北京時間1月9日早間，Apollo平臺研發負責人王京傲在美國拉斯維加斯百度world大會宣布Apollo2.0正式開放，百度COO陸奇進行了開場演講。
百度發布會現場百度發布會現場
Apollo2.0能夠實現簡單城市道路自動駕駛，標志著Apollo平臺包括云端服務、軟件平臺、參考硬件平臺以及參考車輛平臺在內的四大模塊已全部具備。Apollo2.0帶來最完整的解決方案和靈活的架構，首次開放安全服務，并進一步強化了自定位、感知、規劃決策和云端仿真等能力。
王京傲介紹，目前Apollo2.0版本總共有16.5萬行代碼。硅谷自動駕駛創業公司AutonomouStuff在一周內將Apollo1.0車輛升級為“Apollo2.0版本”，實現了晝夜簡單城市道路自動駕駛，充分體現Apollo2.0的靈活性和易用性。
根據王京傲介紹，Apollo對開發者非常友好：
1、本地化做的非常好。1.0版本的時候，我們請CEO去國家會議中心門口來做演示，就是因為我們的本地化做的非常好。
2、規劃技術非常高。經過路口、紅綠燈、掉頭、轉彎時，Apollo2.0路測的時候，有記者說”其實挺無聊的，跟人開車沒什么區別。
3、可以做交通燈甄別，可以看到500米外的物體，可以甄別障礙物。
在Apollo2.0要做仿真甄別，之前要用30分鐘把數據上傳云端，但是現在時間縮短到30秒。Apollo2.0不管日間還是夜間都可以駕駛。Apollo現在不僅在圣荷西和北京，并且在伊利諾州也可以上路。
Apollo還宣布將在2018年全面支持包括NVIDIA、Intel、NXP、Renesas在內的四大主流計算平臺。Apollo已成為全球最具活力的自動駕駛開放平臺，陸續還會有更多合作伙伴加入。
王京傲展示了Apollo Pilot與奇瑞、金龍以及面向殘障人群的出行服務商Access LA等合作伙的落地應用成果，涵蓋了乘用車、公共巴士和共享交通服務等多種場景。百度將與奇瑞在2020年量產L3級別自動駕駛汽車，與金龍客車在2018年量產無人駕駛微循環巴士，并與Access LA在2018年底在洛杉磯為當地殘障人士推出自動駕駛共享出行的試點服務。
此外在演講中陸奇表示“AI is borderless”，技術創新對全球所有人都有利，中國具備人口、數據、產業集群、政策的結構化優勢，盡管從高端技術和人才來講美國仍然是領先的，但中國和美國之間的差距在快速縮小
“百度的AI在以非常、非常快的速度發展，百度在2013年率先創建了一個深度學習研究院，我們2014年在硅谷建立了一個實驗室，叫做Baidu Research，2015年，我們發布了DuerOS的第一個版本，它是一個對話式AI平臺，2016年，百度發布了自動駕駛汽車技術和百度大腦，2017年，我們通過一系列新的AI產品驅動，尤其是發布Apollo自動駕駛平臺明確了公司的AI發展方向”陸奇這樣概括了百度的AI發展歷程。
現場還有一個無人車視頻連線，這次可不是李彥宏上五環，“AI is Changing the World, China Speed”陸奇在拉斯維加斯向全球開發者、合作伙伴展示中國速度的方式——遠程連線北京清晨7點鐘的北京，十輛阿波羅無人車繞百度大廈開跑!
附：陸奇演講全文(中英文對照版本)
Thank you! Thank you very much! Welcome to the first ever Baidu World at Las Vegas!
謝謝!非常感謝!歡迎來到在拉斯維加斯舉辦的首次Baidu World!
Today we are very excited about the opportunity to tell you the latest AI innovations, the new platforms, the new products, and more importantly, I will share with you how Baidu is accelerating AI innovation at China speed, how through open platforms we can work together to change the world and make it a better place.
