What is ‘Machine Learning’?


Technology is evolving at an exponential rate, and a plethora of new terms have entered the sector’s vernacular in recent years. One of those is ‘Machine Learning’ – but what does it mean and how does it affect the business world? Business Leader explains.

The term ‘Machine Learning’ is defined by the study of computer algorithms that can automatically improve user experience through the use of data. It achieves this through the use of innovative Artificial Intelligence (AI).

Through its increased use, it can help predict outcomes through the information it has collected – which can then be utilised in almost every sector that embraces the technology.

How Machine Learning benefits cybersecurity

To greater understand its application in a real-world scenario, Business Leader breaks down the impact it has on one of the hottest topics in business right now – cybersecurity. 

Artificial Intelligence (AI) is defined as ‘the theory and development of computer systems able to perform tasks normally requiring human intelligence’. Machine learning (ML) is a sub-field within AI. The pioneer, Arthur Samuel, promoted the term ML in 1959, as the “Field of study that gives computers the ability to learn without being explicitly programmed”.

In the Cambridge Dictionary ML is referred to as ‘The process of computers changing the way they carry out tasks by learning from new data, without a human being needing to give instructions in the form of a program’. And, in the Oxford Lexico, it is used to describe ‘The use and development of computer systems that are able to learn and adapt without following explicit instructions, by using algorithms and statistical models to analyse and draw inferences from patterns in data’.

Generally speaking, ML relies on mathematical models which are built by analysing patterns in datasets. These patterns are then used to make predictions on new input data. Similar to the way Netflix offers recommendations for new TV series, based on previous viewing experiences, ML is one of the many approaches to AI that uses a system that is capable of learning from experience, and builds upon what has been learnt.


People are often scared or apprehensive about what they do not understand. Although ML is not a new concept for experts in the field, many are only just getting to grips with what it is and how it can be used. Because the term ML is associated with depictions portrayed in the media of power-hungry robots with a thirst for human destruction, many recoil at the thought of utilizing it in business or private use. But the truth is that these portrayals are not accurate representations, and ML is used by the majority of us, on a daily basis, without us even fully recognizing how or where.

Examples of daily ML in action includes the use of portrait mode on your smart phone, social media feeds on applications such as Facebook or Instagram, music and media streaming including BBC iPlayer or Netflix, online adverts tailored to the user journey, pretty much every online game, banking apps, smart devices… the list is endless.

From reinforcement learning, semi-supervised learning, self-learning, feature learning, sparse dictionary learning, anomaly detection and robot learning, there are many different approaches and techniques used. But, on the whole, machine learning can be broadly classified into two classes, known as supervised and unsupervised learning.

Supervised Learning

Supervised Learning is where a machine learns from training data, and maps out inputs and outputs, based on rules provided in said training data, and from inferred functions. In Supervised Learning, the dataset is labelled, wherein there is a target variable. The value of which the ML model learns to predict, using different algorithms. For instance, it may do this based on IP address location, frequency of web requests and so on. From this, an ML model can then predict if the IP was part of say a Distributed Denial-of-service (DDoS) attack, and more.

The main goal is for the machine to extract the information from the unlabelled data sets, that could aid performance and increase productivity.

Covid-19 pandemic is accelerating UK Plc’s use of Machine Learning

Appsbroker, the largest Google Cloud-only Managed Services Provider in EMEA today published a comprehensive report of the health of Machine Learning in the UK private sector.

Based on an extensive survey* amongst corporate UK IT leaders and decision makers, the report is the most comprehensive nationwide analysis of the state of Machine Learning (ML) in the UK. The insight the report provides is overwhelmingly positive, pointing to significant technical and commercial progression, while also underlining the challenges that must be overcome if ML is to be validated as one of the most transformative technologies for a generation.

Four key findings have emerged from the survey data:

  • In the first instance, ML for UK business is no longer a hypothetical possibility; more companies now are validating ML or are using it in a live business context than the proportion of companies that are still assessing its merits and applicability. 69% of UK companies report that ML will be relevant to their organisation in the next 36 months; with exactly a quarter saying it is relevant right now and a further 17% within 12 months. For this reason, the report has dubbed ML as a ‘now thing’ for business, reflecting its advancing maturity and commercial potential.
  • 39% of organisations surveyed are accelerating their ML initiatives in response to the Covid-19 pandemic. Only 11% indicated that they were in any way de-prioritising ML activities while for half of respondents, their ML plans have remained unchanged by the pandemic.
  • While the investment levels businesses are prepared to commit to ML solutions are advancing, with 57% of companies expecting to invest more than £1m over the next 24 months, the report also contains a striking insight that the average UK company adopting ML expects to see a 10x return on investment. If realised, ML will become one of the most productive contributors to enterprise profitability of recent times.
  • The biggest barrier for delivering ML is the lack of in-house skills (cited by 44% of companies).
  • The report disavows a conventional wisdom that ML is destined to devour jobs and replace human endeavour. In fact, twice as many businesses (40%) saw ML as enhancing human roles in business as those who regarded the technology as central to automation and cost-cutting (20%).

Despite the overwhelmingly positive outlook, the report does however issue a stark warning to business that ML must shift constructively from being the preserve of analysts and technicians to a position in the heart of the business. 68% of businesses reported that ML sat within their IT or Data capability, with only 16% of companies confirming that the capability had been more fully integrated into commercial operations.

Many respondents identified a lack of understanding and leadership for ML from senior executives as a major potential stall point for the technology and something that needs to be overcome to ensure ML’s successful evolution as a primary toolset in UK businesses. 

Henry Brown, Head of Data & Machine Learning at Appsbroker said: “The report paints a very positive picture of Machine Learning adoption and value across a diverse spectrum of enterprise Britain. Clearly this picture is changing, but it is good to note that many of the indicators point to growth, increasing maturity and an expectation of very positive value, which is consistent with the experience we have at Appsbroker as we guide businesses through the process of ML adoption.”

Industry sector highlights include:

  • Financial Services organisations are the most likely to see ML as relevant to their business now or in the next 12 months (57%), followed by Healthcare (50%), Retail (48%) and Manufacturing and Industry (46%).
  • The investment champions are undoubtedly Financial Services organisations – almost a quarter (24%) of firms confirmed an intent to spend more than £5m on ML in the next two years, followed by the Technology sector with 21% of companies looking to make a similar level of investment.
  • Over half of Retail (57%) and Financial Services (52%) sector organisations are using ML for Fraud Detection to mitigate high financial risk and crime.
  • The use of ML for customer segmentation and marketing is being deployed most among Professional Services organisations (65%) ahead of the Retail sector which ranks second (with 52%).
  • A lack of in-house skills to embrace Machine Learning is most pronounced in the Healthcare sector (30%) and of least concern to the Manufacturing (13%) and Technology (14%) industries.
  • Of the third-plus businesses choosing to accelerate their focus on ML due to the pandemic, Healthcare and Technology companies are the most prominent sectors with
  • 45% of companies in both categories planning to accelerate their investment in ML.