Should my business embrace Machine Learning (ML) in 2020?
2020 is set to be an exciting year for businesses using technology. Machine learning (ML) has been around in different forms for a few decades but in the last few years it's reached a level of maturity and availability that makes it useful for many more businesses than before.
What is Machine Learning (ML)?
Machine learning is a technique for using algorithms and statistical analysis to perform a task using pattern matching and inference rather than having explicit code written.
Supervised machine learning algorithms use mathematical models trained with historic or sample data in order to predict future outcomes. Unsupervised machine learning algorithms allow the model to independently discover patterns (even patterns which may not have been previously noticed) within data without prior training.
With platforms such as Microsoft's Azure Machine Learning it is relatively easy to create a model, train the algorithm using sample data and deploy ready-made API endpoints to handle requests from external client applications such as web, desktop or mobile apps.
Artificial Intelligence, deep learning, machine learning — whatever you’re doing if you don’t understand it — learn it. Because otherwise you’re going to be a dinosaur within 3 years. - Mark Cuban
Modern tools allow machine learning models to be designed visually, bringing components into the model, defining the logical flow (splitting and biasing data), the types of algorithms to be used (different algorithms are useful for different tasks) and allowing users to compare how one model performs against other variants and choose the one that best suits the needs of the business.
Is it the same as Artificial Intelligence (AI)?
No, machine learning is considered to be a subset of artificial intelligence.
Artificial intelligence is primarily concerned with enabling cognitive abilities in machines (sensing their environment and adapting their behaviour towards it in a similar way to humans).
What can ML do?
Machine learning is concerned with statistical analysis, probabilities based upon training from historic data or pattern matching based upon previous learning by the algorithm.
This provides us with a customisable tool that can be used for automating repetitive analysis tasks that would usually require human intervention such as approving or declining customer finance applications, automating online help-desks providing answers to customer queries and identifying trends and seasonality for your company's sales.
Machine learning is the science and art of giving computers the ability to learn to make decisions from data without being explicitly programmed. - Dr. Hugo Bowne-Anderson
Where is machine learning heading?
Today's machine learning relies on creating models and training them (supervised or unsupervised) to complete very specific tasks. Every time a new task is required, a new model has to be created from scratch.
The "holy grail" of machine learning is a large-scale model which can perform a large variety of tasks. Most of the time the majority of the model would be idle with just the appropriate functionality being used for a required task. Counter-intuitively this way of doing machine learning is more efficient than the current way of doing it.
How can my business benefit from machine learning?
Most businesses hold large amounts of data about their customer, business partners and the industry that they operate in. Imagine being able to tap into that data to improve efficiency, drive growth and reduce costs. Machine learning can spot patterns that may be holding your company back or trends that are appearing in markets that can be leveraged.
Machine learning can automate, improve and accelerate decision making. Once running, machine learning algorithms provide consistent, repeatable, evidence-based decision making. Whereas a human may make the wrong decision due to fatigue, loss of concentration, poor training or demotivation, a machine learning algorithm will apply the same decision making process every time and use the data collected to improve the performance of the model in the future.
Machine learning has become popular in the finance sector for detecting fraud, algorithmic trading and portfolio management and in healthcare, machine learning is being used for medical imaging and diagnosis and in the development of new medicines.
Many well-known companies like Amazon, eBay and Spotify already use machine learning techniques to provide recommended products, special offers and music to their customers. Research indicates that the vast majority of customers prefer to do business with companies offering a personalised experience.
Airlines are using machine learning to improve efficiency, avoid mistakes, fix problems quickly when they happen and increase customer satisfaction. Universities use machine learning to track student progress and identify students who need extra help or whose behavioural patterns suggest that they're more likely to drop out of their course altogether enabling them to take action before things reach a crisis point.
Should my business embrace machine learning?
Without a doubt, yes.
Machine learning presents a huge opportunity to improve the overall efficiency of your business by intelligently automating repetitive tasks, spotting patterns in your business data to allow you to better plan for the future, all while providing your customers with a better service that's tailored to them.
Businesses that aren't using machine learning are in real danger of being left behind. Machine learning has become widely available at a relatively low cost and is becoming more and more popular within business that I have visited.
If you're interested in learning more about how machine learning technologies can help your business, send me a message on LinkedIn or add a comment to this article.