今天我們非常高興能有機會向大家介紹最新AI創新、新平臺和新產品，更重要的是，我將跟大家分享百度是怎樣以中國速度在加速AI創新的以及通過開放平臺我們可以怎樣來共同改變世界，并讓世界變得更好。
As we all know, the AI era has already arrived, and it has profound social and economic impacts. It will drive the upcoming Fourth Industrial Revolution. The reason, Why does it have so much power? (It is) because the AI technology gives us the unprecedented capability. And this picture captures the essence of AI innovation. At its very core, AI technology is about systems that we build using software, hardware and algorithms. But the systems have the capability to automatically learn and acquire knowledge from data and apply that knowledge to achieve our technical goals to outcome we want. And that’s the essence of AI revolution, which is the acceleration of harnessing knowledge. And this is the unprecedented part, because Francis Bacon said, “Knowledge is power.” With the right knowledge, we can solve any problems that have happened to us. But to drive AI innovations, data will play a central role. Data is the carrier of knowledge. Without data, we can’t build the systems that understand human conversations. Only with our data, can we build the systems that can recognize our objects on the road and can be used for autonomous driving platform. So therefore, capturing and harnessing data will be a deciding factor for how broad and how fast AI innovation will proceed. And this is why AI as a technology has already been widely applied across a broad spectrum, and really transforming the existing industries and creating powerful new ones.
我們都知道，AI時代已經到來，而且它具有深遠的社會和經濟影響。它將驅動即將到來的第四次工業革命。為什么?為什么它有如此大的力量?這是因為AI技術給了我們前所未有的能力。這張圖片捕捉到了AI創新的實質。在其核心層面上，AI技術是關于我們用軟件、硬件和算法所打造的系統的技術。但這些系統有能力自動從數據中學習和獲取知識，并利用這種知識來幫助我們實現技術目標。這是AI創新的實質，即我們利用知識的速度的加快。這是前所未有的，因為培根說過“知識就是力量”。通過正確的知識，我們能解決自己所碰到的任何問題。但為了驅動AI創新，數據將起到核心作用。數據是知識的載體。如果沒有數據，我們將無法打造可以理解人類對話的系統。只有通過數據，我們才能打造出能夠識別道路上的物體、可用于自動駕駛平臺的系統。因此，捕捉和利用數據將是決定AI創新發展范圍有多大、速度有多快的一個決定因素。這也是為什么AI作為一項技術已經在很大范圍內得到了廣泛應用的原因，而且它也在真正地改變著現有產業、創建強大的新產業。
The AI technology has already been widely applied across a broad spectrum, transforming and renovating the existing industries and creating powerful new ones. And this is also why “China+ AI” will have a special opportunity to unlock more AI innovations at a fast pace. And this is because structurally China will have a unique advantage. As we all know, the tech industry requires several structural elements—talent, technology, capital, market. Historically, the United States is the only country that has all those elements. But now China has all those as well. Even though technology talent-wise at the top end the United States is still ahead, but the gap between China and the United States is closing, and is closing fast. And the advantage we have is a large population and a fast growing market. And the overall friendly and favorable policy environment is designed to enable more and actually harnessing data and nurture more AI innovations.
AI技術已經在很大范圍內得到了廣泛應用，在改變和革新現有產業、創建強大的新產業。這也是為什么“China+ AI”將有一個特別的機會來快速釋放更多AI創新的原因。這是因為中國從結構上來講將有一個獨特的優勢。我們都知道，科技行業需要幾個結構要素——人才、技術、資本和市場。歷史上，美國是惟一擁有所有這些要素的國家。但現在中國也擁有了所有這些要素。盡管從高端技術和人才來講美國仍然是領先的，但中國和美國之間的差距在縮小、在快速縮小。而我們所具有的優勢是人口多、市場發展快。整體上友好和有利的政策環境也有助于產生和利用更多數據以及培育更多AI創新。
And this is to begin with. China will have a lot more data for AI to work with. What is to start with is people. China has a population of 1.3 billion people with 1.1 billion smartphone users and an online population of 750 million people all in the same market. For any AI innovation that is related to people such as healthcare and education, there is just more data to experiment, to try different ideas, and to realize implementation at a faster pace. China also has the largest manufacturing and traditional industry base. For our industry alone, China has more than 200 million vehicles and over 200 OEMs. There is a lot more data to tap into. There is a lot more opportunities for AI to drive innovation.
