Ai's heat in science and technology through Li's proposal

Today, our world is entering an information world connected to everything, and the rise of big data and the collision of Internet of things have inspired a greater era-the age of artificial intelligence.

30 years ago, Bill Gates resolutely dropped out, founded Microsoft, became the leader of the personal computer popularization revolution, 30 years later, he predicts that the robot will repeat the road of the PC's rise. Li at the two sessions proposed to build the "Chinese brain", hope that through the country-level infrastructure to promote, so that AI will be like today's computer as Universal Society, a more thorough revolution.

Before a huge revolution, the opportunity is enormous, and who can seize the opportunity of artificial intelligence, who will be in the next 30 years to become the vanguard of global technology. At the moment, global technology giants are beginning to speed up the development of artificial intelligence technology to seize this commanding point.

　　Google, through acquisitions, continues to climb the new heights of smart technology

A few days ago, Google researchers have just announced that their newly developed "Deep Q Network" Computing program is expected to advance the AI technology a big step forward, breaking through the previous single algorithm can only correspond to a single task limit, allowing computers to achieve closer to human independent learning and reasoning capabilities. In fact, over the years, Google has been devoting huge manpower, funds to develop artificial intelligence technology and acquisition of artificial intelligence technology companies. Especially for the acquisition of some AI technology innovation companies, Google is a great show of interest.

　　1, enter the image recognition technology

Autofuss, a cross-cutting design company that covers the field of automatic animation and robotics, is the first AI technology company that Google has acquired, but after the successful acquisition of Boston Dynamics, the outside world is finally aware of Google's determination in the field of robotics. In addition, Google has acquired Autofoss's sister company, the mechanical and film Engineering Studio, Bot Dolly, which has its main business of providing automated robotics for Hollywood business films. After that, Google also acquired a photo analysis company Jetpac, the full march into the image recognition technology.

　　2. Deep learning with exerting force

In March 2013, Google acquired the Deep learning Enterprise Dnnresearch, founded by Jeffrey Hitton, a computer science professor at the University of Toronto in Canada, to truly establish Google's leading position in the field of deep learning in AI. At the beginning of 2014, Google also acquired a 400 million-dollar architecture for deep learning algorithms--deepmind,deepmind is a "cutting edge" company that works by creating powerful general-purpose learning algorithms to drive global business and game product development. Shortly thereafter, Google announced that it was expanding its AI program and had hired several leading academics and experts in the field to work with Oxford University to "speed up" the process.

　　3. Attacking driverless cars

The takeover of Holomni was designed to lay out Google's driverless cars, and before Google's takeover, Holomni Zeng issued an acceleration system to improve the robot's propulsion system. The acquisition of the US Industrial Perception Corporation (Industrial Perception) clears the visual barrier of Google's driverless cars, and the company's 3D Vision system (3D vision Systems) helps the robot to identify objects of different shapes to avoid accidental collisions.

　　4. Fully build the robot

Meka, a computer science and artificial Intelligence Laboratory stripped out of the Massachusetts Institute of MIT (MIT), announced it would join Google in a robotic revolution after it was acquired by Google. Subsequently, Google acquired the Meka and professional robotics research and development company Willow Garage co-founder of the humanoid robot and robotic arm manufacturer, Redwood Robot Company (Redwood Robotics). Google has always been enthusiastic about the company's takeover of Robotics, and then acquired Schaft, the Tokyo University's emergency rescue Robotics company.

　　5, competing smart home

Early last year, Google spent 3.2 billion of billions of dollars to buy smart home company Nest;6 Month, Google Nest spent $555 million on the acquisition of the cloud-based home surveillance company DROPCAM;10 month, and the acquisition of Smart Home Center control equipment company REVOLV, this continuous action shows Google's determination to enter the smart home.

　　Apple, still aspiring to create an ecological

As early as 2012, Apple announced a new voice recognition system: Siri, which opens the signal that Apple is officially marching into AI. Apple's watch is a sign of Apple's determination to take AI to the end. But for AI technology, Apple is not as obsessed with technology as Google, but like the iphone, creating an intelligent ecosystem around iOS, Apple has always been a leader in positioning itself.

　　1, HomeKit, to create a smart home platform

Unlike Apple's iphone, Apple has released its smart home platform, HomeKit, which is completely open to any vendor to work with. HomeKit to break the current various intelligent hardware manufacturers fragmented, user experience uneven market pattern, so that the smart home equipment manufacturers can interact at the iOS level, without the direct docking of these manufacturers, the current domestic Haier has joined the HomeKit platform.

　　2, Carplay, Smart car by force

The CarPlay is a new design designed to be the medium between the iphone and the car. After connecting to the iphone, users can access the phone's contacts, emails, notifications, map navigation, and music playback via the CarPlay system, which also enables Siri-based voice control to ensure that the driver is free to operate and drive safely. At present, CarPlay has cooperated with many automobile manufacturers, such as Ferrari, Mercedes-Benz, Volvo, BMW and Ford.

　　Baidu, promoting artificial intelligence popularization

Recently, Baidu Li in the two sessions on the proposal at the national level to build a "Chinese brain" project can be described as a result of the industry and foreign people's close attention and hot discussion. The project is an important research area for intelligent human-computer interaction, big data analysis and prediction, autonomous driving, intelligent medical diagnosis, intelligent unmanned aerial vehicle, military and civil robot technology, etc., supporting the ability of enterprises to build artificial intelligence basic resources and public service platform, open platform resources for different research fields, and efficiently docking social resources, So as to promote the whole nation's innovation.

