Author Archives: Google Devs

Announcing the fourth class of Launchpad Accelerator featuring new countries!

Posted by Roy Glasberg, Global Lead, Launchpad Program & Accelerator

We're back at it again and excited to welcome an inspiring group of startups from all over the world for the 4th class of Launchpad Accelerator.

This time around, startups from Asia and Latin America will be joined by peers from Africa and Europe. In addition to expanding our reach, we’re also expanding our curriculum. We’ll help the startups dig deeper into machine learning and AI, to help them leverage Google’s latest technologies to scale their apps.

Launchpad Accelerator includes intensive mentoring from 20+ Google teams, and expert mentors from top technology companies and VCs in Silicon Valley. Participants receive equity-free support, credits for Google products, and continue to work closely with Google back in their home country during the 6-month program.

Class 4 kicks off July 17th, 2017 at the Google Developers Launchpad Space in San Francisco and will include 2 weeks of all-expense-paid training.

Here's the full list of participating startups (by region):

Africa

Kenya

    Twiga Foods Ltd: A tech-enabled sourcing and distribution platform that replaces informal wholesale markets for the millions of small and medium size food and FMCG retailers in Africa's urban markets.

Nigeria

    Gidi Mobile Limited: gidimo is a mobile learning platform that uses mastery learning & social gamification to deliver personal advancement, in a fun way, and at unprecedented scale
    Flutterwave: Flutterwave is building technology and infrastructure for digital commerce across Africa starting with Rave, an app that helps merchants accept mobile money, cards and bank account payments across 4 African countries.
    Paystack: We help businesses in Africa accept payments from their customers.

South Africa

    JUMO Marketplace: Jumo is the largest scale, lowest cost financial services marketplace for emerging markets

Asia

India

    EdGE Networks: HR solutions provider powered by Artificial Intelligence
    FastFilmz: The Super App for Super Fans of South Indian movies!
    IndiaLends: Credit underwriting and analytics platform for unsecured consumer lending
    RailYatri: Intelligent, big data platform that leverages crowd-sourced content to help long distance travelers
    RecipeBook: Intelligent solutions in food and retail powered by deep learning computer vision
    SigTuple: Smart screening solutions powered by data driven intelligence

Indonesia

    NADIPOS: Restaurant Management Platform helping simplify operations and analytics
    SIRCLO: Empowering businesses to sell online

Malaysia

    HealthMetrics: Assisting companies managing their employee health benefits efficiently and cost effectively

Philippines

    BLOOM: Transforming remittance businesses with blockchain technology
    Honesty Apps: A Do-it-yourself mobile application platform for both iOS and Android

Singapore

    Cicil: We allow Indonesian university students to purchase items online, and pay for them in monthly installments without a credit card!

Thailand

    Piggipo: An application that helps users manage and monitor credit card spending more easily and effectively
    QueQ: Helps users manage time more wisely at crowded places

Vietnam

    eDoctor: Allows users to consult doctors anytime, anywhere

Europe

Czech Republic

    Gamee: A social gaming platform
    Spendee: Helps users understand their finances

Hungary

    PublishDrive: Intelligent SaaS enabled ebook publishing platform to increase sales globally

Poland

    DrOmnibus: Multimedia tools that support special needs education and therapy for children with developmental and behavioral disorders

Latin America

Argentina

    Increase: Simplifies and modernizes how merchants and companies understand, control and manage their income.

Brazil

    Arquivei: Empowers companies with smart data from fiscal documents fetch & analysis
    Contabilizei: Tax reporting and accounting SaaS for small businesses in Brazil
    Contratado.ME: The marketplace that puts candidates at the center of their job search

Chile

    ComparaOnline: The best and the most transparent marketplace for financial products in Latin America

Mexico

    Switch: Low cost, mobile bill pay

Google I/O: New Ways to Put Users at the Center of Your Apps and Payments

Cross posted from Inside AdMob

Posted by Sridhar Ramaswamy, Senior Vice President, Ads and Commerce

I/O is a magical time at Google. Every year, thousands of developers gather in Google's backyard to share new product ideas and learn about our latest innovations in computing.

We're meeting at an exciting time for the developer community. It's a time when consumers have more choices than ever before—like where to shop, what to watch, which games to play and how to communicate with friends and family. Your product needs to stand out. You need tools to help your business grow. And you need to make sure your users are happy.

We think we can help.

