Understand the scope of Machine Learning. Supervised, Unsupervised, Semi-Supervised

Grace Omojola
AI+Club OAU
Published in
2 min readFeb 22, 2021

Introduction

Machine learning is a sub-field of artificial intelligence (AI) that gives systems the ability, without being specifically programmed, to automatically learn and develop from experience.

In order to explore possible underlying patterns concealed in our data, we need to have some observations or data (also known as samples or examples) available for the learning process (model fitting). These learned patterns are nothing more than some function or decision.

Supervised Machine Learning

You only need labeled examples for supervised machine learning tasks, where you must specify the ground truth for your AI model during training. Examples of supervised learning tasks include image classification, facial recognition, sales forecasting, customer churn prediction, and spam detection.

Unsupervised Machine Learning

On the other hand, unsupervised learning deals with instances where you do not know the ground truth and want to identify specific patterns using machine learning models. Customer segmentation, anomaly identification in network traffic, and content recommendation are examples of unsupervised learning.

Semi-Supervised Learning

Semi-supervised learning stands somewhere between the two. It solves classification problems, which means you’ll ultimately need a supervised learning algorithm for the task. But at the same time, you want to train your model without labeling every single training example, for which you’ll get help from unsupervised machine learning techniques.

Summary

Supervised: All observations in the dataset are labelled and the algorithms learn from the input data to predict the output.

Unsupervised: All observations in the dataset are unlabeled and the algorithms learn from the input data about the underlying structure.

Semi-supervised: Some of the dataset’s observations are labelled, but most are usually unlabeled. Therefore, a combination of supervised and unsupervised approaches is normally used.

We are able to conduct analyses of large amounts of data using Machine Learning (ML) models. Data patterns that would be difficult for a human being to recognize can be accurately extracted within seconds using these ML models (in some cases). However, reliable outcomes (good models) usually take a deal of time and resources for model training(the process in which a function or a decision boundary is learned by the model) most of the time.

Thanks for following this far. Hope you liked this article!

I can be contacted via Twitter or the comment section if you have any questions, suggestions, or critiques.

Editor’s Note: AIPlusOAU is a subsidiary of Data Science Nigeria based in Obafemi Awolowo University, Ile-Ife, Osun State. Our aim is to raise a community of Data Scientists, Machine Learning/AI Engineers, and Researchers. We are dedicated to supporting young talents to start and grow their Data Science skills.

Do follow AIplusOAU on Twitter for emerging trends in Data Science, Machine Learning, and AI.

--

--

Grace Omojola
AI+Club OAU

Data Scientist | Python Developer|Technical Writer