You’ve probably heard the term “Machine Learning” being thrown around a lot lately, or certainly a lot more than you did before.
Recent years have seen Machine Learning (ML) become more and more ubiquitous in the tech scene.
And though the phrase has existed since the late 1950s, it is only recently that the potential implications of this technology are starting to be explored properly.
In some ways, ML is being used nowadays as a sort of buzzword with little-to-no awareness of what makes it distinct.
It is often confused with Artificial Intelligence (AI), but can more accurately be described as a form of AI in which computers that are fed extremely large data sets are able to learn as changes in the data occur.
Whereas AI is seen as the science of building machines capable of human intelligence, ML is (according to Stanford University) “the science of getting computers to act without being explicitly programmed to do so.”
Ever wondered how Google knows what you’re looking for, even when you’ve typed something incorrectly, or how on earth you’re even able to have a conversation with chatbots like Siri or Alexa? A lot of it is down to ML and predictive learning.
As a branch of AI, Machine Learning works in tandem with data science, wherein large data sets provide the ability for machines to learn and develop from.
What can Machine Learning actually achieve?
It’s plain to see that many businesses have become aware of this trend for ML and data science, and have thus latched on by figuring out how they can best make it work for themselves.
Whether it’s retail, banking, sports, or more, it’s difficult to think of even one industry where there is no practical application of this technology. Generally, when one thinks of AI or ML however, the mind forms a very futuristic and often unrealistic idea of the concept.
Driverless cars, for one thing, have been a very big talking point for years now. Even considering various setbacks within that market, they remain fresh and exciting ideas that continue to receive a lot of funding.
However, at this point in time, a concept such as a driverless car remains just that: a concept. Without the necessary research and development (of which a great deal more is needed on top of the plenty that has likely already gone into it), applications such as these exist as almost entirely abstract ideas for now.
Cutting edge development such as Amazon’s drone delivery services “Amazon Prime Air”, for instance, are exciting and do serve to fuel public interest in these new technologies.
Nevertheless, there are far more realistic and practical applications capable right now thanks to ML, such as the ability to exponentially improve supply-chain processes – and though this perhaps doesn’t sound as sexy as the fantasy of driverless cars, it’s certainly far more exciting given what can be achieved in the here and now.
What are businesses doing to stay ahead of the curve?
Over time, businesses have become aware that the mass amounts of data they have can be extremely useful and profitable, perhaps even relevant to other parts of the business that they had not previously considered.
An interesting early example of this is Tesco, which existed for years as your run-of-the-mill supermarket. During the mid-nineties, however, the company made the decision to expand into the realms of finance and telecoms, beginning to seek out new opportunities.
This led to Tesco having access to huge amounts of data from its customers, via the cross-sections of these specific and distinct industries. It was also around this time that the company introduced the Tesco clubcard, its own loyalty scheme. Though many competitors had similar schemes, they had been dismissed as unprofitable and did not consider the possibilities that could come from this.
Tesco soon understood the massive potential such a scheme would allow for and used this data to understand exactly what its customers were looking for and to target all the more effectively.
Since then, many companies have followed the way paved by Tesco and others. And today more than ever, businesses are looking into how exactly they can capitalise on the data that they have.
In recent years, there are a number of areas in which these processes are being used especially effectively. For one thing, supply chain logistics are being effectively improved thanks to the implementation of AI. Going forward, ML will have (and already is having, in many cases) a substantial effect on such processes. In order for most businesses to succeed, it’s essential to have a properly managed supply chain, and ML is effectively placed to improve the accuracy and efficiency of supply-chain management.
Because it generally involves gathering massive amounts of data on a day-to-day basis, the capabilities of ML (i.e. predictive analytics, ability to process large amounts of data) are especially exciting and innovative, particularly in large or global business that are likely to handle huge amounts of data relevant to their supply chain.
The possibilities opened up by emerging technologies such as ML have no reasonable end in sight, and practically every industry has a potential use for it. As I’ve already mentioned, there are countless sectors whereby the use of automation can increase productivity and profitability.
ML is affecting even the recruitment industry, as Cathcart Associates has the opportunity to see firsthand.
Though remaining mostly somewhat hypothetical at present, for now there are quite a few rudimentary but novel applications for ML in the recruitment sector.
Recruitment for high-volume positions such as call centres and customer service, for instance, can implement ML to match keywords to phrases in a candidate’s CV, with the basic idea being to judge an applicant’s relevance to a job.
How will the growth of ML and data science be affected by the skills shortage?
It’s clear that there is an increased appetite for ML and data science within the business sector. It would take just a cursory internet search to see that countless organisations are now suddenly on the lookout for data scientists where previously they might not have been.
All this is occurring amidst the growing skills shortage in the UK, which can naturally pose a significant issue.
While it’s difficult to say with certainty, however, the UK is not in an entirely unattractive position going ahead.
Scotland, especially, is well placed to meet these kinds of demands in the face of the skills shortage, being home to the Edinburgh University School of Informatics and Stirling University, both of which are already producing well-qualified data scientists.
It’s a positive outlook, but speculative nonetheless. What remains beyond doubt is that ML and data science are going to be increasingly important in years going forward, with potential beyond our imagination right now.
Most organisations are being sensible and focusing on the practical applications of the technology. Who knows where we’ll be in ten years time, however.
For now though, we’ve barely even scratched the surface.
Sam Wason & Gordon Kaye, Co-Founders and Directors of Cathcart Associates
Image Credit: Zapp2Photo / Shutterstock