This week, my team and I shared 3 new innovations for developers to make it easy for users to pay for your services, build profitable businesses and grow your user base.
Enabling users to pay with Google

Starting today, our suite of payment solutions will be expanding. The Google Payment API enables merchants and developers to turbocharge checkout conversion by offering your users an easy way to pay with credit and debit cards saved to their Google Account. Users will have multiple Google payment options at their fingertips, like a credit or a debit card previously saved via Android Pay, a payment card used to transact on the Play Store or a form of payment stored via Chrome. And they'll be able to use these saved payment options in third-party apps and mobile sites, as well as on Google Assistant when they are on-the-go.

Paying with Google for Panera Bread on Google Assistant

For users, this means faster checkout. Now they'll never miss a deal because they're stuck on a bus and don't want to pull out their credit card in front of strangers. They'll no longer experience the pain of stumbling on a sale that ends at midnight when they're tucked in bed with their credit card out of reach. Users can save time and headache by using credit and debit cards they've already saved to their Google Account whenever they see the option to pay with Google on supported apps or sites.

For developers, this API is a significant innovation that can enable faster checkout, drive more conversions, increase sales and reduce abandoned carts—all with a simple integration. Learn more about our Google Payment API here.
Earn more from your apps with the brand new AdMob
People turn to their mobile devices throughout the day to shop, communicate and stay entertained. For developers, in-app purchases are one way to monetize. Ads are another way.
AdMob was built to support the app ecosystem. With over 1 million apps across iOS and Android, AdMob has paid over $3.5 billion dollars in ads revenue to developers. But there's more we can do to help you succeed.

Today, we're introducing a completely redesigned AdMob. Rebuilt from the ground up, AdMob is now simpler to use and delivers richer insights about your users' in-app experiences.

Simpler to use: We've applied Material Design to all aspects of the AdMob look and feel to deliver an easy-to-use and intuitive experience across the entire platform—on mobile and desktop. You'll get more done in less time. Below you can see how easy it is to pick an app that you're monitoring, check out its key metrics and then quickly take action to fine-tune its performance.


Redesigned AdMob experience 
Deeper insights: We've also integrated Google Analytics for Firebase into the core of the redesigned AdMob so you have quick access to the metrics that matter most for your business. Once you link your AdMob and Firebase accounts, you'll have access to detailed ad revenue data and user insights like time spent in the app and in-app purchases—all in one place.

Google Analytics for Firebase dashboard in AdMob

Know your user, find your user with Universal App Campaigns
Earning money from your app is one piece of the puzzle. You also need to think about how to grow your user base.

Google's app innovations have delivered over 5 billion installs from ads and we are now helping developers drive over 3 billion in-app events per quarter—like users adding something to their cart or reaching level 3 of a game. Developers have gravitated toward Universal App Campaigns (UAC) as the "one stop shop" campaign type that scales your reach and maximizes app installs across Google's largest properties: Google Play, Search, YouTube, Gmail and the Display Network. UAC uses Google's machine learning technology to evaluate numerous signals in real time, refining each ad to help you reach your most engaged users. We're continuing to double down on UAC, with all new innovations being built into UAC to make app promotion even more effective.
Engage users in key moments of discovery with new UAC placements in Google Play 
Android reaches more than 2 billion active devices every month, with Google Play available in 190+ countries around the world. It's the place users come to discover new apps and games. Beyond searching for apps to try, users are increasingly browsing the Play store and finding recommendations for new apps. 
To help those users discover more of your apps, we are introducing new ad placements on the home and app listing pages in the Google Play Store. These new placements, available exclusively through UAC, help you reach users in "discovery mode" as they swipe, tap and scroll in search of their next favorite app. 

New ad placements reach users browsing in Google Play 
Discover more of your best users with new bidding options in UAC 
Some users are more valuable to your business than others, like the players who level-up in your game or the loyal travelers who book several flights a month. That's why we're expanding Smart Bidding strategies in UAC to help you acquire more of these high-value users. Using Smart Bidding, you can tailor bids for your unique business goals - target cost per acquisition (tCPA) or target return on ad spend (tROAS). UAC delivers the right users based on your objectives: installs, events and, coming soon, value. This update starts rolling out to iOS and Android developers and advertisers in the coming months. 
Introducing App Attribution Partners, a new measurement program 
Many developers rely on third-party measurement providers to measure the impact of ads and gain valuable insights about how users engage with your app. To help you take action on these insights in a faster and more seamless way, we are introducing App Attribution Partners, a new program designed to integrate data from 7 global companies right into AdWords.

