blog comfone header

Machine Learning - what's it all about?


When I began in my role as Senior Data Scientist at Comfone in 2020, it became my mission to help Comfone to make progress with machine learning.

Machine learning can have a reputation as something extremely complex and impenetrable, and a lot of people find it 'scary' to approach.

However, I have dedicated a lot of my working life to Machine Learning and Artificial Intelligence, so I know that it's a case of understanding the processes behind, and really asking the right questions.

In this blog, I have shared the most frequently asked questions from colleagues and customers, and provided my answers to these questions.

My goal is to break down some of the 'complexities' within the world of Machine Learning, to help you on your journey to understand it better.

In other words:

The only thing to do with good advice is to pass it on. It is never of any use to oneself.

- Oscar Wilde

What is machine learning?

Machine learning involves computers discovering how they can perform tasks without being explicitly programmed.

The outcome of any machine learning is to create a model which can be used to effectively perform the task.

 What are good problems for machine learning?

Essentially, any problem that is not straightforward. Straightforward business problems can easily be defined as a clear sequence of steps and therefore be automated without the requirement of learning.

This does not mean that all problems are equally suitable.

Problems which require intuition, highly involved reasoning or justification are less suitable for machine learning and still require human involvement.

Furthermore, the technical complexity and risk-return trade-off can make a problem less suitable for machine learning. For instance, if the machine does not effectively 'solve' the problem, and it's deemed 'risky' if it goes wrong, it's better that a human does the task and gets it right 100% of the time. 

 How can we benefit from machine learning?

Throughout a machine learning project there are many great opportunities to learn, both for the machines and for humans.

Just to name a few: domain knowledge is broadened by looking at the data behind a problem, assumptions get verified, processes are formalised or re-evaluated.

And finally, once a machine learning model performs according to our requirements, we can utilise the model to address the problem it is designed for.

We therefore save time for humans, allowing them to concentrate on other complex, creative or intuitive tasks. 

 How do we start on this journey?

The first step is to discuss the challenges you are facing with a machine learning expert.

I also recommend reading blogs and articles. Especially the ones about your field of expertise can provide great machine learning examples.

This should quickly give you a good sense of what is possible and provide you with the next steps to develop a machine learning solution.

Once a solution is developed, the sky really is the limit on what you might be able to achieve with Machine Learning!

Staying positive during a pandemic
Why are TDRs so interesting for Comfone engineers?

Related Posts

By accepting you will be accessing a service provided by a third-party external to