The two-day Google Developer Conference 2018 officially kicks off today. The conference will cover 12 topics such as machine learning, artificial intelligence, Android and ARCore, and will be presented to you in the form of 60 lectures.

In the keynote session at the beginning of the conference, Google Greater China President Bo Meng Shi and subsequent speakers said that 5.5 million Chinese developers have been following Google I/O in the past, and they have always expressed the same belief: AI is starting to make a difference in all sorts of areas. Google has always believed in THE power of AI and has always wanted to make it better for everyone.

“Google wants to use Android and TensorFlow to help developers around the world create hundreds of thousands of great apps,” He said. At the event, Google also demonstrated two programs, universal app advertising and Google Growth Guide, aimed at helping developers better.

Shortly before the conference, Fei-fei Li, who helped build Google’S AI Center in China, returned to Stanford as a professor. He was replaced by Andrew Moore, former dean of CMU’s School of Computer Science, as head of Google cloud AI.

More specific content will also see the next two days of each sub-forum in the wonderful speech. The following are the highlights of the conference mentioned in the keynote speech

1. Pictionary song: As a warm-up, Google officially opened a pictionary song game contest, inviting more than 2,000 audience to participate at the same time. Finally, the person who guesses the fastest and most in the same round will win the prize provided by Google as 21 pictures. Andrew Bowers, Google’s director of product, also announced the launch of the “Friends to Play” game today.

Since its launch two months ago, the song has been used by more than a million users, with 500 million drawings correctly guessed by AI. In addition to picture-guessing songs, Google has also used such games for arts and culture projects, for example, by allowing users to draw an object and match it with the closest artwork.

2. Android system: AI has profoundly changed the latest version of Android system, making it more intelligent, stable and compatible. For example, adaptive batteries. Android can use machine learning to determine which apps are commonly used, saving CPU energy by 30%.

The number of Android developers in China has tripled in the past two years, and developer revenue in China grew by 50% last year. Google also explained how to use the new Android Jetpack framework to speed up app development, and how to use the Kotlin static programming language to simplify code.

3. TensorFlow: TensorFlow is an open source software library for machine learning for a variety of perceptual and language understanding tasks. TensorFlow is currently used by 50 teams to research and produce many Of Google’s business products, such as voice recognition, Gmail, Google Albums and search, many of which used its predecessor, DistBelief.

Currently, it has 17 million downloads and 39,000 submissions worldwide; 1,600 people have contributed and 11,500 questions have been answered, which is why Google keeps the platform open.

4. Fiutter: As Google’s new cross-platform framework, Flutter has many advantages through its own rendering engine, such as high performance and the ease of building a smoother UI. The conference introduced Fiutter’s current partners and detailed new features using The example of Idle Fish, which is used by 50 million users.

5. Wear OS: Google released the first generation of Wear OS in 2015. Now, More smartwatches are carrying Wear OS, and It hopes to provide users with a more perfect platform. Google is also looking forward to working with Chinese developers and actively adding partners to create pleasant experiences for users.

Currently, smartwatches equipped with The Chinese version of Wear OS have the fitness function “Xiaowan Sports” by default. Right swipe to bring up the intelligent assistant “Little Ask Assistant”; Slide can easily browse information, timely processing wechat and other notices.

6. Fairbase: Firebase provides cross-platform solutions for data storage and helps developers provide application quality. Currently, it has 1.2 million monthly active applications. ML Kit is a machine learning suite for Firebase that allows developers to deploy new features to the cloud without updating their apps.

7. Ar Core: ARCore has been applied in the products of xiaomi, Huawei, Samsung, Vivo and many other manufacturers, and 175 million mobile phones around the world can use ARCore. It’s expected that by 2020, Chinese developers will account for 25% of the global AR revenues. To this end, Google has built Sceneform, Cloud Anchors and other features that make it easier for developers to write AR apps and allow everyone to come together and play AR games.

A look back at the highlights of Google’s previous conferences

Google Cloud Next 2018

At Google Cloud Next 2018, Fei-Fei Li and Jia-Fei Li announced another milestone in their Google Cloud: AI call center, AutoML’s natural language and translation service, and TPU 3.0.