Welcome to adjust, Adways, AppsFlyer, Apsalar, CyberZ, Kochava and TUNE... we're thrilled to have them onboard!

AdWords' integration with these partners ensures that you have consistent, reliable and more granular data where you review app metrics. Now you can take action with confidence and stay on top of your business performance.
As consumers live more of their lives online, it's increasingly important for developers to build user-centric experiences in everything that you do—from the apps you design, to the experiences you deliver, to the ways you help people transact. We know it's not always easy, so Google is here to help.

We look forward to continuing on this journey with you.

The Modern Mobile Web: State of the Union

Cross posted from the Chromium Blog

Posted by Rahul Roy-chowdhury, VP Product Management, Chrome
What a difference a year makes. Last year at Google I/O, we shared that the mobile web was open for business. New technologies such as AMP and Progressive Web Apps (PWAs) were bringing new capabilities, better performance, and a streamlined workflow to the mobile web.
Fast forward one year later: more than two billion AMP pages have been created and "PWA" has proved to be far more than a buzzword—it’s now the way that many businesses around the world are building for mobile devices. For more details, take a look at the video from Google I/O on the latest mobile web state of the union, or read below on how these technologies are making the modern mobile web mainstream.
Momentum
Summing up all the great success stories from around the world in a single post is a tall order, but here are some highlights.

To improve the performance of Wego's mobile site, the company built AMP pages using amp-install-serviceworker to transition to a fast PWA experience. Average page load time decreased from 12 seconds to less than one second, and conversion rates increased by 95%.


When Forbes rebuilt their mobile website as a PWA, they began by re-thinking what their experience could look like on a phone. Instead of minimally updating their underlying site, Forbes integrated PWA technologies to provide an immersive, app-like experience. They saw immediate improvements and engagement rates have more than doubled since launch.


Ola, the leading cab aggregator in India, built a PWA and noticed that 20% of users who book using their PWA had previously uninstalled their app. By reducing the amount of storage space needed, the PWA allowed them to effectively re-engage with users that otherwise would have been lost.
Another success story is Twitter Lite, a PWA which minimizes data usage, is resilient on unreliable mobile networks, and is less than 1MB of space on a device. Twitter's new mobile experience is also optimized for speed, with up to 30% faster launch times as well as quicker navigation throughout the site. They've found that users are spending 2.7x more time on site, and as a result are seeing 76% more tweets on the new PWA than their previous mobile site. Twitter is seeing incredible re-engagement with 1 million sessions initiated a day from icons added to the Android homescreen.
Polished Experiences
Users expect a lot from their mobile devices, and we've added tons of APIs over the past year to meet that demand. The mobile web can support more use cases and get more done than ever before. A few highlights:
  • Improved Add to Homescreen: Earlier this year we unveiled Improved Add to Homescreen, integrating PWAs much deeper into the Android operating system. Now, in addition to being displayed on the homescreen, PWAs are also displayed in the app launcher and Android settings alongside native apps, and can also open in response to users clicking links in Chrome or other apps.
  • Payments: Checkout can be a complicated process. To improve payment flows on the web, we launched a one-tap payment API called Payment Request. Using this API allows web apps to support credit cards and Google payment mechanisms such as Android Pay. We also just announced that it is now possible to integrate this API with additional payment apps.
  • Media Consumption: Over 70% of internet traffic is video. To allow great mobile web media experiences we have given the users more control over playback with the Media Session API, improved full screen playback with the Screen Orientation API, and we’re filling out features for offline with Background Fetch. To learn more, see our mobile web media best practices and see how the APIs can come together at our PWA for Media demo.
Tooling
We’ve also been working hard to improve and extend the set of tools that let you build engaging experiences on the web.
Lighthouse is a new automated tool for measuring the quality of a web experience. It runs nearly 100 audits against your web app, checking everything from page performance, to byte efficiency, to accessibility, and gives you a summary score. New integration with Chrome's DevTools means you’ll be able to run Lighthouse audits without leaving the browser.
Polymer 2.0 is the next major release of the Polymer library, re-built from the ground up to take advantage of the best new features of the modern web platform. This release uses new Web Component API’s that have shipped in Chrome and Safari. It’s completely modular and best of all - it’s now 10% faster and 80% smaller.
Chrome is committed to making sure that you can develop easily, engage with your users, and build a thriving business around the web. For the latest news, subscribe to our YouTube channel and follow us on Twitter @ChromiumDev.

