Monthly Archives: May 2019

Content API for Shopping live webinar on May 31st (Mandarin, Chinese)

We are running a live webinar this Friday, May 31, on the Content API for Shopping. This webinar is for our Chinese-speaking developers that are interested in using the API to manage products and Shopping ads at scale.

The webinar will cover topics on:
  • Content API for Shopping use case
  • Merchant Center setup
  • Content API features
  • Best practices
  • API request/response demo
You will be able to submit questions throughout the live event. We will answer questions during the Q&A session in the webinar.

Note: The webinar content will be delivered in Mandarin, Chinese.

The webinar will begin at 10:00am (China Time, GMT +8) on Friday, May 31, 2019. Please contact us via the Content API forum if you have any questions or need help.

If you’d like to review some of our previous webinars covering the Google Ads API, you can find them at the following links:

Google Maps adds dish-covery to the menu

When you’re trying out a new restaurant, Google Maps already shows you how to get there, photos, reviews and ratings. But as you scan the menu, you’re probably wondering, “What’s good?” Now Google Maps is making it easy to “dish-cover” a restaurant’s most crowd-pleasing meals with a new popular dishes feature.

Popular dishes from down the street and around the planet

Whether you’re craving a tea cocktail in Brooklyn or sampling Pacific Rim cuisine in London, Google Maps now highlights a restaurant’s most popular items on the menu, so you can place your order with confidence.

The popular dishes feature is powered by a machine learning algorithm that matches dish names, provided by Google Maps users, with relevant photos and reviews. This creates a handy arrangement of a restaurant’s most popular meals right at your fingertips in Google Maps.

Simply pull up a restaurant on Google Maps to find its popular dishes in the overview tab. Feeling extra peckish? Dive into the menu tab to scroll through all the most-talked about meals, and tap on a popular dish to explore reviews and photos. In a country where you can’t read the language? Maps will also translate the reviews for you too.  

Animation showing popular dishes feature on Google Maps

The new popular dishes feature makes it easier to discover a restaurant’s most talked-about meals.

At the end of the day, this feature is  made possible because of contributions from people around the world who want to help others using Google Maps. So if you want to pay it forward to the next diner, simply take a photo of your meal (before you’ve scarfed it down!) and add a dish name so others can know what’s good on the menu.

From getting you directions to helping you find a popular dish, Google Maps is there for your dining journey. You’ve heard of “farm-to-table”—now we bring you from “door-to-dish.”

It’s all part of Google Maps becoming even more helpful as you travel, dine and get things done. From a redesigned Explore tab in Google Maps that helps you find things to do, group planning features that make it easier to coordinate with friends and family, and now popular dishes, Google Maps is helping you get out and enjoy the world.

Starting today, popular dishes is live on Google Maps for Android devices worldwide and will be rolled out to iOS users in the coming months.

Announcing Webmaster Conference — India!

In the next three years, half a billion Indian language users will be on the web. This unprecedented growth presents a huge opportunity for local language content creators to expand their reach and to improve the local language web ecosystem.

Last year, we went to 11 cities with our Google Search Conference and engaged with over 700 content creators who joined us from across the length and breadth of the country. Looking at the overwhelmingly positive response from the publishers, we are now excited to bring back this conference, fresh and rebranded to Webmaster Conference.

This year, Webmaster Conference will be held in 15 cities and in five Indian languages -- Bengali, Hindi, Marathi, Tamil and Telugu, besides English.

Webmaster Conference series schedule:

Primary Language
New Delhi
June 17
June 19
June 21
June 24
June 26
July 15
July 17
(for women only)
July 19
July 22
July 29
July 31
Bangalore, for women only
August 2
August 5
August 7
August 9

Agenda and Topics
The conference aims to help you better understand industry best practices for building great websites, creating high-quality content that is safe for users and making content easily accessible and visible on Google Search to the growing number of local language Internet users.

It includes talks on what’s latest on Google Search, the new Google Search Console and various other tools, as well as Google’s underlying principles and policies for Search as well as AdSense. The topics will be selected based on community needs and requests.

We believe it’ll also serve as a great platform for the participants to connect with Googlers from the Search and AdSense teams as well as industry experts and influencers engaged in building the local language web ecosystem.

Who can participate
The conference is open to everyone who creates content on the web. If you create content online, run or own a website, are a publisher or a webmaster, manage SEO or are aspiring to start your online presence in one of the India local languages, then you shouldn't miss this conference!

Register your interest!
If you are interested in participating, please register here!

Registration for each city event closes 10 days before the respective event date.

