Keynote Speaker Sundar Pichai
Sundar Pichai. the chief executive officer of Google Inc., takes the stage.
Goal: Make Android into a Timeless Machine with New Features
AI offers new ways to solve problems for users around the world.
Use cases for Deep learning
- Healthcare–early diagnosis and deep learning for diabetic retinology; in addition,
eye scan can predict strokes, noninvasively - Help doctors predict imminent health events (24-28 hours before the event occurs) through retianl scans
- Looking to listen–allowing the hearing impaired to disambiguate voices on TV
- Devices that allow typing in Morse Code: putting dots and dashes into text and voice; can even be used when skydiving; uses AI for predictive suggestions for wording
Smart Compose Gmail
Gmail: new feature: Smart Compose uses machine learning to suggest phrases as you type, hit tab for auto-completing, allows you to focus on content
Suggested Actions Smart Phone
Use AI for Photos, one swipe lets you send pictures to the people in the picture
- Fix brightness
- Convert document to PDF
- Change background color
- Change black and white photos to color
Next Generation ChipsTPU 3.0
8x more powerful than last year with water-cooled data centers
Google Assistant
- Better Voice Assistance
- Wavenet action now closer to human speech
- Added 6 new voices to Google Assistant
- More globally accessible
- Coming later this year: John Legend: Lots of different use cases: recipes, songs, rich voice variation, news update
Hey Google! Expanded Capabilities
More natural-understand social dynamics
- Showed the Italian Grandma using Hey Google! from Youtube–what not to do
- Understands requests without Hey Google
- Continued Conversation–available in next few weeks, can respond to multiple action requests, understands AND in context, known as coordination reduction in Linguistics
- Family friendly–children’s stories, Pretty Please, teaches children manners by rewarding kids for saying Please and polite requests, user tested on Google employees’ kids
Smart Displays
Lillian’s talk: Bringing Voice and Visuals–Best of Google to bring voice and visuals together
- Local news, live sports, in addition to usual how-to videos
- Combining videos and text for cooking and other instructions
- Rich immersive response
- Bringing controls for heat
- Partnering with Dunkin’ Donuts and Starbucks other restaurants knows your usual order and usual place
- Visual Snapshot of my day
- Share MTA to friends and family with maps
Google Assistant – Enhance Getting Things Done
Google assistant can call your real salon to book a haircut as a specified date and time range in real-time in the background
- Google Duplex
- Deep learning that can understand nuances
- Make 1 phone call to update millions of users
Search and Digital Wellbeing
- Help you understand your habits
- Dashboard-display how you use your phone
- Example: YouTube will remind you to take a break
- Notifications
Family Link
Helps kids make smart decisions, train for secure, kind online experience
New Google News
Google Supports Quality Journalism
These are challenging time for the journalism industry. Google’s actions:
- Help journalists to innovate programs
- Combining multiple sources with Google News
- Google News and AI to bring best of journalism has to offer
- Provide deeper insights
Android News Briefings
You can toggle News displays:
- 5 top stories in the news
- Personal selection based on your preferences through reinforcement learning
- Headlines around the world
- Let the stories speak for themselves
- Built on Material Themes
Newscasts
- Easy ways to get the basics and decide where you want to dive into deeply
- Able to get many different perspectives and depth of choice
- Full coverage
- Temporal Co-locality: news organized in real time
- AI takes a story and shows the timelines and details based on context, perspective, fact checks, 360-degree view of headlines
- Provides the same content to everyone, full coverage from trusted sources
- Help support sources you love and help discover new sources and follow publications,
- Subscribe through Google: once you’re signed in to a source, no need to log in
Android 10 Years Later
10 years ago, Google launched Android with the simple and bold idea that it was open to everyone. Androids fueled the shift from desktop to mobile>
Goals:
- Open to change and differences
- New Frontiers
- New Friends, Innovations
Android-P Beta – Available Today!
Google wants your input!
Google is building technology to help people every day with every-day activities, combining AI and OS:
- Intelligence
- Extend batteries–adaptive batteries, use AI to use of app patterns
- Auto-brightness that takes into account personal preferences –Adaptive brightness
- UI more intelligent-predictive actions and apps using Actions API
- Google Search-action to appropriate entities
- Slices-new API, interactive snippets, versatile, app actions, partnering with multiple vendors
- Machine Learning
- ML Kit-new APIs, face detection, cross-platform, Android and IOS
- Works with existing apps like Loseit, providing the camera with food analysis
Android Simplifies its UI
- Improved gesture recognition
- 5 predictive apps
- second swipe shows all apps and overview
- quick stoke gesture
- Refers back to info in the previous app
- Smart text selections enable multitasking
- Volume control-located in-app, except for ringer volume
- Rotation under user’s control–ease of use
Making it easier to merge real life and digital life
Sameer Samat works to help users find the right balance between digital and real life; 70% people want more help with striking the balance
Meaningful Engaging Dashboard
- Displays how you use your phone, what and when you’re engaging in apps
- Controls how and when you use apps
- Improve do not disturb–Shush when the phone is face down so you can be more present
- List of important people can always get through
- Wind Down mode-at a self-selected time, Google switches the phone display to grayscale
Google Maps to Assist with Everything
Smarter Maps that can be added through photos, for rapid changing places, isolated places, if your business is open, parking, and wait times.
Updated version keeps you informed of what’s happening in your areas. Better decisions on-the-go.
For you: based on your areas, trending cafes, new places, including:
- Your Match-recommendations just for you based on your preferences, habits
- Share choices for planning with friends
- Small businesses: support small businesses by connecting you with them, daily posts, events, updates, place orders with 1 click
- Mapping EveryWhere combined with AI and cameras: provide context, spatial directions, combined with street view and maps, pointers >>>>> to help you stay oriented, a helpful guide who can show you the way.
- Visual Positioning System-uses visual features that show you where you are and how to get to where you want to go
Google Lens
Lens will be in the camera app, provide information with the camera, providing these features:
- Smart text selection, copy and paste text from the real world into your phone, analyze ingredients on every dish on the menu
- Style Match: Use the camera with AI to match styles in clothing, furniture, shapes, colors, and more
- The lens works in real time- lens surface proactively that what you see – using machine learning, for example, pointing at a camera at a music poster and see the video of the artist.
Self Driving Cars-Waymo
WAYMO in use in Phoenix now!
Waymo has driven 6 million miles on public roads, simulated driving of 5 million miles to model behavior and mitigate accidents.
Soon you can call a self-driving car to take you to public transport
Waymo can do the following for you:
- Helps people who can’t or don’t want to drive to get around, helps kids get to sports practice, and more.
- Improve driving through AI and Machine Learning-enables
- AI improves safety by 100X!
- AI and ML combine sensors with prediction, mapping, and perception.
- Prediction: perceiving and classifying objects, pedestrians, people in manholes or costumes
- Provides sensor data in all forms in real times
- Predicts what other drivers might do-example, cars running a red light, model understands unusual behavior and accommodates for safety
- ML assists differentiating objects from the background, for example, driving in snow and spotting moving and parked cars