Google I/O 2017: Empowering developers to build the best experiences across platforms

By Jason Titus, Vice President, Developer Product Group

It's great to be in our backyard again for Google I/O to connect with developers around the world. The 7,200 attendees at Shoreline Amphitheatre, millions of viewers on the livestream, and thousand of developers at local I/O Extended events across 80+ countries heard about our efforts to make the lives of developers easier -- allowing them to focus on the problems they're trying to solve by minimizing the pain points of building a product.

Earlier this morning, our CEO Sundar Pichai talkedabout our various billion-user platforms. Whether it's Android or Chrome or the mobile Web, our success would not have been possible without the developer community. And during our Developer Keynote, we covered our heavy investments in tools and services for developers who build on our platforms every day.

We have a lot to cover over the next three days. Let's take a closer look at the major developer news at I/O so far:

Platforms that connect developers to billions of users around the world

  • Android O Developer Preview 2 — Get a look at the next release of Android O focused on fluid experiences that make Android even more useful, and our efforts to optimize battery life, startup time, graphic rendering time, and stability. Early adopters can opt in to the Android O Beta Program at android.com/beta and run Android O now.
  • Project Treble — Last week, we also introduced a new Android framework designed to help reduce the time and effort it takes device makers to upgrade a phone to a new version of Android, starting with Android O.
  • Android Go — We're optimizing Android to run smoothly on entry-level devices, starting with the O release. We're also designing Google apps to use less memory, storage space, and mobile data, including apps such as YouTube Go, Chrome, and Gboard.
  • Kotlin— Android is officially supporting the Kotlin programming language, in addition to the Java language and C++. Kotlin is a brilliantly designed, mature, production-ready language that we believe will make Android development faster and more fun.
  • Android Studio 3.0 Canary — Our new preview includes three major features to accelerate development flow: a new suite of app performance profiling tools to quickly diagnose performance issues, support for the Kotlin programming language, and increased Gradle build speeds for large sized app projects.
  • Mobile Web — AMP and Progressive Web Apps (PWAs) are re-defining modern mobile web development. AMP gets content in front of users fast and PWAs deliver app-focused experiences that are reliable, fast and engaging. We're seeing success storiesfrom all around the world - travel company Wego has rolled out a successful AMP based PWA and Forbes has seen user engagement double since launching a PWA. If you're wondering how good your current web experience is, you can use Lighthouse - an automated tool for measuring web-page quality. Be sure to tune in this afternoon for theMobile Web: State of the Union talk to hear more about building rich mobile web experiences.

Infrastructure and services to take mobile apps and the Web to the next level

  • Firebase— At last year's I/O, we expanded Firebase to a full mobile development platform with products to help you build your app and grow your business. Over a million developers now use Firebase, and we're doubling down on our efforts to simplify more every-day developer challenges. We're giving more insights to understand app performance through Firebase Performance Monitoring, introducing integration between Hosting and Cloud Functions, adding support for Phone Number Authentication, and continuing to improve Analytics in a number of ways. We've also started open sourcing our SDKs.
  • Mobile web developer certifications — At I/O'16 we launched the Associate Android Developer Certification. This year, we're adding two new certifications for web developers: the Mobile Sites Certification and the Mobile Web Specialist Certification.

Powerful tools to acquire and engage new users; grow successful businesses

  • Google Play Console — We announced several powerful, new features and reports in the Play Console to help developers improve their app's performance, manage releases with confidence, reach a global audience, and grow their business. The Play Console also has a new name, to reflect its broadened business uses, and a fresh look to make it easier to get things done.
  • Android Instant Apps — We opened Android Instant Apps, a new way to run Android apps without requiring installation, to all developers. Now anyone can build and publish an instant app. There are also more than 50 new experiences available for users to try out from a variety of brands, such as Jet, New York Times, Vimeo and Zillow.
  • Payments, Monetization & Ads — We introduced a Google Payment API that enables developers to give their customers the ability to pay in apps and online with credit or debit cards saved to their Google Account. New AdMob integration with Google Analytics for Firebase helps them monetize efficiently and updates to Universal Apps Campaigns will help them grow their user base.

New interfaces to push the limits of what's possible

  • Actions on Google for the Google Assistant — We brought Actions on Google to phones, introduced new features and functionality, improved our SDK and more. We also launched the Actions Console, a new developer console that helps developers work as a team, and collect data on app usage, performance and user discovery patterns. This new console is integrated with the Firebase and Google Cloud consoles.
  • VR and AR at Google — We'll have more to share on the latest Daydream platform features and developer tools during our "VR and AR at Google" session tomorrow (May 18) at 9:30 AM PT in the Amphitheatre and on the livestream.

