In the second half of this year, Google’s external recruitment information shows that it is recruiting talents in the AI ​​direction in China. According to the recruitment information, Google recruits jobs in the field of machine learning, including machine learning researchers, machine learning technology directors, cloud machine learning product managers.
This move also makes the outside world guess whether Google is preparing to return to China. In 2010, Google announced its withdrawal from the Chinese mainland market. Before it exited the mainland market, Google once occupied 35.6% of the Chinese online search market. The rumors that Google returned to China in these years have never been interrupted, but they have not returned yet.
Since the release of the AI ​​First strategy, Google has become more determined on the road to artificial intelligence. Not only the Google Assistant smart assistant, but also released a variety of AI hardware including mobile phones, headphones and smart speakers at the autumn conference to build an AI ecosystem. Under the constant AI threat theory of Tesla CEO Musk, Google said that it focuses on AI's cutting-edge research and solving practical problems.
In recent years, Google has held APAC (Asia Pacific Regional Media Conference) every year. As a technology giant moving from Mobile First strategy to AI First, the focus of this media conference is naturally artificial intelligence, "Made with AI."
At the November 28 meeting, Google Brain head Jeff Dean said that Google's ultimate goal in the field of artificial intelligence is three: using artificial intelligence and machine learning to make Google's products more useful (Making products more useful); Helping businesses and external developers use artificial intelligence and machine learning for innovation (Helping others innovate); providing researchers with better tools to solve the major challenges facing humanity.
AI+ software + hardwareCurrently, machine learning is used in most of Google's products. Such as Google Photos Cloud Photo Gallery, which uses image recognition technology to provide face detection and automatic photo classification; Google Lens is based on image recognition and OCR technology to instantly identify and provide content related to smartphones; Google Maps You can get more detailed information about the area through the street and street view data, as well as the difficulty of parking; after receiving the mail, Gmail and Inbox will provide users with a reply (Smart Reply); automatic in YouTube Subtitles (Auto capTIons) automatically add captions to more than 1 billion videos through machine learning; Google Translator uses Neural Machine TranslaTIon.
Google Assistant is a voice assistant launched on May 19, 2016. The core is speech recognition. Pracious Gupta, director of engineering at Google Assistant, said the product is based on Google's experience in machine learning, natural language processing and search.
Among these products, Google Translate may be the most used by Chinese users. Jeff pointed out that the past translation system used a simpler statistical translation model consisting of 500,000 lines of code. In 2016, the Neural Network Machine Translation System (GNMT: Google Neural Machine TranslaTIon) was officially applied to Google Translate. Jeff said the system consists of only 500 lines of TensorFlow code. After using the new system, the translation accuracy has been greatly improved, "comparable to the results achieved in the past decade." Jeff mentioned that the most obvious improvement in translation results is the Japanese-English translation.
However, Google is not the first company to use a neural network machine translation system in translation. On the open day of Baidu's machine translation technology in 2016, Dr. Wu Hua, co-chairman of Baidu Technical Committee and technical director of Natural Language Processing Department, said that Baidu was the first to release the world's first neural network more than a year ago (2015). The Machine Translation System (NMT) overcomes the shortcomings of traditional methods of segmenting sentences into different segments for translation, making full use of context information to encode and decode sentences as a whole, resulting in a smoother translation.
Wu Hua said at the time that Google Translate was strong on statistical-based machine translation, but Baidu had to lead in neural network-based machine translation. In addition, Google Translate is centered on English, and Baidu translation is centered on Chinese.
An important part of building an ecosystem is to integrate the ingredients. Google is also working hard to bring hardware, software and AI together. At this fall conference, Google released nine hardware products, including smart phone Google Home Mini / Google Home Max, notebook Pixelbook, smartphone Pixel 2 and Pixel 2 XL, Google Pixel Buds headset, these new hardware are related to AI, The integration of Google's Smart Voice Assistant, Google Assistant, highlights Google's ambitions from software to hardware in the AI ​​space.
Among them, Google Home also has Voice Match, which can identify different voices through machine learning, allowing up to six users to connect to the same Google Home. Google Home Max also uses the AI ​​technology Smart Sound to automatically adjust the sound quality based on where you are. Google's first wireless Bluetooth headset, Pixel Buds, also provides easy access to Google Translate, using speech recognition and translation technology for real-time translation.
Unlike Huawei and iPhone X smartphones, which use dual cameras, Google's Pixel 2/2 XL combines machine learning and computational photography to analyze images and separate the subject from the background. Although only one camera is used, it also has a portrait mode function that softens the background when shooting portraits. Usually, this requires a multi-lens professional camera.
In addition to the use of AI for internal products, Google also offers enterprises and developers three innovative tools: TensorFlow, Cloud Machine Learning APIs and Tensor Processing Unit (TPU) computer chips.
