With machine learning skills, you can work in data science, AI and medtech to name a few. Here, we give some pointers on how to get started.
Machine learning is a subset of AI that is used in a lot of real-world scenarios including customer service, recommender algorithms and speech-recognition software.
As machine learning is so widely used it is a great area to get familiar with. A very simple way of explaining machine learning and how it works is to think of it as computers imitating the way humans learn using algorithms and data.
Let’s take a look at some of the concepts you should know in machine learning. You may end up honing in on one of these areas down the line after you’ve learned some of the basics.
Terminology
Neural networks
Neural network architecture is often also referred to as ‘deep learning’. It consists of algorithms that can mimic the way human brains learn to process and recognise relationships between large data sets.
You’ll find neural networks used in sectors such as market research and any industry that interacts with large data.
There are three main types of learning in neural networks. These are: supervised learning, unsupervised learning and reinforcement learning. We’ll take a look at the difference between supervised and unsupervised a little further on in the piece.
Regression analysis
This consists of a set of machine learning methods that predict a continuous outcome variable based on the value of one or multiple predictor variables.
Regression analysis can be used for things like predicting the weather or predicting the price of a product or service given its features.
Clustering
Clustering does what its name says in that its main purpose is to identify patterns in data so it can be grouped.
The tool uses a machine language algorithm to create groups of data with similar characteristics. It can do this much faster than humans can.
Supervised v unsupervised
Supervised machine learning relies on labelled input and output data, but unsupervised does not. Unsupervised machine learning can process raw and unlabelled data.
Clustering uses unsupervised machine learning because it groups unlabelled data.
Skills for machine learning
As we have identified, machine learning professionals interact with data quite a bit. As well as software engineering knowledge, they should have some data science skills.
This piece by Coursera on machine learning skills recommends that people learn data science languages like SQL, Python, C++, R and Java for stats analysis and data modelling.
That brings us on to maths; you will need a fairly solid grounding in statistics and maths to be able to understand the data science components of machine learning.
Being able to critically think about why you’re using certain machine learning techniques is also pretty important, especially if you need to explain your methods and reasons to colleagues with a non-tech background.
Earlier this year, Yahoo’s Zuoyun Jin gave us some tips for learning, based on his experience as a machine learning research engineer.
Learning resources
If you want to brush up on your Python for machine learning, this guide on SiliconRepublic.com points you in the direction of some handy resources.
In terms of gaining a basic overview of machine learning, you might want to check out some online beginner’s courses. This Understanding Machine Learning programme from Datacamp says it provides an introduction with no coding involved.
If you are looking for something more advanced, this course by MIT gives learners an introduction to machine learning as well as ways the tech can be used by businesses. It’s mainly geared towards applying the techniques in a business context.
Last but not least, Google’s Machine Learning Crash Course is a 25-lesson programme that features lectures on the topic from Googlers.
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.