It's important to us that developers are successful. In addition to building products that help solve developer challenges, we're on the ground in over 130 countries, growing and expanding the developer community through programs such as Women Techmakers & Google Developer Groups (GDGs). We're also investing in training programs like Google Developers Certification and courses through Udacity and other partners to help developers deepen their technical capability. We're also excited to announce two large multi-product developer events, Google Developer Days, which are planned for Europe (September 2017 in Krakow, Poland) and India (December 2017 in Bangalore, India). If you are interested to find out more, sign up for updates on g.co/gdd2017.

During Google I/O, attendees and viewers have an opportunity to dive deep into a number of these areas with 14 content tracks and 140+ breakout sessions -- covering Android to Assistant to VR -- and all livestreamed. We've also launched over 70 codelabs to get developers up and running with our latest APIs today.

Whether it's Android, Chrome, Play, VR/AR, the Cloud, and the Mobile Web — we're constantly investing in the platforms that connect developers to billions of users around the world. Thank you to the continued support and feedback from the developer community.

From Actions on Google to the SDK, the Google Assistant is getting better for developers

Posted by Brad Abrams, Product Manager for the Google Assistant


Five months ago, we announcedan early version of the Actions on Google developer platform for Google Home. Since then we've been focused on growing your audience, expanding the platform's features and improving the developer experience overall. With today's news, we're excited to bring the platform to phones, introduce new features and functionality, improve our SDK and continue to work together to create great apps for Google Assistant users near and far!

Introducing Actions on Google on phones


Since the platform launched in December, it's been exciting to see creative and interesting apps come to life on Google Home, from working with FitStarto getting the latest news with CNBC. Today, we're bringing Actions on Google to the Assistant on both Android phones and iPhones.


With apps for the Assistant now available on phones, you'll have the opportunity to grow your user base and build apps for entirely new use cases -- things that weren't suited to a voice-only interface, like shopping for clothes or ordering food from a lengthy menu. Moving to a screen also means users can easily get things done with their app through new UI elements like image carousels, lists, and suggestion chips.


You can build and deploy apps for the Assistant on the phones starting today - here's the documentation.


We'll also soon be launching Actions on Google in English in the UK and later this year French and German, and other languages.


Adding Transactions and Payments

Our goal with the Assistant is to help you get things done - and that doesn’t just mean asking questions or listening to information. We also want to make it easy to complete purchases.

To enable payments in your apps for the Google Assistant, you have two options. You can use Google-facilitated payments that are free, easy to integrate, and allow you to leverage the hundreds of millions of cards that are already available with Google. Or, you can use the payment method the user has already provided you. With this second option, we recommend using our new seamless account linking solution that enables users to sign into an existing account or create a new account with just two taps.


But a transaction isn't over when the user pays, it includes things like tracking an order, modifying or reordering. That's why the Assistant now allows users to see all their transactions in a single history view. And we also built an order-updates feature to make it easier to re-engage. With it you can send status updates like when a car arrives to pick them up, when their food is delivered, or their prescription is ready.


You can start building and testing transactional apps today, and they'll be available to Google Assistant users on phones soon.

Offering Better Tools and Discovery

With all of these new features, getting the basics is more important than ever and we know that great tools and being discovered are top of mind.


We're invested in offering a great developer experience, so today we're also launching a new developer console. This console helps you work as a team, and collect data on your app's usage, performance and user discovery patterns. It's integrated with the Firebase and Google Cloud consoles, so that you share data within your apps.


Additionally, we're rolling out a new app directory. Users can access it with a single tap from the Google Assistant and it has both categories and user ratings. Each app's directory page is also shareable on the web so that you can promote your app for new and existing users and they can share it with their friends.


With this update, they can also create a shortcut to your app, so instead of saying "Ok Google, ask Forecaster Joe what's the surf report for the Outer Banks", a user can just say their personal shortcut, like "Ok Google, is the surf up?" to easily re-engage with your app.


While we're confident these features will help improve discoverability, our work is not done! We'll continue to add new features and improve the discoverability of your apps over time.

Updating the Assistant SDK

Last month we introduceda preview of the Google Assistant SDK. Today, we're continuing to make it better by adding a number of new features.


With hotword support, developers can now build devices that are triggered by "Ok Google" rather than a button or some other physical action. We're also adding the ability to have both timers and alarms. So a user can now say "Ok Google, set a timer for 60 seconds" on any device with the Google Assistant built-in.


