Google IO => AI, 2018

Last year at google IO ‘AI first approach’ was promoted by google, moving away from ‘Mobile first approach’. Today’s Google IO was a reflection of that. Google’s path is ‘AI must approach’.

Less on products such as android or flutter or firebase or any usual google developer tools. All the product announcements, demos and solutions were on AI.

Let’s take a look at the highlights of Google IO 2018.

Google Digital skills – Google investing high on training students on digital skills.

Google AI – Opening more AI centers around the world. Some of the use cases where google AI powered solutions are implemented –

  1. Healthcare
    • Diabetic retinopathy diagnosing. With the same image able to predict various factors like age, sex, smoking habits, A1C, BMI and even blood pressure using machine learning.
    • Help doctors predict medical events – Predicting when the patient might get sick again.
  2. Accessibility
    • Closed captioning – ML – Looking to listen. Combining visual and audio cues presenting in a more understandable form.
    • Morse code – Gboard adds morse code as input
  3. Gmail
    • Smart compose – Smart auto complete. Not just words. But the entire sentence is also based on the context
  4. Photos
    • Suggested actions – Actions based on the context of the photo
      • Photo corrections
      • Sharing options
  5. TPU 3.0
    • 8X powerful than 2.0
  6. Google assistant
    • Wavenet – 6 new voices added to the google assistant
    • Coming to 40 car brands
    • Natural conversation – Continued conversation. No more ‘Hey Google’ to invoke assistant every-time in the conversation
    • Multiple actions – 2 or more unrelated operations can be combined as a single sentence
    • Pretty please – Family friendly assistant encouraging kids to use ‘Please’
  7. Visual assistant
    • Illustration on how voice and video can help in cooking
    • Assistant using more of visual elements in it reply
    • Assistant coming to navigation in google maps
    • Google duplex – Assistant getting things done on your behalf and also helping small business get noticed in a competitive market
  8. Digital wellbeing
    • FOMO -> JOMO
      • Understanding the habits
      • Focus on things that only is important
      • Switch off and wind down
      • Finding balance for your family
    • Dashboard – Time spent on phone
    • Family link – Managing kids screen time on phone
  9. Re-imagined Google News
    • Customized news for you – Material theme – News cast provides information from multiple source and provides basic informations
    • News over a time line – Context based full coverage to help you understand the in-depth information
    • Support news publisher – Newsstand – Seamless subscription to magazines
  10. Android P Beta
    • Intelligence – On device ML and data privacy.
      • Adaptive battery using ML to understand what app will be used in near future and not
      • Adaptive brightness uses ML to understand your brightness pattern based on environment and activities
      • App actions – ML predicts the action and provides options to the user based on their usage
      • Slices – Developers can use slices API to show slice of the app based on the context under search
      • ML Kit – API’s from firebase. On device capability. Some examples include – Image labeling, Text recognition, face detection, Barcode scanning, Landmark detection, Smart reply, etc. Built on top of tensorflow lite as inbuilt models. Available on both android and iOS platforms
    • Simplicity
      • Single home button – Navigation changes
      • Changes to volume buttons
      • Screen rotation control – Rotation confirmation
    • Digital wellbeing
      • App timer – Set daily timer for an app and you will witness the app icon goes grey once when the app is used more than the specified limit
      • Do Not Disturb – Silences the visual interruption too. Shush mode just lets your phone on DND mode by just flipping the phone screen facing down
      • Wind down – Set the phone to grey mode after the preset time
  11. Maps
    • Improvements on showing the waiting time – Find your parking. Addition of motor cycle as one more medium of navigation
    • New ‘For you’ and ‘Your Match’ uses ML to provide customized information based on your interest
    • Create shortlist and vote in real-time – Easy to select and choose places among friends
    • Brilliant navigation with camera support for walking navigation
    • VPS – Visual positioning system also helps in navigation
  12.  Google Lens – This will be integrated right into the camera apart from google assistant and photos.
    • Smart text selection – Copy & paste from the real world camera
    • Style match -Search and show available matched options on the go
    • On device intelligence and cloud TPU helps in answering the real time results
  13. Self driving cars
    • Deep learning – Neural networks with multiple layers reduced error rate by 100 times
    • Vision system + Lidar system + Sensors
    • Self driving in severe weather conditions like snow. Applying machine learning reduces too much noise created by snow flakes
Google AI 2018 – Highlights

It will be interesting to know how these modern digital skills can be used to develop solutions that would modernize and sophisticate the way people lead their life.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.