Google AI team improves Linkify on Android Pie

The Google AI team strives towards solutions that make the user’s life simple. With the launch of Tensorflow, the deep learning team has taken the power of machine learning to almost all the Google Products. Smart replies on Gmail, better photographs on Google camera, translation and text completion are some of the features that we have been using for a while. And now, Android Pie stands at the top of the line for an enhancement to Linkify.

Android Pie

Google has been at the forefront of AI development the past couple of years. It has embedded AI deep within Android  9(codenamed Pie) to make the smartphones personalized and intuitive for users. Though most of these features are subtle and don’t bring about a massive difference in usage, some enhancements are noteworthy.

The Android Linkify adds clickable links to texts. This allows users to take faster action. For example, if your friend sends you a text message containing a phone number. The phone numbers are usually clickable, and upon clicking, it opens it up in the dialer app. Here’s an example on the phone

linkify

The linkify here is identified based on a set of rules. If none of the rules match, then no link will be added to it. And so, we wonder why a phone number is not getting highlighted.

Now you know!

With the Android Pie, the Linkify no longer works based on a set of rules. Google has used machine learning to train the Android linkify system to identify different types of data and add clickable links appropriately. This is called Smart Linkify.

Smart Linkify

android Pie

The most notable feature of the Smart Linkify feature is the ability to detect addresses and turn them into clickable links. These addresses are auto-detected and you can directly open them through Google Maps. Along with addresses, the existing features such as phone numbers, email, etc are detected in a much smarter way.

The reason they are smarter is because of the Neural network that has been put in place. It will not follow rules but try to understand what is being talked about. It tries to identify patterns by analyzing text around the area of interest. Here is an example:

If you received a text which read, “Your order is shipped, and the confirmation number is 857-555-3556.” Here the area of interest would be the number “857-555-3556” as it looks like a phone number. Earlier versions of Linkify would have recognized this as a phone number, but the Smart Linkify on Android Pie will identify that the context does not hint for that number to be a phone number and it will not add the clickable link to it. This is the power of AI.

If you are interested in learning how to build AI-powered applications and how machine learning works, I would recommend you to check out this course on Machine Learning and Deep Learning.

Android Pie Developers

It’s good news for Android Pie developers as they can efficiently use all of these features while building mobile applications. The Smart Linkify feature has been open sourced and available as a TensorFlow API on the Android Pie.

Source: Google AI Blog

akshay pai

I am a data science engineer and I love working on machine learning problems. I have experience in computer vision, OCR and NLP. I love writing and sharing my knowledge with others. This is why I created Source Dexter. Here I write about Python, Machine Learning, and Raspberry Pi the most. I also write about technology in general, books and topics related to science. I am also a freelance writer with over 3 years of writing high-quality, SEO optimized content for the web. I have written for startups, websites, and universities all across the globe. Get in Touch! We can discuss more.

Leave a Reply