這是我們發展的基礎。中國將擁有更多的數據供AI利用。首先是人。中國有13億人，其中包括11億手機用戶和7.50億網民，而且他們都在同一市場上。對如醫療和教育等與人有關的任何AI創新來說，我們有更多數據來實驗、來嘗試不同想法、來以更快速度實施。中國還擁有最大的制造業和傳統產業集群。僅僅對我們這個行業來說，中國就擁有超過2億輛汽車和超過200家整車廠。我們有更多的數據可供挖掘。我們有更多的機會來讓AI推動創新。
In addition, China also has a favorable overall environment, because the Chinese government has a good grasp of what AI can do, and this is a historical opportunity for China to ride the AI wave to become an innovation power. And the government has shown a lot of commitment and willingness to invest for long term. AI is already part of the government’s five-year plan. Just a few months ago, the government published a comprehensive white paper for systematically investing in AI. At the same time, the government is also much more willing to apply the policies to support and nurture AI innovation, such as the support of the national AI platforms and policies that are designed to enable autonomous driving test and R&D.
另外，中國還有一個有利的整體環境，因為中國政府對于AI能做什么有很好的把握，這是讓中國借助AI浪潮來成為創新強國的一個歷史機遇。而且政府也顯示了進行長期投入的很大誠意和意愿。AI已經是政府五年計劃的一部分。僅僅幾個月前，政府發布了一個關于對AI進行系統性投入的全面的白皮書。與此同時，政府也非常愿意通過政策來支持和培育AI創新，如對國家AI平臺的支持以及旨在讓自動駕駛汽車測試和研發能夠得以進行的相關政策。
This is all great news for all of us around the world, because AI is global and technology innovation benefits everyone on the globe. Wherever a new cancer treatment is discovered, it saves lives everywhere. Anytime a new way of learning is discovered, it helps everyone across the globe to become better educated. For all of us working in the tech industry in China, doing AI innovations will enable us to embrace the opportunities, to lead, and to serve more people around the world, because together we can change the world and make it a better place.
這些在世界范圍內對于我們大家來說都是好消息，因為AI是全球性的，技術創新對全球所有人都有利。不管是哪里發現了癌癥的一個新療法，它在所有地方都可以挽救生命。任何時候一個新的學習方法被發現，它都可以幫助全球每個人受到更好的教育。對我們在中國科技行業的所有人來說，進行AI創新將使我們能夠擁抱機會、引領發展和服務世界各地更多的人，因為通過共同努力我們能夠改變世界、使其成為一個更美好的地方。
Baidu is exactly what all of this is about, working super hard to drive all innovations. Here is a quick introduction of the company:
百度正是為所有這一切而生，我們正在極為努力地推進所有這些創新。這里我對本公司做一個簡單介紹：
For all of you, it’s simple—Baidu essentially is the Google of China. Because of such a heritage, just like Google, Baidu is clearly the leading AI company in China, as the best AI technology, the overall best talent, and the vast amount of data assets are coming from the leading suite of online services.
對大家來說，很簡單——百度實質上就是中國的谷歌。由于這樣一種傳統，就像谷歌一樣，百度顯然是中國領先的AI公司，因為最好的AI技術、最好的人才和大量數據資產都來自我們領先的在線服務。
Baidu’s AI is also developing at a very, very fast pace. Baidu is among the very first to create a deep learning institute in 2013. We created a Silicon Valley lab, Baidu Research, in 2014. In 2015, we launched the first DuerOS version, which is a conversational AI platform. In 2016, Baidu debuted the self-driving car technology and also the Baidu Brain AI platform. In 2017, we pivotally tied the company towards AI with a series of new AI engines, especially the launch of the Apollo autonomous driving platform.
百度的AI也在以非常、非常快的速度發展。百度在2013年率先創建了一個深度學習研究院。我們2014年在硅谷創建了一個實驗室，叫做Baidu Research。2015年，我們發布了第一個版本的DuerOS，它是一個對話AI平臺。2016年，百度發布了自動駕駛汽車技術和百度大腦AI技術平臺。2017年，我們通過一系列新的AI創新引擎，尤其是Apollo自動駕駛平臺的發布，確定了公司的AI發展方向。
Going forward, Baidu is an AI company. AI is at the center of driving everything we do at Baidu. To begin with, we use AI to transform and renovate Baidu’s current core businesses—search, personalized feed in Mobile Baidu, and also video entertainment in IQIYI. For those of you, IQIYI is essentially the Netflix of China and much more.