While we are still in the early days of the SDK and the platform, we're continuing to work on creating a more comprehensive developer experience. We're also looking to bring the platform to new devices, including those powered by the Google Assistant SDK.



Announcing a new developer competition

Last but not least, we're also launching the first developer competition for Actions on Google. As part of this competition, there will be more than 20 prizes for the best apps for the Google Assistant! So start building - we can't wait to see what you come up with.


We're excited for the road ahead and look forward to working with you to develop new apps for the Google Assistant.

Introducing the TensorFlow Research Cloud

Posted by Zak Stone, Product Manager for TensorFlow
Researchers require enormous computational resources to train the machine learning (ML) models that have delivered recent breakthroughs in medical imaging, neural machine translation, game playing, and many other domains. We believe that significantly larger amounts of computation will make it possible for researchers to invent new types of ML models that will be even more accurate and useful.
To accelerate the pace of open machine-learning research, we are introducing the TensorFlow Research Cloud (TFRC), a cluster of 1,000 Cloud TPUs that will be made available free of charge to support a broad range of computationally-intensive research projects that might not be possible otherwise.
The TensorFlow Research Cloud offers researchers the following benefits:
  • Access to Google's all-new Cloud TPUs that accelerate both training and inference
  • Up to 180 teraflops of floating-point performance per Cloud TPU
  • 64 GB of ultra-high-bandwidth memory per Cloud TPU
  • Familiar TensorFlow programming interfaces
You can sign up here to request to be notified when the TensorFlow Research Cloud application process opens, and you can optionally share more information about your computational needs. We plan to evaluate applications on a rolling basis in search of the most creative and ambitious proposals.
The TensorFlow Research Cloud program is not limited to academia — we recognize that people with a wide range of affiliations, roles, and expertise are making major machine learning research contributions, and we especially encourage those with non-traditional backgrounds to apply. Access will be granted to selected individuals for limited amounts of compute time, and researchers are welcome to apply multiple times with multiple projects.
Since the main goal of the TensorFlow Research Cloud is to benefit the open machine learning research community as a whole, successful applicants will be expected to do the following:
  • Share their TFRC-supported research with the world through peer-reviewed publications, open-source code, blog posts, or other open media
  • Share concrete, constructive feedback with Google to help us improve the TFRC program and the underlying Cloud TPU platform over time
  • Imagine a future in which ML acceleration is abundant and develop new kinds of machine learning models in anticipation of that future
For businesses interested in using Cloud TPUs for proprietary research and development, we will offer a parallel Cloud TPU Alpha program. You can sign up here to learn more about this program. We recommend participating in the Cloud TPU Alpha program if you are interested in any of the following:
  • Accelerating training of proprietary ML models; models that take weeks to train on other hardware can be trained in days or even hours on Cloud TPUs
  • Accelerating batch processing of industrial-scale datasets: images, videos, audio, unstructured text, structured data, etc.
  • Processing live requests in production using larger and more complex ML models than ever before
We hope the TensorFlow Research Cloud will allow as many researchers as possible to explore the frontier of machine learning research and extend it with new discoveries! We encourage you to sign up today to be among the first to know as more information becomes available.

Get ready for Google I/O 2017

Posted by Mónica Bagagem, Product Marketing Manager
We’re excited to be hosting Google I/O 2017 next week at the Shoreline Amphitheatre! The agenda for May 17-19 is packed with rich, technical content. Here are some tips to help you make the most of it.

Attending in person?

Everyone is guaranteed a spot for the keynotes but seating will be pre-assigned on a first-come, first-served basis during badge pick-up. Your seating section will be noted on your badge. Badge pick-up starts on Tuesday, May 16th, between 7AM - 7PM PDT at the Shoreline Amphitheatre. Plan to come by early to get the best seats!
Sessions start at 2PM after the Developer Keynote ends, and are roughly 40mins in length. To help make it easier for you to attend your favorite talks and minimize lines, you can reserve seats for sessions now via our web app, Android app and iOS app using your Google I/O registration email address. Additionally, App reviews and select Sandbox demos will be reservable onsite on a first-come, first-served basis at the beginning of each day.
Beyond attending technical Sessions, you’ll have the opportunity to check out our latest product demos and speak directly with Google engineers throughout the Sandbox space; during Codelabs where you can complete self-paced tutorials; and at Office Hours where you can get specific questions answered 1:1 with Googlers.
Remember to save some energy for the evening! On Day 1, we’ll host an After Hours Block Party from 7-10PM. It will include dinner, drinks, and lots of fun, interactive experiences throughout the Sandbox space: our very own comedy club, an international food market & pizza party, several musical performances, a VR drive-in, a Museum of Developer Art, to name just a few! On Day 2, we’ll have an After Hours Concert from 8-10PM (don’t worry, we’ll feed you dinner, too!). Stay tuned - we’ll be announcing the talent closer to I/O.
Don’t forget to to check the Mountain View weather forecast for each day; we recommend bringing a jacket for the evening festivities as it can get chilly after dark. Although all Sessions and Sandboxes will take place in climate controlled structures, Shoreline Amphitheatre is an outdoor venue - so come prepared for whatever mother nature might have in store!
Finally, you can find directions, shuttle schedules, biking, parking, and carpooling info here.

