Machine learning is everywhere. In fact, it’s already at the point where you’re probably using it in your day-to-day life without ever thinking about it. When done well, machine learning adds value by making our interactions with technology more seamless and automatic. When done really well we don’t need to interact with the technology at all – all the work is done in the background on our behalf.
Machine learning is a practical application of AI, an innovative approach to technology that we also use daily. These are just some of the ways in which machine learning, through AI, is now enhancing our daily lives and the way that we work, live and play.
First, though, what is machine learning?
Machine learning is complex technology, but a simple concept: basically, it refers to applications of technology where the software “learns” how to do something by being fed a lot of data. For a simple example, consider a maths problem. By feeding the application the correct answer to a maths problem, the application can “learn” to recognise an incorrect answer instantly, and without human input.
The latter is important: the point of machine learning is to “teach” it to operate without human supervision. That way, the technology can run 24/7 and do things at far greater speed than if a human was involved in the process.
How does that then benefit us? In short, it allows us to automate a lot of the mundane tasks in daily life, freeing us up to get a lot more done with our days, with greater convenience. Here are just some examples of how you’re probably encountering machine learning every day.
1. We now know exactly how long it will take us to get somewhere.
If you’re older, then you probably remember carefully looking up the schedule for trains, or keeping an eye on the traffic report for an indication that congestion was going to slow down your trip. From there you would calculate when you needed to leave home to be there on time.
Now, however, you just pick up your phone and look at the maps application where the app will have already recorded the traffic and/or any disruptions to public transport, and calculated for you how long it will take to get to your destination (and therefore when you need to get going to be on time). This is only possible because the apps have been “trained” through machine learning to make those predictions and calculations on your behalf.
2. Our lives are more secure now.
This application of machine learning is controversial, as there are privacy considerations, but when used correctly and ethically, it can significantly improve the safety and security of venues and other public areas. Machine learning can be “trained” to identify faces in a crowd, and from there cross check those faces against databases to raise red flag notifications with security or the authorities as necessary. For example, a store could use this technology to identify a frequent shoplifter, or a club could use it to identify someone that should have been banned from entry.
Taking steps one step further, it’s increasingly possible for the technology to also “learn” how to read intent and behaviour. This has application at big events or sensitive locations, such as sporting fixtures or airports, to detect mannerisms that might be tied to a terrorist attack.
3. Cities run much more efficiently
In recent years, the more advanced and innovative metropolises around the world have started investing in a concept called “smart cities”. What this basically means is the deployment of IoT sensors and other devices to collect data on how a city is operating, and then leveraging machine learning applications to cause technology end-points to respond to the data.
A really simple example of this in action is traffic management. In a smart city, sensors and cameras deployed at traffic lights will monitor congestion, and adjust the timing of when lights would change to facilitate a better flow of traffic through the lights. To be able to do that, the machine learning application has to be “taught” how to recognise when congestion in one direction is getting a bit too much, and then the appropriate response to ease that congestion.
So, if you find yourself spending less time frustrated at lights, that could be because your city is investing in machine learning to become a smart city.
4. We find the information that is of interest to us much more quickly
Machine learning is of intense interest to digital marketers, social media companies, streaming entertainment services and online retailers, and the reason is simple; the more tailored the information that you see, to more likely you are to have a good customer experience, spend longer on the service, and spend money.
All digital companies collect customer data, and they use that to personalise the experience. Say you’re interested in tennis, and spend time on a tennis group in Facebook, or use #tennis hashtags on Twitter. You’ll start getting advertisements for tennis equipment and/or tickets to tennis events served to up to you. After clicking on those ads and buying a tennis racquet, you’ll start getting email advertisements from that retailer when tennis balls are going on sale.
None of this is happening because someone at the social network and retailer is stalking you and manually building a profile of your interests. Rather, it’s that your data is being analysed by AI and machine learning, and then you’re being specifically targeted with more information related to your interests. That in turn makes it easier to find the things you want to buy when you need them.
5. It’s getting easier to keep you healthy.
Medicine is a data-driven industry, and they’re embracing machine learning in a big way to help improve patient outcomes. You’ll find machine learning used in everything from diagnostics to the planning of treatments and therapy. Machine learning can also be used to assist with effective monitoring of patients in a facility, so that rather than relying on nurses rounds, the monitoring application can raise an alert if there’s an anomaly detected.
Machine learning can never completely replace doctors, nurses and other skilled practitioners, of course. There’s far too much of the human involved in medical treatment. However, machine learning can assist doctors to work more efficiently and with greater accuracy.
6. It’s helping you to stay secure online.
Security is a big concern online. The number of cyber attacks are accelerating, and they’re becoming more sophisticated. What’s more, we’re working remotely more often now, and therefore we’re no longer protected by the corporate network in the office.
This doesn’t mean we need to stop working remotely, or that our mobile phones and other “personal” devices can’t be used for work as we are using them now. Most IT security companies are investing heavily in using machine learning, hosted in the cloud, to detect threats in real time. The machine learning algorithm is taught what “good” data looks like, and how malicious code behaves. It then continually scours the Internet for signs of the latter, and if it detects any it will raise the alert (so a solution can be found by the vendor), and will proactively block that data from your devices, protecting and isolating them from the malware.
With the speed that new threats are coming online, machine learning’s real-time, 24/7 capabilities are really the only way that security companies can continue to protect their companies – including your devices – from the latest concerns.
7. Self-driving cars
Okay, so this isn’t an everyday application of machine learning just yet, but it’s likely to be in the not-too-distant future, and in many ways represents the holy grail of the technology. Just consider the number of “thought” processes a self-driving car will need to undertake. Not only would it need to be able to “read” the traffic conditions and respond appropriately. It will also need to be able to choose the most efficient route to a destination, so it needs to have machine learning-capable mapping technology into it.
Perhaps most critically of all – and where a lot of the debate around self-driving cars is centered – is how the car should respond in a life-threating incident. Say a crash is inevitable – the machine learning application needs to be taught how to take steps to protect as many lives as possible, and if the statistical chance is that there will be fatalities, the application needs to have a triage system in place so it can determine who to protect first. This is actually a major point of ethical discussion within the AI industry, and self-driving cars are a clear example of why it is so important to get it right.
These are just some examples. Whether you’re using home automation devices from Amazon and Google, virtual assistant tools like Apple’s Siri, chatbots when looking for customer support, or cloud services to work and play, you are using AI and machine learning. It is prevalent, pervasive, and making our lives fundamentally better and more convenient.