There are two core highlights of this product: 1. Cloud AutoML has two new functions of language classification and translation. Many media groups already use them; 2. AI that has passed the Turing test can be put into commercial use. The scene shows how AI can perfectly help e-commerce users to return goods through voice communication.

In January, Google announced its Cloud AutoML, which allows you to train a customized machine learning model even if you don’t know much about machine learning.

Fei-fei Li has mentioned his resignation in moments of friends. Cloud AI has been highly rated as a leading team in the industry by Forrester, MIT Technology Review and Forbes.

“We created together multiple influential products include AutoML, Contact Center, AI DialogflowEnterprise, Vision/researched/NL/Translation APIs, Cloud AI platform and so on. Our industry solutions (retail, healthcare, energy) are also gaining traction.”

Google I/O 2018

Google announced the merger of Google Research with Google.AI to create a new Google AI that will take charge of all Google Research projects and ARTIFICIAL intelligence. Google AI aims to make AI tools available to everyone, advance technology in the field, and commit to AI applications and new areas.

Google Assistant, which announced it is available in more than 25 countries and regions and can control more than 500 million smart devices, was also given the ability to carry out multiple rounds of conversations at the conference, as well as the ability to identify the people being spoken to.

There is also TPU, Google’s dedicated chip for machine learning, designed for TensorFlow, Google’s deep learning framework, which is only used in Google products such as Street View, DeepMind and AlphaGo.

Google News has also been completely reconfigured for iOS, Android and the Web to help users subscribe to multiple publication sources with the help of ARTIFICIAL intelligence and machine deep learning.

The Waymo CEO also described how Google’s brain and machine learning technology could be used to solve various problems at the self-driving center, such as allowing the self-driving car to spot pedestrians next to it and avoid danger by predicting the actions of pedestrians and cars on the road.

Google Developer Conference 2017

At the conference, Fei-fei Li, chief scientist of Google Cloud’s AI and machine learning team, announced the launch of Google’S AI China Center in Beijing. The center is led by Li Feifei and Dr. Jia Li, head of Google Cloud research and development. Feifei Li will lead the research work at the center, as well as co-ordinate the work of Google Cloud AI, Google Brain and the Local Team in China.

In addition to publishing its own research results, Google AI China is also looking forward to building on local collaborations in China to provide high-quality AI and machine learning education support to a wider audience of students and researchers.

In addition to this, Google announced a series of key numbers: Google has developed a very powerful TPU Pod that can currently do 11.5 PetaFlops per second and 4 terabytes of memory.

There are 2 billion active Android devices and 82 billion apps installed on Google Play worldwide. There are more than 140 carriers that offer withholding payments, covering 900 million mobile devices. Add it all up, and the number of people spending money on Google Play grew 30 percent last year.

Google Smart Assistant works on millions of devices, including voice-activated speakers like Google Home, eligible Android phones, Android TV, Google Pixelbooks, headphones and Android Wear, Android Auto is coming soon.

At the same time, Google also released the Google Assistant SDK, which empowers Google Smart Assistant into custom hardware projects. The Google Assistant SDK works on Linux, Android Things, and any other platform that supports gRPC.

Google I/O 2017

The main focus of Google’s conference was AI, which has been integrated into every aspect of Google’s products and software systems. Google, for example, unveiled its own DEDICATED AI chip, TPU3.0, which has eight times the computing power of TPU2.0 and can solve more problems faster and smarter.

In addition, Google Assistant got a good look at the conference, which introduced six new sounds and announced that it now supports 30 voices and is available in 80 countries.

In addition, Google News uses AI to intervene in the way people read News, recommending relevant reading to users and pushing other relevant News.

The Beta version of Android P is officially launched. More than half of the adaptive phone manufacturers mentioned by Google are from China, with Xiaomi, OnePlus and Vivo becoming the first batch of supporting manufacturers.

The conference also featured a recent update on Waymo’s driverless cars, which are already on the road in Phoenix, Arizona, and have 25,000 vehicles on the road each day to verify various road conditions. The intervention of AI has reduced driverless errors by more than 100 times, making driverless cars safer.

[For more information on artificial intelligence, please pay attention to the wechat public account “Rebuild_ai”]