point, the student is now expected to start his own "learning" process by evaluating the information provided. Starting from the information provided to him, he will get better with each question by applying these experiences to the problems he encounters for the first time. The more and more accurate information the school provides to the student at first, and then the more different questions the student solves, the better the result will be. Machine learning follows a similar process.

Types of Machine Learning

To understand how machine learning works, you’ll need to understand different machine learning methods and algorithms, which are basically sets of rules that machines use to make decisions. Below, you’ll find the three most common types of machine learning:

Supervised Learning

Supervised learning models make inference based on labeled training data. Each training sample includes an input and a desired output. A supervised learning algorithm analyzes this sample data and makes a prediction.

This is the most common and popular approach to machine learning. It’s “supervised” because these models need to be fed manually tagged sample data to learn from. Data is labeled to tell the machine what patterns (similar words and images, data categories, etc.) it should be looking for and recognize connections with.

For example, if you want to design a system that decides whether the images are human faces or not, you must first feed the machine with data by entering different human faces into the system and marking them as "human faces" and the images without faces as "not human faces".

Unsupervised Learning

Unsupervised learning, another type of learning, is a machine learning technique that uses an algorithm to predict an unknown structure on unlabeled data. To make it more understandable, I would like to give an example here. Let's say you want to guess the occupations of passers-by. By looking at the clothes or movements of the people on the street, you grouped them all in your mind and started to think about their profession. Here you make predictions about people you know nothing about based on unlabeled data. Grouping people by occupation based on your guesses is an example of unsupervised learning.

Semi Supervised Learning

In semi-supervised machine learning, training data is divided into two groups. A small amount of labeled data and a larger amount of unlabeled datasets.

In this case, the model uses labeled data as input to make predictions about unlabeled data, resulting in more accurate results than normal supervised learning models.

This approach does not require large amounts of labelled data, so it is faster to create, more cost-effective, and ideal for companies with large amounts of data.

How Machine Learning Works

To understand how machine learning works, you first need to know what a "label" means. For example, it is necessary to label whether a consumer review is about customer service, product features, ordering and delivery process, or price. This is also called data labeling.

When working with machine learning text analysis, you would feed a text analysis model with text training data, then tag it, depending on what kind of analysis you’re doing. If you’re working with sentiment analysis, you would feed the model with customer feedback, for example, and train the model by labelling each comment as Positive, Neutral, and Negative.

In its simplest form, the machine learning process includes these steps:

  • Feed a machine learning model training input data. In our case, this may be customer reviews we collect from media such as Twitter.
  • Label the training data with the desired output. In this case, teach your sentiment analysis model that each comment is Positive, Neutral, or Negative. The model transforms the training data into text vectors, that is, numbers that represent data features.
  • Test your model. Algorithms are trained to associate feature vectors with tags based on manually tagged samples, then learn to make predictions as they process unseen data.

Once your new model has been tested and meets your standards and criteria, it is ready to work on any new data. If the model doesn't have the desired accuracy, you'll need to continue training. (If you want to learn how to optimize the accuracy rate, you can check out this article.) Moreover, as human language and sector-specific language morphs and changes, you may need to continually train your model with new information.

🍍 Also see: What is Machine Learning, and How is it Used for Text Classification?