今后，百度將是一家AI公司。AI處在推動我們在百度所做所有事情的核心位置。首先，我們利用AI來改變和創新百度當前的核心業務——搜索、“手機百度”的個性化信息推送、以及愛奇藝的視頻娛樂服務。愛奇藝實質上就是中國的Netflix，但所提供的內容要多得多。
In addition, we are gonna build a number of new AI-enabled businesses, especially Apollo for autonomous driving and DuerOS for conversational AI. These are huge business opportunities for Baidu to grow into. They will also fully leverage Baidu’s core competency and assets. And all those are built on top of the common technology platforms of Baidu Brain and Baidu Intelligent Cloud. Baidu Brain is clearly the most complete AI platform in China. As the low level infrastructure, over the years, Baidu has developed first-class cloud infrastructures and strong data assets and data processing capabilities.
另外，我們還要打造若干新的AI賦能的業務，尤其是用于自動駕駛的Apollo和用于對話式人機交互的DuerOS。這些對于百度來說都是可以把握的巨大商機。它們也都將充分借助百度的核心競爭力和核心資產。而且所有這些都建立在百度大腦和百度智能云的通用技術平臺之上。百度大腦顯然是中國最全面的AI平臺。作為底層基礎設施，多年來，百度打造了一流的云基礎設施和強大的數據資產及數據處理能力。
In other parts of AI competition, Baidu is very strong in AI algorithms, and we have powerful AI platforms for deep learning on both car side and on devices. Baidu Brain is also the first platform to clearly separate the perceptual layer and the cognitive layer. Both of them offer very comprehensive capabilities for all sorts of applications. Putting all together, Baidu Brain is a potent platform that is fueling massive innovations at Baidu and also for all other car innovations at our partners. That’s why as a result Baidu Brain today is already the most vibrant AI platform in China. Our third-party developer is growing very, very rapidly. The daily call volume to the platform is skyrocketing. And at the same time, the number of applications, the variety, all different types of AI innovation simply is growing at a massive scale and a rapid pace.
在AI競爭的其他領域，百度在AI算法方面實力非常強，我們擁有供汽車和智能設備使用的強大的深度學習AI平臺。百度大腦也是第一個將感知層與認知層明確分開的平臺。二者對所有種類的應用都能提供非常全面的能力。綜合來講，百度大腦是推動百度進行大規模創新的強大基礎平臺，對我們合作伙伴所進行的其他所有有關汽車的創新來說它也是一個強大平臺。這也是為什么百度大腦今天已經成為中國最有活力AI平臺。我們的第三方開發者在非常、非常迅速地發展。我們平臺的日均調用次數在急速增長。與此同時，應用的數量和種類以及所有不同類型的AI創新也都在以巨大的規模快速增長。
Third-party applications simply are promising. And this is very symbiotic between Baidu and the “China+ AI” macro-environment, because with more third-party innovations, it helps strengthen the platform. At the same time, a stronger platform will drive more, and accelerate the pace, of AI productization and AI commercialization for the core AI-enabled businesses for Baidu.
第三方應用很有前景，這是百度與“China+ AI”宏觀環境之間非常密切的共生關系，因為通過更多第三方創新，它可幫助增強這一平臺。與此同時，一個更強大的平臺也將有助于推動更多AI創新，加快百度核心AI業務的產品化速度和商業化速度。
This is the main part of today’s presentation, which is to unveil the new AI platforms and showcase new AI products, specifically Apollo for autonomous driving and DuerOS for conversational platforms. And our fundamental approach for those is through open APIs and open platforms to build innovation ecosystems together to accelerate the pace of innovation.
這是我今天演講的主要部分，即發布新的AI平臺和展示新的AI產品，特別是用于自動駕駛的Apollo平臺和用于對話的DuerOS平臺。而我們用于這些平臺的根本方法是，通過開放APIs和開放平臺來共同打造創新生態系統，以加快創新速度。
眾視IoT記者親臨拉斯維加斯現場，
（點擊圖片進入專題）
微信掃一掃
