Machine learning is changing the way we interact with our mobile devices. Our phones and tablets are now powerful enough to run software that can learn and react in real-time. This has opened up the door to some cool applications.
Startups and tech giants are all starting to use machine learning in mobile app development, and they’ve come up some interesting ideas. Let’s take a look at 10 examples you can download to your phone right now.
1. Snapchat Filters
Snapchat started taking machine learning seriously when they acquired the Ukrainian computer vision company Looksery for $150 million. They use Looksery’s clever facial tracking algorithm to find your face in your snaps and add things like glasses, hats and doggy ears.
Recognizing a face is easy for humans but difficult for computers. Explicitly programming a computer to recognize a face is almost impossible. Instead, Snapchat has its algorithm look at thousands of faces to slowly learn what a face looks like. Each picture has all facial features such as eyes and nose marked by humans. Check out this video to see how it works in more detail.
2. Oval Money
Oval money takes a different approach. The app uses machine learning to help save you money. The CEO explains that “Oval combines machine learning with the lessons users can teach one another to create collective intelligence”.
By looking at your spending habits and collective knowledge from all users, Oval creates a money saving strategy that’s smart and easy for you to follow. Early users in the UK saved hundreds of pounds in just a few months. And, as it’s a machine learning app, so it gets better every time you use it!
3. Carat App
Battery life has always been a problem with mobile devices. Power storage just isn’t improving at the pace of other technologies. This has encouraged some researchers to look into other ways of improving the battery life of your phone.
Carat monitors all kinds of activity on your phone and gives you suggestions on how to reduce power usage – and not just obvious things like “turn your brightness down”. Developed by Ph.D. students, Carat learns how you use your phone and can actually tell when there is a problem. You’ll be notified when one of your apps is broken and needs re-downloading, or when your phone is due for a restart.
Just as Shazam can hear a song and tell you the artist and title, LeafSnap aims to tell botanists the species of a tree from a photo of a leaf. Leaves are by far the most common type of fossil, and determining the species of these fossils is hard work.
By looking at 1000s of photos of leaves, LeafSnap’s algorithm has learned to identify many of them. Like many machine learning applications right now, it’s not 100% accurate. But, it’s good enough to be a great tool for scientists and it’s improving all the time.
Machine learning isn’t all about science. Apps like Dango are attempting to tackle the real problems in life, like finding the perfect emoji.
Dango uses deep learning (a form of machine learning) to actually understand what you mean when you type. It’s learned from looking at millions of comments and messages that use emojis, and it can even understand things like emotions and jokes. With this knowledge, it then suggests emojis and GIFs to enhance your texting.
We all have that task we’ve been meaning to do but just can’t seem to squeeze in the time. That’s what ImprompDo can help you with. The app monitors things like your location and what your doing and learns the best time to remind you to do a task. It even takes care of things like prioritization. There are other similar apps like Google Calendar.
7. Aipoly Vision
Computer vision has improved a lot in the last few years. Mostly thanks to machine learning. Apps like google photos can recognize what’s in an image and tag your photos.
Aipoly is slightly more ambitious. The mobile app can recognize objects in real time from your phone’s camera. Just point at an object and Aipoly will tell you what it thinks it is. The plan is to help the blind and visually impaired with day to day tasks. It’s certainly not perfect yet, but like most machine learning applications, it gets better every time someone uses it.
8. Swiftkey Neural
SwiftKey is an app that makes typing on mobile devices easier. There are a few apps like this. They basically look at the last few words you typed and take a guess at what the next word will be. The problem is that these apps aren’t very good.
SwiftKey Neural aims to change that by using a machine learning technique called neural networks. This method allows the app to get a much deeper understanding of the context of a conversation, and give better suggestions. This is part of a bigger trend of smart applications that make using mobile devices easier and faster.
9. Sea Hero Quest
This one is in a completely different category to the other apps. Rather than using machine learning to power a cool feature, Sea Hero Quest gathers data from users. This data is then used by scientists to train machine learning software that helps with dementia research.
The game is very carefully designed to test the player’s spacial awareness. Every time you play, your information and scores are anonymously used to gain a better understanding of the human brain. That’s right, just by playing an addictive game you can help cure a disease that affects 45 million people worldwide!
10. Train Your Tic Tac Toe
Most of the machine learning mobile applications I’ve introduced are already trained when you download them. This last one you have to train yourself.
When you first start to play tic tac toe against this app, it’s terrible at it. At first, it only knows the rules of the game, and that’s it – no strategy or tactics at all. However, as you play more games, it slowly starts to learn how to play. After many games, it gets really good. This is an awesome app if you’re interested in actually watching a computer learn.
Do want to start building machine learning into your apps? Development teams like devteam.space can help you out. Who knows, maybe your app will be the next Snapchat!
Lolita Rogers is a technical writer with five years of experience in IT. Her prime area of expertise is Mobile App Development.