Human development has experienced the era of agriculture, handicrafts, and then to modern industrial times, and we are now in the information age, but the future of mankind will certainly belong to the era of artificial intelligence. Liu Yan that Li's proposal is to stand in the long-term development of science and technology in the future, hope that the country can build a infrastructure, so that more enterprises at a lower cost to do more innovation, so that scientific research institutions, private companies, SOEs and even entrepreneurs can come to this platform, and promote the domestic artificial intelligence technology to the popularization. In recent years, Baidu has also been making efforts to this end.

　　1. Deep Learning leads the world

Deep learning is a complex model that is closest to the human brain and is currently the closest way to artificial intelligence. At present, whether Baidu, Google and other search giants, such as Apple, Microsoft and other international it giants, have poured into the research and development of deep learning technology, and Baidu has shown greater determination.

If Google's deep learning layout focuses on industrial layout, the means is mainly capital operation, then the idea of Baidu can be said to be "I-based", focusing on talent and technology input. Baidu has set up two deep learning labs in Beijing and Silicon Valley, and invited the former "father of Google Brains" to join Wunda, forming a team of luxury experts composed of Kaiyu, Zhang Yi, Xu Wei and other talents. From the current research and development output, Baidu in the deep study of the level of research and investment in the use of both sides in the forefront of the world, in the vanguard of the advantage position.

　　2, build Baidu Brain

The algorithm-based "Baidu Brain" is a representative of artificial intelligence, deep learning, it is Baidu Artificial intelligence products, fusion of deep learning algorithms and many other technologies, to achieve real-time learning and growth, it has 20 billion parameters, which constitute a large set of deep neural network.

Through the thinking of Baidu's brain, the relative error rate of speech recognition decreased by 20%-30%, the relative error rate of scanned text image generated Chinese text decreased by 30%, similar image search effect significantly more than similar products, through big data significantly improve the ads click-through, improve the quality of web search and sorting. At the same time, Baidu's brain can also be used to provide power for the application of big data through reasoning decision, and the social value appears initially. In addition to predicting the World Cup, you can also predict the college entrance examination, city tourism, tourism comfort, disease and so on.

　　3. Speech and image recognition technology

At present, the focus of Baidu on artificial intelligence in the field of speech recognition, natural language understanding, computer vision, and so on, these with Baidu search combined closely. Through the voice, image recognition technology, the use of mobile search threshold significantly reduced. Wunda introduced the Baidu speech recognition technology said, in the noisy environment, Baidu speech recognition performance, more than the Google Speech recognition API, Microsoft's "Bing Voice" technology, as well as Apple. On the right rate, Baidu can be 10% higher.

In the voice search, if the user hears a strange song, pick up the phone can tell what song, ordinary listeners will import Baidu music listen to the song, and the singer fans can be "direct" through the ticket service to import tickets to choose the seat and payment. In the image search, if the user takes a bunch of flowers that do not know the name, want to know more information, Baidu Brain will be shooting flowers, according to user attributes, through the Baidu brain to recommend students to the Baidu Encyclopedia, and for flower growers can be pushed to the nearby florist through the "direct number".

　　Microsoft misses the phone and doesn't allow it to lose artificial intelligence

For Microsoft, missing a smartphone has made it regret it, and he will never allow himself to miss artificial intelligence. Microsoft's Virtual Personal assistant service, launched at the 2014 build conference, is a harbinger of the company's development in the field of machine learning and artificial intelligence.

Recently, Microsoft's researchers in the academic paper said that Microsoft's latest deep learning system in the 2012 version of the imagenet image database, the error rate is only 4.94%, compared to 5.1% of the average person, this is the first time AI in the recognition of image error rate above the human level. It shows Microsoft's strength in deep learning technology. Microsoft's deep learning system, unlike the GPU-built components of Google's brain, uses more expensive CPUs that are powered by Microsoft's Azure cloud services.

In Microsoft's view, artificial intelligence and machine learning are the company's secret weapon to regain the pinnacle of technology. In addition, Microsoft is also studying "stealth user interface Technology", allowing users to "understand" their intentions before touching or giving voice commands.

　　IBM, determined to be at a higher level of AI

August 29, 2014, IBM formally announced the use of its Watson system, Watson is a supercomputer, composed of 90 IBM power 7 servers in parallel. Compared with Google, Microsoft's AI, it began to simulate human neurons from the hardware chip architecture, based on IBM's "DeepQA" technology development, can listen to human problems, in a vast array of data to find answers and natural language answers.

Now, this technology has been applied to areas such as healthcare, which integrates a large number of data analysis techniques, including deep learning. In addition, IBM has made a lot of efforts around "cognitive computing", and combined four universities for research cooperation, which is the core of deep learning.

　　Facebook to build smarter social platforms

In December 2013, Facebook established an AI lab. The lab is led by New York University professors, deep learning experts Jahn he, who want to build a digital assistant for Facebook. This digital assistant is based on deep learning techniques to identify your drunken photos and prevent you from uploading them to the social network. Their long-term goal is to develop technologies that closely analyze Facebook content outside of many photos, allowing AI to automate the process.

While other U.S. tech giants like Twitter bought deep learning startups Madbits, Yahoo bought deep Learning company Lookflow and image labeling company Iqengine, the Giants love AI technology all illustrate its bright future.

On the whole, the U.S. technology giants have entered the artificial intelligence technology research and development, and most of the domestic internet companies are just immersed in the internet and mobile internet era, only Baidu has been forward-looking forecasts to a new era. Be doing, to become a global technology innovation center, the domestic Internet technology companies and even the whole society must grasp the present opportunity, in artificial intelligence technology to pay more efforts.

　　Liu Yan, founder of the state of purchase, public number: Bangkeji

Ai's heat in science and technology through Li's proposal