Google launched the artificial intelligence system TensorFlow in 2015 and announced open source, after which TensorFlow became the most popular machine learning tool on the open source community GitHub. In addition to TensorFlow, other deep learning tools include Caffe, CNTK, Theano and more. In China, in September 2016, Baidu also announced that its deep learning open source platform PaddlePaddle is open on Github and Baidu brain platforms.
Faced with these competitions, Jeff responded at the media exchange meeting that each platform has its own advantages and disadvantages. This competition is good for different people. "Tensorflow open source software is based on the Apache 2.0 license. Everyone can do whatever they want, whether it's a big business or a startup. This may be one of the reasons for Tensorflow's success. We see a very healthy ecosystem. We have also learned a lot from other open source platforms, and constantly improve the Tensorflow platform to make this platform better."
Focus on realities and research issuesMachine learning is Google's focus in the field of artificial intelligence. Google believes that it is better to write programs that enable computers to learn how to become intelligent, rather than writing smart programs directly. However, as the machine becomes more intelligent, will the machine really be conscious and replace humans? Tesla CEO Elon Musk and the famous British physicist Hawking have warned the AI.
And Jeff believes that these concerns are too far away. "We can use many of these technologies when deploying secure AI systems. I think this may be the area we need to pay most attention to in the short term. Some of the current concerns are still too late. Early. We should now focus on solving the immediate problems.†This is also one of the three goals of Google AI mentioned above: solving humanity's big challenges.
Currently, Google is using machine learning to address issues such as healthcare, energy and environmental issues. For example, the Google Medical Imaging team product manager worked with hospitals in India, Thailand, and the United States to develop a tool to help diagnose eye diseases caused by diabetes through machine learning. In protecting birds, researcher Victor Anton collected 50,000 hours of audio and converted it into a spectrum, which was analyzed more quickly and efficiently by TensorFlow to identify the birdsong in the spectrum.
For the unemployment problem brought by AI, Jeff said that the technology development in the past two hundred years will encounter such problems. He is optimistic about this. "Every time there is a new development in technology, instead of human labor, we will There is a new, interesting area of ​​expertise to harness this technology. We will have new jobs that we may not be able to imagine at the moment. Who can think of social media as it would have been?
Don't worry about competitors continuing to recruit people in ChinaIn early November, Eric Emerson Schmidt, chairman of Google’s parent company, Alphabet and former Google CEO, said to Paul Sharay of the New American Security Center at the Artificial Intelligence and Global Security Summit: “I think we (US) will continue to maintain its leading position in the next five years, and then China will quickly catch up."
Google has long invested a lot of manpower and financial resources to develop artificial intelligence products. The industry has speculated that Google is using its own AI products as a springboard to re-enter China. From this point of view, this guess is not groundless.
Eric’s view stems from the “New Generation Artificial Intelligence Development Plan†issued by the Chinese government. The plan proposes that by 2020, the overall technology and application of artificial intelligence in China will be synchronized with the world’s advanced level; by 2025, the basic theory of artificial intelligence Achieving major breakthroughs, some technologies and applications have reached the world's leading level; by 2030, artificial intelligence theory, technology and applications have reached the world's leading level, becoming the world's major artificial intelligence innovation center.
Faced with the Sino-US AI competition, Jeff said that many companies around the world are interested in machine learning and AI. Many governments are very aware of the potential of AI, conducting research in stages and pragmatically, and building an ecosystem. “The Chinese government advocates AI and formulates policies; the US government may not be as organized as China, but we (the United States) also have a lot of research and corporate ecosystems in the company and academia.â€
Machine learning and AI not only affect computing science, but also all walks of life. Therefore, in the world, some governments and companies are recruiting relevant talents, which directly brings competition in the talent pool. On the other hand, it also involves the training of AI talents. Jeff said, "I believe that more people will have such skills over time, and they can solve many problems with this skill."
How does Google view competitors like Baidu? Jeff said that Google is primarily concerned with the next generation of cutting-edge research issues, hiring talent and providing them with the best computer hardware to solve practical and interesting problems. "We are not worried about competitors, we are concerned about our own research," he added, adding that Google will continue to recruit AI-related talents in Shanghai and Beijing.
At the 2016 Go Meeting in Wuzhen, Google first admitted to recruiting AI teams in China. Related recruitment positions, machine learning software engineer, machine learning technology supervisor, machine learning researcher, cloud machine learning product manager, etc. However, the number of recruits is not shown in the recruitment information. Google said that there is no specific statistics on the number of AI employees in China.
Desktop style 12v and 24v series can be used in many different electronics. EMC, LVD, FCC, RoHS are available in our company. OEM and ODM are available, samples are free for testing, all our products have 2 years warranty.
Our products built with input/output overvoltage protection, input/output overcurrent protection, over temperature protection, over power protection and short circuit protection. You can send more details of this product, so that we can offer best service to you!
Led Adapter, Lcd Adapter,Speaker Power Adapter,Lcd Power Supply
Shenzhen Waweis Technology Co., Ltd. , https://www.laptopsasdapter.com