Wen Yi Xiao Yi The United States local time on May 8th at 10 a.m. (Beijing time 9 on the morning of 1 oclock), 2018 Google I/O Developers Conference in California mountain city officially opened. NetEase intelligent first time for everyone to comb the key content of this conference, lets take a look. |GoogleAssistant (Google assistant): you can make a phone call to help you book a restaurant. Googles AIFirst has entered the third year, and this I/O developers conference also fully bet on AI technology. Google CEO Sundar Pichai (SundarPichai) first came to power, and it was a little surprising that the first topic he had thrown was not AndroidP, but GoogleAssistant (Google assistant). Pichai showed a review of AIs promotion of all walks of life, such as AI technology to help detect retinopathy, help disabled people to speak properly, automatically predict input text, and also mention the speed of TPU3.0 promotion. Then, Peter began to focus on the upgrading of GoogleAssistant. First, Google assistant interactive voice has been upgraded to six new voices. Second, support multiple rounds of continuous dialogue and make dialogues more natural. Third, it supports multiple replies of automatic decomposition dialogues, for example, when users interact with Google assistants, two questions are asked in one sentence, and Google aides can automatically answer them separately. Fourth, Google aides combine the sound with the screen in a greater degree, with a particular reference to the smart display that Google releases on CES2018 (also known as a smart phone with a screen). Fifth, and the most dazzling point, Google aides have added a number of services, such as selling voice interaction points, asking the route, booking a barber shop, and booking a hotel. A special point is that Google assistants support voice duplex technology. Users need only to ask for demand. Google assistants can make a direct call to a barber shop, a hotel, and send a notification to the user, without the need for anything from the user itself. However, considering the different people, different environments and different contents, the performance of Google assistants voice duplex technology in the actual situation remains to be tested. According to the data released at the conference, Google aides have landed 500 million devices, support 30 languages, more than 80 countries are available, and have 130 thousand childrens stories. |GoogleNews (Google News): using algorithms to support in-depth reporting The second item that focuses on upgrading is Google news. This time Google news also introduced a large number of AI technology, which can automatically detect and screen the most important news, select the most concerned content according to the users preferences, find the relationship between people, places and affairs through Newscasts technology, help the users to understand the background and find the depth of analysis. Overall, the upgrade of Google news is mainly to understand users habits and focus on important events. At the same time, users can find their favorite news sources and personalize them in Google news. The new Google news will start next week in 127 countries around the world, and will support Android and iOSs two operating systems. |AndroidP: Open Beta downloads, people who cant sleep with mobile phones are saved. This year is the ten year anniversary of the Android system, and AndroidP is still one of the highlights of this conference. Pichai first thanked developers for their support for Android over the past ten years, and then opened the three upgrades to AndroidP, intelligent, simple, and digital. First of all, AndroidP is more intelligent, especially through machine learning technology to predict the next action of users. On AndroidP, the two functions of adaptive battery and adaptive luminance are added, mainly based on large data and deep learning technology of Google. It is intelligent to judge the users use of App, automatically change the load of CPU and the brightness of the mobile phone, and make the cell phone more long. In addition, AndroidP has also improved the mobile home page search, not only to predict which application is used by the user, but also to predict which function the user uses and give priority to display. Google also mentioned that the machine learning kit MLKit has supported cross platform use. Secondly, AndroidP changed the bottom navigation bar and added a new gesture operation. Specifically, the mobile desktop cancels the traditional bottom three buttons and turns into a single bar virtual key. The slide on the mobile desktop can be transferred out of the application list, and the horizontal slide can be transferred to multitask. Thirdly, AndroidP joined the DashBoard function and Shush mode. DashBoard can count the time the user spends on each App, and analyze it, remind the user to use the time of the mobile phone, and set the application time limit. This is a good help for the user who likes to watch the mobile phone before sleeping. When the cell phone reaches the time limit set by the user, it will turn the screen into gray, so that the user will not play the cell phone too late. The Shush mode is an upgraded version of no disturb. After setting this mode, the phone will not only be disturbed by the sound vibration, but will not even be notified by the message. The Shush mode can be activated by a flip phone. Google announced at the meeting that the AndroidPBeta beta version was officially open and downloaded. In addition to his own Pixel series, the first batch also supported eight mobile phones such as millet, OPPO, vivo, NOKIA, SONY and EssentialPhone. |GoogleMaps (Google map): enhance service and social attributes Google map has also been upgraded at this press conference. The first is that the Google map can tell the user whether it is difficult to stop at the destination by predicting traffic and analyzing the data outside, which way of transportation is easier in the past. Second, Google maps also enhance social properties, not only to recommend new cafes and restaurants in the community, but also to provide restaurant selection functions at a friends lunch time. Again, Google maps add a very cool feature, using hardware and technology, such as mobile camera, navigation, visual positioning system, and other hardware and technology, to accurately position and guide. The user only needs to focus the mobile camera on the building road card, which can identify which direction the user should go. In this process, the most important thing is to use the GoogleLens function. At present, GoogleLens can not only identify objects such as traffic signs, posters, menus, but also support style matching and search for similar style objects. Waymo: object awareness and action prediction upgrade This is Waymos first Google I/O conference, and WaymoCEO John Klavsk (JohnKrafcik) describes how to use Google brain and machine learning to solve all kinds of problems in unmanned research and development, such as driving unmanned vehicles to pedestrians, and so on. Waymo responsible people think that the perception of unmanned vehicles to the surrounding objects and the environment is very important, through continuous data training, predict the actions of people and cars on the road. Waymo revealed that its pilotless car has been running 6 million miles, and has been driving 5 billion miles in the companys simulation system, and the team is exploring unmanned driving in the snow. Pay attention to NetEase intelligent public address (smartman163), to interpret the big events, new ideas and new applications of large companies in AI field. Source: NetEase intelligent responsible editor: Wang Chao _NT4133 Waymo revealed that its pilotless car has been running 6 million miles, and has been driving 5 billion miles in the companys simulation system, and the team is exploring unmanned driving in the snow. Pay attention to NetEase intelligent public address (smartman163), to interpret the big events, new ideas and new applications of large companies in AI field.