• Tuesday , 12 December 2017

Futuristic Machine Learning Implementations in Mobile Applications

The future is now, and the future is machine learning (ML) and recurrent neural networks. At least in terms of mobile app trends that are expected to continue from 2017 into 2018.

So let’s take a look at the various scenarios in which app developers and companies are using artificial intelligence (AI) in their mobile applications to see if we can’t find some worthwhile features to inspire our own projects.

  1. Facial Recognition

First up is Snapchat’s purchase of the Ukrainian startup Looksery. This company developed a way to recognize faces and then apply various filters and visual effects to the people that the software found. This means you can tell the app to add a pair of glasses, and the app will automatically know where to place this, based on the facial features.

If you think about it, not only does it know the difference between a human face and anything else, but it knows the exact shape and dimensions of every facial feature, and can use this data to map new information onto the existing properly.

  1. Smart banking

Oval Money is a clever little app that uses ML in a collective sense. By connecting your bank account to the mobile money-saving application, you can compare your expenses with other people’s, and find out where you’re paying more than the rest of the users – and get notified. This has meant people have saved several hundreds of dollars each month in their budgets, simply by changing up their expenses.

The fact that all expenses are shared with each other, and compared to each other means that if enough users start using this app, we could all be driving down the prices on everything from insurance to groceries. By realizing your neighbour only pays half the amount for his car repairs as you do, you can switch to that other mechanic and put your savings to better use for instance.

  1. Saving the whales

A team of researchers, developers and all-around nature lovers have been experimenting with machines learning to recognize the melodic language of the whales. Because the shipping industry is increasingly expanding their reach, so are whales getting hit by ships.

In 2006 alone more than 35 whales were recorded being hit by an ocean going vessel, which prompted the team to try and develop a whale-sonar based on audio recognition.

  1. Support Ticket monitoring and handling

The well known developer forum Stackoverflow aims to develop a machine learning application that will predict which questions will be answered and which will be closed within a given time. This has been done to combat the large number of questions that are either below the quality average, or deemed unanswerable for any number of reasons.

  1. Using Twitter to predict psychopaths

By putting a machine to crunch data from a number of Twitter profiles, back in 2012 a competition was launched for a machine learning program to identify psychopaths based on their public Twitter metadata. Profile picture, language, use of words and topics written about was among the different sets of data analyzed. Who knows which government will take this to the next level?

Conclusion

So while we can’t yet hold a meaningful conversation with an AI, machines are evolving every day with all of the major technology brands such as Google, Facebook, Apple and Microsoft investing heavily into the industry. It’s only a matter of time before an AI becomes clever enough to pass as a human in more than just the turing test.

Related Posts