Other parts of this series:
Artificial Intelligence (AI), Machine Learning (ML) and Natural Language Processing (NLP) are all topics of considerable interest, however to the average person or senior business executives and CEO’s it becomes increasingly difficult to distinguish between these capabilities. Business executives want to understand how it’s truly going to improve business, provide a better customer experience or generate operational efficiencies.
In his post ‘What on earth is AI anyway’ Kris outlined what AI and Machine Learning really are, giving some detail on the concepts that sit behind the buzzwords. In short, AI describes a collection of algorithms that can perform tasks that are normally thought to require a level of intelligence – playing chess, using an arm to pick up an object, recognising someone’s face in a photo – that sort of thing. ML is a subfield of AI that deals with systems which learn how to perform tasks for themselves based on training data, rather than being explicitly programmed how to do so. Almost all of the compelling applications of AI that have appeared in the last few years are based on ML and so it’s the most important subfield of AI.
So, how does ML work? How can an algorithm learn from data? And why should I care? Well, I think that having an intuition for how it works is helpful in understanding where and how it can readily be applied to solve a task, and where it can’t and why. So, even if you’re more interested in the applications of AI than technical aspects of how it works, getting the detail is valuable.
In our below blog post, Martin Benson explores more about what exactly machine learning is covering the four main ingredients of machine learning, including:
- Training Data
- Loss Function
- Optimising Algorithm
- Learned Model