What is machine learning?

[RichardHatheway]

There’s been a lot of talk recently about machine learning (ML) and how it’s going to help businesses drive innovation, develop new products and services, and become more competitive. But before we get too excited, let’s answer the question that a lot of people ask, “What exactly is machine learning?”

This blog will provide you with a quick overview of machine learning so that you understand what it is and what all the excitement is about.

ML – a quick definition

Machine learning is a branch of artificial intelligence (AI) that focuses on computers using data and mathematical algorithms (aka, models) to imitate the way humans learn. Machine learning effectively automates the process of analytical model building and allows computers to adapt to new scenarios independently, based on the analysis of the data that is input. ML can be used in almost any area that has a defined set of rules or data points can be evaluated.

The model goes through an iterative process where the computer takes in data, evaluates it, performs some type of calculation, and then delivers an output. That output is then compared to an expected output value to see how close it is to the expected value. The model is then adjusted by the computer based on how close to or far away from the forecast value the model output actually is.

Over time, machine learning models gradually improve the accuracy of their output by going through this iterative evaluation process and effectively fine-tuning the model. Once the model has been fine-tuned to the point that the output regularly meets or exceeds the expected output parameters, it’s ready to be used in a production or real-world environment.

Types of machine learning models

Machine learning models are typically broken out into three main categories. These categories are based on the type of training and amount of structure provided to the model, as well as how much autonomy the model has to provide the output.

These three categories are:

• Supervised machine learning 

This type of machine learning model uses labeled datasets to train the algorithms to predict outcomes accurately. The model analyzes data, compares it to the labeled dataset, and then creates an output.

An example of this type of model is a fruit scanning and sorting machine that looks for bruised fruit prior to them being packed. The labeled dataset that is provided has pictures of perfect fruit and blemished fruit. As the system scans the fruit on a moving conveyor belt, it compares the pictures it takes of the fruit on the conveyor belt to the examples (labeled dataset) that it has been provided and picks out fruit that are not perfect.

• Unsupervised machine learning

This type of machine learning model analyzes unlabeled datasets (i.e., raw data), classifies the data, creates categories based on the patterns that it sees in the data, and then puts the data into the relevant categories. 

Let’s use the same example as before, but this time the system must scan the fruit coming down the conveyor belt and create its own categories from the data it gathers before it can begin sorting the fruit. After scanning the fruit, it creates three categories: unblemished fruit, slightly blemished fruit, and damaged fruit. Once the categories have been created, the model knows how to separate the fruit and can begin scanning them in the production environment and separating them accordingly.

• Semi-supervised machine learning

This type of machine learning model is basically a compromise between the previous two types of models, as the model uses a small, labeled dataset to guide the classification of the unlabeled data.

An example of this type of machine learning model is another fruit scanning and sorting machine that runs in parallel to the one we used as an example above, but this belt has a mixture of three different types of fruit (Granny Smith, Golden Delicious, Red Delicious apples) on it. The small, labeled dataset provides examples of the three different types of apples moving along the belt, but that is all. The system then uses that information, together with the categories that it determines by scanning the fruit, to come up with a system of being able to sort and separate each of the three types of apples, as well as to categorize them into unblemished, slightly blemished, and damaged apples.

These examples help you understand how machine learning can be applied in the real world. ML models can be used to automate processes, which in turn saves time and helps increase productivity.

Now that you understand the differences between the three types of machine learning models, you’ll know what type of ML model needs to be built for a specific situation, as well as the type of data and supervision each model requires.

To learn more about machine learning, please visit any of the following resources:

• HPE glossary of enterprise IT terms: What Is Machine Learning?
• HPE Ezmeral Learn-On-Demand: Intro to AI & ML
• HPE Developer: Machine Learning 101 – Introduction to ML Concepts
• HPE Developer Blog: Artificial Intelligence and Machine Learning: What Are They and Why Are They Important?
• HPE Developer Blog: Types of Machine Learning – Part #2 in the Intro to AI/ML Series
• HPE Developer Blog: Demystifying AI, Machine Learning and Deep Learning

HPE Ezmeral on LinkedIn | @HPE_Ezmeral on Twitter

@HPE_DevCom on Twitter 

[SFERRY]

Also see the free on-demand courses series on Machine Learning, avaible at Ezmeral Learn On-Demand. Four courses covering an intro to artificial intelligence, details on the types of ML, and checklists on how to plan and execute successful machine learning projects.

• Intro to AI/ML
• Machine Learning project success planning
• Machine learning pipeline: prepare
• Machine learning pipeline: implement