Machine learning (ML) is a type of artificial intelligence (AI) that allows software applications to carry out actions that they’ve not been explicitly programmed to do. Machine learning algorithms use existing data to predict new outputs allowing it to ‘learn’ patterns to predict outcomes.
Imagine you want to teach a computer to understand human language, specifically the sentiment of a sentence, whether it is positive, neutral or negative. Just like teaching a computer to recognise cats, you would need a lot of examples. So you would give the computer a dataset of sentences, some labeled as “positive”, some as “negative” and some as “neutral”.
For example, you might show the computer a sentence like “I had a fantastic holiday” and tell it “this is a positive sentence.” Then you would also show it sentences like “I had a terrible time at the concert” and tell it “this is a negative sentence.”
As the computer “reads” more and more examples, it starts to recognise patterns in the words and phrases that are used. The computer might notice that words like “fantastic” and “great” are often used in positive sentences, while words like “terrible” and “horrible” are often used in negative sentences. The more examples it sees, the better it gets at identifying the sentiment of new, unseen sentences.
This is just a simple example of what a machine learning model for natural language processing (NLP) can do. NLP is a very broad field, and there are many different tasks that can be performed using machine learning, like language translation, text summarisation, named entity recognition, and more.
In summary, Machine learning is a way to teach a computer to understand natural language text, like written sentences or spoken words, by training it with a lot of examples, labelled with the desired output, the computer learns how to recognise patterns and make predictions, like determining the sentiment of a sentence in this example.