Once you are shortlisted, you will receive an invitation email from us confirming the venue details. Please note that this is an invite-only event and filling out the form does not guarantee your spot at the conference.

We hope to see you at one of the events! If you have any questions, please go through the FAQs here.

Posted by Syed Malik Mairaj, Trust & Safety Outreach Lead South Asia

Beta Channel Update for Desktop

The beta channel has been updated to 75.0.3770.66 for Windows, Mac, and, Linux.

A full list of changes in this build is available in the log. Interested in switching release channels?  Find out how here. If you find a new issue, please let us know by filing a bug. The community help forum is also a great place to reach out for help or learn about common issues.

Srinivas Sista
Google Chrome

Investing in Australia’s Connectivity and Digital Economy

Over the last three years, Google has invested US$47 billion (approx. A$67.8 billion) in CAPEX to improve our infrastructure globally. Today we’re happy to announce that the INDIGO cable system, which connects Sydney, Perth and Singapore, is ready for service!

Developed in partnership with AARnet, Indosat, Singtel, SubPartners and Telstra, INDIGO will strengthen connectivity between Australia and Southeast Asia, boosting business capabilities within the region. It will enable better internet connectivity, meaning faster and more reliable services for all of our users.

When combined with our Cloud Regions in Sydney and Singapore, INDIGO will play a major role in supporting Google Cloud customers in Asia Pacific. We launched our Sydney Cloud Region in 2017, at the time our fourth region in Asia Pacific.

Today, we have grown our Asia Pacific cloud infrastructure to seven regions, with two more to come in Seoul and Jakarta within the first half of 2020. Australia’s demand for digital connectivity continues to grow rapidly following a boom in digital trade. According to a report by the Export Council of Australia and the Hinrich Foundation, digital trade enables A$43 billion of economic value in the domestic economy.

Google is investing and expanding its infrastructure to help Australia realise this potential and deliver on the current and future needs of its users and customers nationwide.

The new cable spans 9,200 kilometers, or the same length as 92,000 rugby fields! It’s the first subsea cable across the Great Australian Bight with a design capacity of around 36 terabits per second (Tbps).

Next up - JGA-South! 

We aren’t stopping here. The JGA consortium cable is under installation to be ready for service by Q1 2020. When it is, it will connect, Guam and Australia through 7,080 kilometres of undersea fiber optic cable system.

Stay tuned for more in the coming months on JGA-South and how we’re connecting Australia to the growing Internet ecosystem across Asia Pacific!

Posted by Ashish Ahuja and Kwok Peng Won, Global Network Infrastructure, Google Cloud

Chrome Beta for Android Update

Hi everyone! We've just released Chrome Beta 75 (75.0.3770.67) for Android: it's now available on Google Play.

You can see a partial list of the changes in the Git log. For details on new features, check out the Chromium blog, and for details on web platform updates, check here.

If you find a new issue, please let us know by filing a bug.

Ben Mason
Google Chrome

EfficientNet: Improving Accuracy and Efficiency through AutoML and Model Scaling

Convolutional neural networks (CNNs) are commonly developed at a fixed resource cost, and then scaled up in order to achieve better accuracy when more resources are made available. For example, ResNet can be scaled up from ResNet-18 to ResNet-200 by increasing the number of layers, and recently, GPipe achieved 84.3% ImageNet top-1 accuracy by scaling up a baseline CNN by a factor of four. The conventional practice for model scaling is to arbitrarily increase the CNN depth or width, or to use larger input image resolution for training and evaluation. While these methods do improve accuracy, they usually require tedious manual tuning, and still often yield suboptimal performance. What if, instead, we could find a more principled method to scale up a CNN to obtain better accuracy and efficiency?

In our ICML 2019 paper, “EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks”, we propose a novel model scaling method that uses a simple yet highly effective compound coefficient to scale up CNNs in a more structured manner. Unlike conventional approaches that arbitrarily scale network dimensions, such as width, depth and resolution, our method uniformly scales each dimension with a fixed set of scaling coefficients. Powered by this novel scaling method and recent progress on AutoML, we have developed a family of models, called EfficientNets, which superpass state-of-the-art accuracy with up to 10x better efficiency (smaller and faster).

Compound Model Scaling: A Better Way to Scale Up CNNs
In order to understand the effect of scaling the network, we systematically studied the impact of scaling different dimensions of the model. While scaling individual dimensions improves model performance, we observed that balancing all dimensions of the network—width, depth, and image resolution—against the available resources would best improve overall performance.

The first step in the compound scaling method is to perform a grid search to find the relationship between different scaling dimensions of the baseline network under a fixed resource constraint (e.g., 2x more FLOPS).This determines the appropriate scaling coefficient for each of the dimensions mentioned above. We then apply those coefficients to scale up the baseline network to the desired target model size or computational budget.