Attending remotely?

Even if you’re not at Shoreline, you can still participate in I/O from afar! Here’s how:
  • I/O Extended: Find an I/O Extended event near you to watch the keynotes with your community, participate in hackathons, codelabs, and much more.
  • Livestream: Tune into the livestream throughout the 3 day festival on desktop and mobile.
  • I/O Live Widget: If you want to bring the livestream and the #io17 social conversation to your audience, you can customize and embed our I/O Live widget on your site or blog.
  • I/O Guide: Follow our Guide, Timothy Jordan, as he tours the venue and gets the inside scoop. You can find him on any of our livestream channels throughout the event, in-between sessions.
  • #io17request: Between May 17-19, send us your questions about I/O via English-language tweets that include the #io17request hashtag. A team of Googlers across Android, Chrome, Assistant, VR, Machine Learning, and more will track down answers to your burning questions.
  • I/O in photos: Be sure to follow out our real-time I/O photo album from Shoreline!

Check out our FAQ page if you need more info and join the conversation at #io17. See you veryyyyy soon!

Build beautiful apps and websites with modular, customizable UI components

Posted by Adrian Secord and Omer Ziv, Material Design

Material Components lets you build easily for Android, iOS, and the web using open-source code for Material Design, a shared set of principles uniting style, brand, interaction, and motion.

These components are regularly updated by a team of engineers and designers to follow the latest Material Design guidelines, ensuring well-crafted implementations that meet development standards such as internationalization and accessibility support.

Accurate

Pixel-perfect components for Android, iOS, and the web

Current

Maintained by Google engineers and designers, using the latest APIs and features.

Open-source

The code on GitHub is available for you to contribute or simply use elements as needed

Industry standards

Also used in Google's products, these components meet industry standards, such as internationalization and accessibility

Material Components are maintained by a core team of Android, iOS, and web engineers and UX designers at Google. We strive to support the best of each platform by:

  • Supporting older Android versions with graceful degradation
  • Developing iOS apps that use industry standards like Swift, Objective-C, and storyboards
  • Integrating seamlessly with popular web frameworks and libraries

With these components, your team can easily develop rich user experiences using Material Design. We'll be continually updating the components to match the latest Material Design guidelines, and we're looking forward to you and your team contributing to the project. To get the latest news and chat with us directly, please check out our GitHub repos, follow us on Twitter (@materialdesign), and visit us at https://material.io/components/.

Updating developer identity guidelines and registration processes to protect users

Posted by Naveen Agarwal, Identity Team

Last week, we took immediate action to protect users from a phishing attack that attempted to abuse the OAuth authorization infrastructure.

Today, we’re supplementing those efforts to help prevent these types of issues in the future. These changes may add some friction and require more time before you are able to publish your web application, so we recommend that you plan your work accordingly.

Updating app identity guidelines

As our Google API user data policy states, apps must not mislead users. For example, app names should be unique to your application and should not copy others'.

To further enforce this policy, we are updating our app publishing process, our risk assessment systems, and our user-facing consent page in order to better detect spoofed or misleading application identities. You may see an error message as you’re registering new applications or modifying existing application attributes in the Google API Console, Firebase Console, or Apps Script editor as a result of this change.

New review processes and restrictions on web apps requesting user data

We have also enhanced our risk assessment for new web applications that request user data.

Based on this risk assessment, some web applications will require a manual review. Until the review is complete, users will not be able to approve the data permissions, and we will display an error message instead of the permissions consent page. You can request a review during the testing phase in order to open the app to the public. We will try to process those reviews in 3-7 business days. In the future, we will enable review requests during the registration phase as well.

You can continue to use your app for testing purposes before it is approved by logging in with an account registered as an owner/editor of that project in the Google API Console. This will enable you to add additional testers, as well as initiate the review process.

