Google AI is opening centers around the globe. Last year they began studying retinopathy in India. The AI systems are finding additional information from eye scans, such as cardiovascular risks.
They are also using AI to predict medical events. 76% success in predicting short term regression
Youtube uses AI to detect who is talking when conversation s overlap. They showed an example of a sports talk show where hosts talked over each other.
Gboards are alternate keyboards. They added Morse code to help a user with CP that communicates with a hand built Morse system.
They are developers and worked to increase word prediction.
Other uses of ML and AI
- Smart compose in Gmail predicts the conversation, not just the word
- Google photos can auto color old black and white images
- Google assistant is a key focus for AI. The goal is to have natural conversation.
- Wavenet is a new audio voice platform that better replicates natural voices.
- Continued conversation allows you to start a session with Google assistant and you do not have to repeat hey Google with each utterance.
- Family assistant now has pretty please,which encourages polite behavior
- Visually assistive incorporates visual displays and go on sale in July.
Google duplex allows Google assistant to make a live call to a business and create an appointment. It uses natural language and feels seamless for the business.
- Digital well being
AI is being incorporated at base level
- Adaptive battery uses on device machine learning to adapt app usage to what it expects is needed. This has a 30% increase in battery life.
- Auto brightness is now using on device machine learning to adapt screen brightness to users preference.
App actions are predicting what actions a person will do, such as making a call or start a run.
Developers add an actions.xml file to let Android p know what is available. This would be good for people that regularly review accounts in the evening.
ML Kit taps into on device machine learning andtensor flow. It will also work on iOS
TurboTax was listed as a featured partner
Only a single home button instead of three buttons within Android P design. I noticed presentations throughout Google IO still showed three button designs.
Walking directions incorporated cam we ran view to easily know directions and what is nearby.
Visual position system uses features in the environment to locate your position. This works when GPS is not available.
Google lens is being added directly to camera app on featured phones
Smart text selection will detect signs and more. Now words can be connected, such as copy and paste from the view of a document. Quickly get description of menu items
It’s not what is this, but what is similar. Like finding a similar outfit
Real time results
ML on device and cloud cpus are allowing the real time results within Google lens
These will arrive in the next few weeks
Waymo is the only company with a fleet of fully self driving cars.
Detecting objects is difficult due to various situations, such as a construction worker exiting a sewer
People do not always behave as predicted
The use of many sensors helps them predict nonstandard behavior