Comparison of different scaling methods. Unlike conventional scaling methods (b)-(d) that arbitrary scale a single dimension of the network, our compound scaling method uniformly scales up all dimensions in a principled way.
This compound scaling method consistently improves model accuracy and efficiency for scaling up existing models such as MobileNet (+1.4% imagenet accuracy), and ResNet (+0.7%), compared to conventional scaling methods.

EfficientNet Architecture
The effectiveness of model scaling also relies heavily on the baseline network. So, to further improve performance, we have also developed a new baseline network by performing a neural architecture search using the AutoML MNAS framework, which optimizes both accuracy and efficiency (FLOPS). The resulting architecture uses mobile inverted bottleneck convolution (MBConv), similar to MobileNetV2 and MnasNet, but is slightly larger due to an increased FLOP budget. We then scale up the baseline network to obtain a family of models, called EfficientNets.
The architecture for our baseline network EfficientNet-B0 is simple and clean, making it easier to scale and generalize.
EfficientNet Performance
We have compared our EfficientNets with other existing CNNs on ImageNet. In general, the EfficientNet models achieve both higher accuracy and better efficiency over existing CNNs, reducing parameter size and FLOPS by an order of magnitude. For example, in the high-accuracy regime, our EfficientNet-B7 reaches state-of-the-art 84.4% top-1 / 97.1% top-5 accuracy on ImageNet, while being 8.4x smaller and 6.1x faster on CPU inference than the previous Gpipe. Compared with the widely used ResNet-50, our EfficientNet-B4 uses similar FLOPS, while improving the top-1 accuracy from 76.3% of ResNet-50 to 82.6% (+6.3%).
Model Size vs. Accuracy Comparison. EfficientNet-B0 is the baseline network developed by AutoML MNAS, while Efficient-B1 to B7 are obtained by scaling up the baseline network. In particular, our EfficientNet-B7 achieves new state-of-the-art 84.4% top-1 / 97.1% top-5 accuracy, while being 8.4x smaller than the best existing CNN.
Though EfficientNets perform well on ImageNet, to be most useful, they should also transfer to other datasets. To evaluate this, we tested EfficientNets on eight widely used transfer learning datasets. EfficientNets achieved state-of-the-art accuracy in 5 out of the 8 datasets, such as CIFAR-100 (91.7%) and Flowers (98.8%), with an order of magnitude fewer parameters (up to 21x parameter reduction), suggesting that our EfficientNets also transfer well.

By providing significant improvements to model efficiency, we expect EfficientNets could potentially serve as a new foundation for future computer vision tasks. Therefore, we have open-sourced all EfficientNet models, which we hope can benefit the larger machine learning community. You can find the EfficientNet source code and TPU training scripts here.

Special thanks to Hongkun Yu, Ruoming Pang, Vijay Vasudevan, Alok Aggarwal, Barret Zoph, Xianzhi Du, Xiaodan Song, Samy Bengio, Jeff Dean, and the Google Brain team.

Source: Google AI Blog

Gmail confidential mode launching on by default on June 25, 2019

What’s changing

Earlier this year, we announced the launch of Gmail confidential mode for G Suite in beta. On June 25, 2019, this feature will become generally available and your users will have the option to send emails with confidential mode.

At launch, confidential mode will be set to default ON for all domains with Gmail enabled, unless you choose to disable this feature (see instructions below).

Who’s impacted

Admins and end users

Why you’d use it

Confidential mode provides built-in information rights management controls in your emails by allowing senders to create expiration dates and revoke previously sent messages. Because a sender can require additional authentication via text message to view an email, it’s also possible to protect data even if a recipient’s email account has been hijacked while the message is active. Additionally, with confidential mode, recipients don’t have the option to forward, copy, print, or download their content or attachments.

How to get started

  • Admins: As an administrator, you can determine whether your users can use confidential mode by navigating to Apps > G Suite > Settings for Gmail > User settings and selecting from the following options:
    • Disable: This option will prevent users in your domain from sending emails with confidential mode.
    • Enable now (beta): This option will enable confidential mode, which is currently in beta. The feature will remain enabled when it enters into general availability on June 25, 2019.
    • Enable when the feature is in general availability: This is the default option and will enable confidential mode when this feature becomes generally available on June 25, 2019.
  • End users: Once Gmail confidential mode launches, you can follow the steps in this Help Center article to send & open confidential emails.

