Most people can agree that our machines are becoming faster, better, and, ideally, cheaper. But more intelligent? That’s where some people draw the line and many outside of technology spheres get a bit nervous. Well, here’s your spoiler alert: you are already highly reliant on intelligent machine learning.
When you understand what machine learning is, you can begin to see its potential and appreciate what it does to enhance our lives in and outside of work.
Arthur Samuel is attributed with coining the term machine learning and defined it as: giving computers the ability to learn without being explicitly programmed. In other words, the computer becomes intelligent by recognizing patterns and data sets on its own, and improves with experience.
Today, machine learning is ushering in the democratization of data. We no longer need highly specialized data scientists to develop highly specific algorithms. Machine learning has the capability to provide a more general approach that can be applied to a wide variety of problems. This prevents engineers from spending time “reinventing the wheel” and instead allows them to spend time understanding what type, level, and frequency of data is needed.
Machine learning algorithms are best utilized for processing large amounts of data and different data sets – work that would require significant effort for a person to input and then attempt to extrapolate meaning.
For example, imagine attempting to accurately predict the price of a home for sale. To do that you need to know the square footage, its key features including number of bedrooms and bathrooms, size of the lot, and so on. By combining five or 10 features with the previous sale price (single data point) and comparing against: homes on the same block (small sample), neighborhood (larger sample), community (even larger sample), an accurate prediction on sale price emerges. With machine learning we can use thousands of data points to gain access to valuable insights at our fingertips.
THE THREE TYPES OF MACHINE LEARNING
A common question from C-suite leaders who are curious about machine learning is – how can it benefit my business and how do we know if it’s right for us?
The short answer is: it’s right for any type of business that makes decisions based on data. And machine learning is only going to become more integrated into business practices going forward.
The first step in implementing AI is determining where it will be most beneficial. That knowledge comes by understanding which of the three types of machine learning can best solve a pressing business problem.
Supervised learning turns data into real and actionable insights, delivering a clear answer from the data fed to the machine algorithm. This type of machine learning is often used to identify people or objects in images, but can also be used to recognize and prevent unwanted outcomes. For example, many of us have received alerts from our credit card companies to verify a purchase that is outside our normal spending habits. This application uses supervised learning to detect what is a normal spending habit, things like location and price, and then take appropriate action when there is a deviation.
Unsupervised learning does not start with the answers in mind; instead, you’re leaning more into discovering what the data reveals. With this type of machine learning, the algorithm is processing vast amounts of data and making correlations: for the purpose of improving experiences, reducing the use of resources, and/or increasing success. Think about Amazon or your favorite online retailer. If you just bought a suit, the algorithm might recommend a shirt or an accessory. This is machine learning that finds correlations and positions data as a prompt like “customers also bought” to help drive higher sales.
Reinforcement learning is the training of machine learning models, through reward and penalty, to make a series of decisions. Remember the hot and cold game you played as a kid? This is the same idea. By giving positive and negative feedback – cold, cold, warmer, warmer, hot! – the machine and its algorithm can learn to choose the right option over the wrong one. For example, if we are teaching a robot how to vacuum your home, we would give it a series of rewards or reinforcements as it chooses correctly or penalties if it doesn’t. It might be a small reward for every new section the robot cleans, a big reward when the entire room is clean, and a negative reward when the robot cleans a spot that has already been completed.
Turn your data into immediate business value?
Our AI assessment will help you better understand your data, identify feasible AI opportunities, and provide strategic recommendations on how you can add a whole new level of intelligence to your business model.Learn more
MACHINE LEARNING AND THE IMPACT ON JOBS
So, what happens when accounting gets a new payroll system or sales invests in a lead generation tool powered by machine learning? What happens to those jobs?
This is a fear that demands an answer. Machine learning isn’t the job killer we’ve imagined it to be, rather it is better viewed as a role enhancer.
When clerical responsibilities such as invoicing and billing benefit from machine learning, the tasks get streamlined, but the accounting work itself doesn’t disappear. Instead, individuals, who previously spent time on tedious and recurring tasks, shift to do more valuable work and solve problems. Repetitive tasks should be automated, instead of making humans do mundane, unrewarding work. As a result, people can spend time doing higher value work. In short, jobs don’t go away; they get redefined. And employees remain, because their insights, problem-solving skills, and empathy are invaluable and not machine replaceable.
FACING NEW OPPORTUNITIES
Recently I had a request to provide a photo of me and my wife. Thanks to facial recognition – a machine learning algorithm – I was able to avoid a long search across all of my pictures in Google Photos. By saying “Hey Google, find a picture of me and my wife”, Google’s algorithm found photos with just the two of us in it. This saved me a ton of time searching for a picture and allowed me to spend my time looking for one that looked good. With a deep photo archive, what could’ve taken 30 minutes, or an hour only took seconds.
While this was a personal task, it was still a time-saver. Now think about voice recognition, text recognition, the online retail predictor, your navigation app – all these little things add up and save time so we can get to what matters.
Similar use cases exist for businesses. There are likely internal functions that your team could stop doing with the help of machine learning – and plenty more external opportunities as you look to enhance the experiences of customers and partners. To start the process of considering AI for your business and your customers, learn about the 5 considerations before diving into AI here.
When we are willing to learn what the intelligent machine can do on our behalf, we will all be significantly more productive and happier doing the work we were designed to do.