We also recommend developers review our earlier post outlining their responsibilities when requesting access to user data from their applications. Our teams will continue our constant efforts to support a powerful, useful developer ecosystem that keeps users and their data safe.

TensorFlow Benchmarks and a New High-Performance Guide

Posted by Josh Gordon on behalf of the TensorFlow team

We recently published a collection of performance benchmarks that highlight TensorFlow's speed and scalability when training image classification models, like InceptionV3 and ResNet, on a variety of hardware and configurations.

To help you build highly scalable models, we've also added a new High-Performance Models guide to the performance site on tensorflow.org. Together with the guide, we hope these benchmarks and associated scripts will serve as a reference point as you tune your code, and help you get the most performance from your new and existing hardware.

When running benchmarks, we tested using both real and synthetic data. We feel this is important to show, as it exercises both the compute and input pipelines, and is more representative of real-world performance numbers than testing with synthetic data alone. For transparency, we've also shared our scripts and methodology.

Collected below are highlights of TensorFlow's performance when training with an NVIDIA® DGX-1™, as well as with 64 NVIDIA® Tesla® K80 GPUs running in a distributed configuration. In-depth results, including details like batch-size and configurations used for the various platforms we tested, are available on the benchmarks site.

Training with NVIDIA® DGX-1™ (8 NVIDIA® Tesla® P100s)

Our benchmarks show that TensorFlow has nearly linear scaling on an NVIDIA® DGX-1™for training image classification models with synthetic data. With 8 NVIDIA Tesla P100s, we report a speedup of 7.99x (99% efficiency) for InceptionV3 and 7.91x (98% efficiency) for ResNet-50, compared to using a single GPU.
The following are results comparing training with synthetic and real data. The benchmark results show a small difference between training data placed statically on the GPU (synthetic) and executing the full input pipeline with data from ImageNet. One strength of TensorFlow is the ability of its input pipeline to saturate state-of-the-art compute units with large inputs.




Training with NVIDIA® Tesla® K80 (Single server, 8 GPUs)

With 8 NVIDIA® Tesla® K80s in a single-server configuration, TensorFlow has a 7.4x speedup on Inception v3 (93% efficiency) and a 7.4x speedup on ResNet-50, compared to a single GPU. For this benchmark, we used Google Compute Engine instances.

Distributed Training with NVIDIA® Tesla® K80 (up to 64 GPUs)

With 64 Tesla K80s running on Amazon EC2 instancesin a distributed configuration, TensorFlow has a 59x speedup (92% efficiency) for InceptionV3 and a 52x speedup (82% efficiency) for ResNet-50 using synthetic data.

Discussion

During our testing of the DGX-1 and other platforms, we explored a variety of configurations using NCCL, a collective communications library, part of the NVIDIA Deep Learning SDK. Our hypothesis before testing began was that replicating the variables across GPUs and syncing them with NCCL would be the optimal approach. The results were not always as expected. Optimal configurations varied depending not only on the GPU, but also on the platform and model tested. On the DGX-1, for example, VGG16 and AlexNet performed best when replicating the variables on each of the GPUs and updating them using NCCL, while InceptionV3 and ResNet performed best when placing the shared variable on the CPU. These intricacies highlight the need for comprehensive benchmarking. Models have to be tuned for each platform, and a one size fits all approach is likely to result in suboptimal performance in many cases.

To get peak performance, it is necessary to benchmark with a mix of settings to determine which ones are likely to perform best on each platform. The script that accompanies the article on creating High-Performance Models was created not only to illustrate how to achieve the highest performance, but also as a tool to benchmark a platform with a variety of settings. The benchmarks page lists the configurations that we found which provided optimal performance for the platforms tested.

As many people have pointed out in response to various benchmarks that have been performed on other platforms, increases to samples per second does not necessarily correlate to faster convergence, and as batch sizes increase it can be more difficult to converge to the highest accuracy levels.

As a team, we hope to do future tests that focus on time to convergence to high levels of accuracy. We hope these numbers and the guide will prove useful to you as you tune your code for performance.

We'd like to thank NVIDIA for sharing a DGX-1 for benchmark testing and for their technical assistance. We're looking forward to NVIDIA's upcoming Voltaarchitecture, and to working closely with them to optimize TensorFlow's performance there, and to expand support for FP16.

Thanks for reading, and as always, we look forward to working with you on forums like GitHub issues, Stack Overflow, the [email protected]list, and @TensorFlow.