Additional details

How Gmail confidential mode works with Vault and eDiscovery
When a user sends a confidential message, Gmail replaces the message body and attachments with a link. Only the subject and body containing the link are sent via SMTP. This means that if your users send or receive messages in Gmail confidential mode, Vault will retain, preserve, search and export confidential mode messages. The message body of received messages will be accessible in Vault only if the sender of the message is from within your organization.

Helpful links


Rollout details

G Suite editions

  • Available to all G Suite editions

On/off by default?

  • This feature will be ON by default beginning on June 25, 2019.

Stay up to date with G Suite launches

Migrating from position to impression share bidding strategies in Google Ads

In June 2019, we will shut off the ability to add new Target Search Page Location and Target Outranking Share bidding strategies. This will affect both the AdWords API and the Google Ads API. Later this year, existing campaigns on these strategies will automatically be migrated to the new Target Impression Share bidding strategy, which offers more flexible and granular controls to optimize for your desired impression share and search page location.

Read on to see how this will affect your API usage.

Google Ads API Users
Please start using the TARGET_IMPRESSION_SHARE BiddingStrategyType rather than TARGET_OUTRANK_SHARE or PAGE_ONE_PROMOTED.

If you attempt to create a TARGET_OUTRANK_SHARE or PAGE_ONE_PROMOTED after they are removed, you will get an error.

AdWords API Users
If you attempt to use the AdWords API to add a TARGET_OUTRANK_SHARE or PAGE_ONE_PROMOTED bidding strategy after they are removed, you will get an error.

All existing campaigns using the old bidding strategies will be automatically migrated to use Target Impression Share for both APIs. The AdWords API does not support the Target Impression Share bidding strategy, so we recommend migrating to the Google Ads API if you need this functionality.

If you have any questions about this change or any other API feature, please contact us via the forum.

Season of Docs Now Accepting Technical Writer Applications

Season of Docs is excited to announce that technical writer applications are now open!

In their applications, technical writers can submit project proposals based on the project ideas of participating organizations, or propose their own ideas. Refer to the guidelines on the website for how to create a technical writer application. The technical writer application form is located here:

The deadline for technical writer applications is June 28, 2019 at 18:00 UTC.

What is Season of Docs?

Documentation is essential to the adoption of open source projects as well as to the success of their communities. Season of Docs brings together technical writers and open source projects to foster collaboration and improve documentation in the open source space. You can find out more about the program on the introduction page of the website.

During the program, technical writers spend a few months working closely with an open source community. They bring their technical writing expertise to the project's documentation and, at the same time, learn about the open source project and new technologies.

Mentors from open source projects work with the technical writers to improve the project's documentation and processes. Together, they may choose to build a new documentation set, redesign the existing docs, or improve and document the project's contribution procedures and onboarding experience.

How do I take part in Season of Docs as a technical writer?

First, take a look at the technical writer guide on the website, which includes information on eligibility and the application process.

Explore the list of participating organizations and their project ideas. When you find one or more projects that interest you, you should approach the relevant open source organization directly to discuss project ideas.

Then, read create a technical writing application and submit your application using this form: The deadline for technical writer applications is June 28, 2019 at 18:00 UTC.

Is there a stipend for participating technical writers?

Yes. There is an optional stipend available to the accepted technical writers. The stipend amount is calculated based on the technical writer's home location. See the technical writer stipends page for more information.

What kind of mentor will I be working with?

Season of Docs mentors are not necessarily technical writers, and they may have little experience in technical communication. They're members of an open source organization who know the value of good documentation and who are experienced in open source processes and tools.

The relationship between you and your mentors is a collaboration. You bring documentation experience and skills to the open source organization. Your mentors contribute their knowledge of open source and code. Together, you can develop technical documentation and improve the open source project's processes.

What if I have a full time job and don't have many hours per week to devote to Season of Docs?

In the technical writer application, there is an option to apply for a long-running project, which allows technical writers to complete their project in five months instead of the standard three months. This must be agreed upon with the open source organization before work commences.

If you have any questions about the program, please email us at [email protected].

General Timeline

May 29 - June 28Technical writers submit their proposals to Season of Docs.
July 30Google announces the accepted technical writer projects.
August 1 - September 1Community bonding: Technical writers get to know mentors and the open source community, and refine their projects in collaboration with their mentors.
September 2 - November 29Technical writers work with open source mentors on the accepted projects, and submit their work at the end of the period.
December 10Google publishes the list of successfully-completed projects.

Join Us

Explore the Season of Docs website at to learn more about participating in the program. Use our logo and other promotional resources to spread the word. Examine the timeline, check out the FAQ, and apply now!

By Andrew Chen, Google Open Source and Sarah Maddox, Google Technical Writer