Machine learning or artificial intelligence? Over the past few decades, these two terms have been interchangeably used, all the more so because of its capacity to improve productivity, planning, execution, and profits for companies across the globe.
Let’s allay the confusion first. Machine learning and artificial intelligence are connected. In fact, machine learning is a subset of artificial intelligence. It uses statistical techniques and computing power to create systems that can learn from the databases and databases of data (read Big Data) available to it.
Machine learning has the capacity to evolve. Machine learning algorithms make use of various programming techniques to process large amounts of data, detect patterns and trends, and extract useful information. This information is used to learn, adapt, and create new versions that have improved performance when compared their previous iterations.
Currently, Machine learning algorithms are trained using three methods: supervised learning, unsupervised learning, and reinforcement learning.
Michael Jordan, professor of Computer Science and Statistics at the University of California, Berkeley noted in a medium post that “Although not visible to the general public, research and systems-building in areas such as document retrieval, text classification, fraud detection, recommendation systems, personalised search, social network analysis, planning, diagnostics, and A/B testing have been a major success — these are the advances that have powered companies such as Google, Netflix, Facebook, and Amazon.”
In a recently published study by the Economist Intelligence Unit for SAP, the survey of 360 organisations showed that around 68 percent of survey respondents were already using ML, in some form or the other to enhance their business.
Today, machine learning is being used in many amazing ways to impact our daily lives, business decisions, and operations, both visible and behind the scenes. Here are some wonderful and inspiring examples of machine learning in action.
There’s a good chance that you may remember or have read about the chess match between Gary Kasparov and IBM’s Deep Blue. Deep Blue won. If you are tempted to chalk that down to coincidence, consider the ancient Chinese game of Go. It is considered much more difficult for computers to master than chess. And yet, in 2016, Google DeepMind’s AlphaGo defeated the world Go champion Lee Dedol. AlphaGo was not only trained by observing matches of world champions repeatedly, but also by having it play against itself in numerous matches.
Amazon has been investing deeply in artificial intelligence and reaping benefits from machine learning since the 1990s. In the words of Amazon itself, “ML algorithms drive many of our internal systems. It’s also core to the capabilities of our customers’ experience – from the path optimisation in our fulfilment centres and Amazon’s recommendations engine to Echo powered by Alexa, our drone initiative Prime Air, and our new retail experience, Amazon Go.”
Google and other search engines use machine learning to observe and adapt to the search patterns of users. They track how long a person spends on a page of results, whether they go to the second or third page to find more information, and also if they go past the second and third page without clicking on any of the links. If the person spends most of their time on the first page, Google concludes that the results displayed were accurate. If not, the search results are further adapted until the user is satisfied.
Netflix uses machine learning to curate its large collection of TV shows and movies. The movie recommendations you get while watching Netflix come from machine learning from each user’s streaming history and the watching habits of millions of its users with similar tastes.
Yelp is a crowd-sourced review platform rich with ratings, reviews, and pictures of restaurants, gyms, venues, bars, and more. Yelp uses machine learning to sort through these image uploads and group them into categories such as menus, food, outside, inside, etc. This helps users find the information they are looking for quickly rather than scrolling through loads of images.
Waymo is Google’s autonomous vehicle project offspring. Waymo’s ambition is to create driverless cars. However, achieving this feat needs some serious artificial intelligence support. Through machine learning, Waymo cars can sense their surroundings and make predictions about how other elements/ people will behave. This helps them navigate through the roads with minimal risk. The varying degrees of variables on the road make advanced machine learning indispensable to the success of Waymo.
Duolingo is a revolutionary language learning application designed to work just like a game. Duolingo makes language learning fun by using interactive graphics and simple quizzes. However, it maintains its effectiveness by relying on machine learning to determine how much time a user can go before they need a refresher course. Armed with this information, Duolingo pings users to take a refresher course periodically, ensuring that learning remains fun and achievable to them.
Two evident ways in which Facebook uses machine learning is in recommending people you may know and facial recognition. Facebook notices the friends you connect with, profiles you frequently visit, your interests, the posts you like, and groups you share with other people to make intelligent friend suggestions. Facial recognition of people in the pictures you upload is also a complex application of machine learning.
Thomson Reuters, the news and data services group, and Tamr, an enterprise data-unification company partnered and used machine learning to unify more than three million data points with a 95% accuracy. They were not only successful in reducing the time needed to manually unify the data but were also able to cut labor by 40%.
Ever wondered how Pinterest manages to make similar pin recommendations so well? Pinterest makes use of computer vision that relies on machine learning to identify objects in images with similarities.
Google Maps and other dynamic route map apps use machine learning to make real time suggestions on route changes depending on real time traffic information obtained from the smartphones of multiple users.
Uber ATG engineering lead, Jeff Schneider revealed that they used machine learning to predict rider demand and accordingly project prices in surge hours. This helps the app estimate the price of the ride more or less accurately, setting the right expectation with users.
GlaxoSmithKline, the pharmaceuticals group, used machine learning algorithms to sift through comments left by parents on forums and messaging boards about vaccines to develop content specifically designed to address these concerns and allay their fears.
In the retail sector, Walmart, which used machine learning to optimise home delivery routes, prevent theft, and offer stellar customer service, just stepped it up a notch with their facial recognition machine learning software. The software detects frustration in the faces of customers at the checkout point and directs customer service representatives to intervene.
Vale, a Brazilian mining group, uses machine learning to optimise maintenance and significantly reduce purchase requisition rejections due to operational delays. They successfully reduced rejections by 86% with the help of machine learning.
Across the world, enterprises are now harnessing machine learning to deliver significant benefits to customers, employees, business owners, and investors. The vast amount of data generated in this digital age is being put to use through machine learning to create an efficient and optimised future for all. If you haven’t invested in machine learning yet, consider giving us a call to learn how it can improve your business.