Google I/O 2017: Everything important that Google announced today

谷歌I/O 2017大会:谷歌宣布的所有重大事件

Recode · 2017-05-26 14:37

Recode · 2017-05-26 14:37

At Google’s 2017 I/O keynote today, CEO Sundar Pichai introduced new products and shared more information about the company’s “AI first” future. Here’s a running list of what happened that matters.

在今天谷歌2017年I/O大会上,CEO Sundar Pichai介绍了新的产品并分享了很多关于公司“AI第一”的未来。这里有一份关于所有重要事件的清单。

Google is rethinking “all” of its products for an AI-first world. That’s the high-level promise from Pichai, and the change Google must successfully navigate to continue its dominance. Examples: Google Search now ranks differently using machine learning, Google Maps Street View automatically recognizes signs, video calling uses machine learning for low-bandwidth situations, etc.

谷歌为了AI第一的世界重新考量“所有”产品。这是来自高层Pichai的承诺,并且是谷歌为了继续保持优势地位必须成功探索的改变。例如:谷歌搜索(Google Search)现在利用机器学习可以呈现不同的搜索排位,谷歌地图街景图(Google Maps Street View)能自动识别标志,利用机器学习在低带宽地区进行视频聊天等。

Google can now use your camera as an input device. Google announced a “Lens” feature for its Assistant service that will tell you information about what’s in front of your phone camera: What type of flower you’re looking at, info about the restaurant across the street, etc. This combines a bunch of tech buzzwords — AI, AR, machine learning, computer vision — into one feature, and highlights how Google’s massive collection of data about the world can be used together. It's also another example of how Google is trying to get ahead of the latest preference for photos versus text.

谷歌现在可以把你的摄像头作为输入设备。谷歌在它的助理(Assistant)服务上宣布了一个“镜片”功能,它会告诉你在你手机摄像头前面任何东西的信息:你正在看的花是什么种类,街对面那家餐厅的信息等。这结合了许多科技的热词—AI, AR, 机器学习,计算机视觉—在一个功能里,并且强调了谷歌将如何使用它所收集的如此庞大的数据。这同样也是另一个谷歌关于最新的在照片还是文字的选择中试图领先的例子。

Google’s latest user numbers: There are now 2 billion active Android devices, 500 million active Google Photos users and 1.2 billion videos and photos uploaded to Google Photos every day. Users navigate more than 1 billion kilometers per day with Google Maps. Google Drive has 800 million users. And people watch 1 billion hours of YouTube videos a day.

谷歌最新的用户数量:目前有20亿台活跃的安卓设备,5亿位谷歌照片(Google Photos)的活跃用户并且每日有12亿条视频和照片上传到谷歌照片。用户每日利用谷歌地图(Google Maps)导航超过10亿公里。谷歌云盘(Google Drive)有8亿用户。人们每天共观看10亿小时的YouTube视频。

Don’t like writing emails? Don’t have to. Gmail on Android and iOS is getting an automated-response feature called Smart Reply, which lets you choose up to three responses that are based on the received email. You can add more to a response after selecting it if you want, or you can just send the automated response as-is.

不喜欢写电子邮件?没关系。安卓和iOS系统上的Gmail现在有一个自动回复功能,叫做智能回复(Smart Reply),它可以根据你的收信让你选择最多三个回复。你可以选择之后在回复里添加更多内容,或者直接就用现成的自动回复。

More Google in more places, in more languages. Google Assistant is rolling out to the iPhone — though it won’t (and can’t) replace Siri as the default assistant — and will be available in more languages like French and German.

更多地方的谷歌,更多语言的谷歌。谷歌助理(Google Assistant)在iPhone上也开始登陆—虽然它不会(也不可能)取代Siri作为默认助理—并且会有更多的语言比如法语和德语。

Hands-free calling on Google Home. “Hey Google, call Mom” will cause the device to dial out with your personal number when it detects your voice. No setup required, according to today’s presentation. (Amazon launched calls for Alexa this week, too.) It’s the new landline phone.

免手动的呼叫谷歌家居(Google Home)。“嘿谷歌,打给老妈”当它探测到你的声音的时候,将会让设备使用你的个人号码打电话。根据今天的讲述,不用任何设置。(亚马逊也会在这个星期让Alexa打电话)。这是全新的座机。

Google Photos has a cure for laziness. Google’s photo service, which the company says has 500 million users, can now suggest what photos you should share — and with whom — with a feature called suggested share. It can also automatically share photos with an option called shared libraries. The big idea: Most photos just sit there on your phone. It’s better for you and your friends (and for Google!) if you share more of them.


Super Chat for YouTube. While watching a YouTube livestream, you can now pay to have your comment highlighted as a “super chat” message. This makes the message stand out — so the creator is more likely to see it in an endless stream of messages — and also generates some more cash for the YouTuber.


Google is working with partners to launch standalone VR headsets. That means everything for a VR experience will be built into the headset itself — no phone or PC required. The headsets, running Google’s Daydream platform and made by HTC and Lenovo, are slated to ship later this year.


Google for Jobs. Google is taking on LinkedIn with job listings in its main search product. If you search for “retail jobs,” for example, Google will know where you’re searching from and show jobs in your area.


Apps and transactions on Assistant. Developers can now build apps or “actions” that run on Google Assistant on Android and iOS. Already, developers have been able to build actions for Assistant on the Home device. Developers will also be able to build transaction features for Assistant, which will soon be available on phones with Assistant.


Android Go. Google is launching an initiative called Android Go to better tailor Android to low-connectivity devices. Starting with the release of Android O, the latest version Android not yet released publicly, devices with 1 gigabyte or less memory will receive versions of apps like YouTube and Chrome that use less memory. The software is also supposed to have features tailored for users who speak multiple languages.

安卓Go。谷歌现在正在发起一个专门为低连接装置设计的叫安卓Go(Android Go)的计划。从安卓O的发行开始,这是安卓最新的还未面世的产品,1G或者更少储存空间的设备将可以使用更小内存版本的应用程序,如YouTube和Chrome。这个软件同时也对多语种的用户提供了专门的功能。

Indoor mapping. Google is introducing something called visual positioning service, or VPS, that will allow you to map indoor locations using its Tango AR platform. An example of what VPS can do is tell a user the exact location of a product in a store.

室内定位。谷歌正在介绍一种叫视觉定位的服务(VPS),它将可以通过Tango AR平台来对室内位置定位。比如告诉一个用户某一个产品在商店里的具体位置。

Machine learning hardware for the cloud. Google introduced the second version of its Tensorflow Processing Unit, or TPU, called the Cloud TPU. That’s the hardware it uses for Tensorflow, its open-source machine learning software, which developers can use to build their own AI-powered tools and apps.

云端的机器学习硬件。谷歌介绍了Tensorflow Processing Unit(TPU)的第二个版本,叫做云端TPU(Cloud TPU)。这是Tensorflow所使用的硬件,它的开源机器学习软件,这个开发者们可以用来创建他们自己的AI驱动工具和应用。(5月17日)


本文转载自Recode ,如若转载,请注明以下出处:




文章评论( 0 )


请先登录在评论 没有账号?注册



评论 4 分享