In 2016, it was called the first year of artificial intelligence. In this year, the development of artificial intelligence intelligence ushered in a wave of climax. Governments have put artificial intelligence into government planning, raising the development of artificial intelligence to an unprecedented height. On December 2, Baidu released the 2016 annual search report, and the term “artificial intelligence†has been integrated with “blue thin mushroomsâ€. The popular words were selected together as one of the “Top Ten Keywords†in 2016. Let's take a look at the development of artificial intelligence in 2016 with AsiaOTT.
NO.1 Go Battle AlphaGo4: 1 victory over Li Shishi
Attention: ★★★★★
In March 2016, the artificial intelligence AlphaGo, which was developed by Google DeepMind, played against the Korean Go and the Korean national team Li Shishi, and finally won the global attention with a 4:1 victory over Li Shishi. Go is the most complicated professional game invented by humans, and artificial intelligence has already mastered quite a skill. Experts say that this moment is ten years earlier than expected.
After the game, the Korean Chess Academy awarded the AlphaGo honor to the nine-segment segment. AlphaGo's world ranking rose to second place shortly after the game. By July 17, the world's number one Go master Ke Jie had a poor performance in the cross-strait world championships, and points fell. AlphaGo surpassed Ke Jie, ranking first in the world.
AlphaGo utilizes the Monte Carlo tree search combined with two deep neural networks to evaluate the strengths and weaknesses of a large number of neutrons with a valuation network, and to select the singularity by the chess network. In this design, the computer can combine the long-term inference of the tree diagram, and can learn spontaneously and intuitively like the human brain to improve the chess.
Go has long been regarded as the "final fortress of human wisdom" in the chess game, and the victory of AlphaGo has made the public re-recognize the potential and prospects of artificial intelligence development.
NO.2 White House publishes artificial intelligence strategy report and planning
Attention: ★★★★★
On October 12, 2016, the White House released a strategic report entitled “Preparing for the Future of Artificial Intelligenceâ€, and also released the US National Artificial Intelligence Research and Development Strategy Plan.
The issuance of these two documents is the result of the Obama administration's efforts to promote the emerging technology industry, including the artificial intelligence industry, and to work closely with the Silicon Valley technology community in recent years. Quartz commented that the report and the strategic plan show that the White House is not eager to conduct large-scale supervision of artificial intelligence research and development, but will first develop application standards in vertical areas such as transportation and finance.
After Donald Trump was elected president of the United States, the White House followed up on December 20th to release the "Artificial Intelligence, Automation and Economy" strategy report, pointing out that artificial intelligence will have a positive impact on productivity growth, and should not be The potential threat of the existing order in the labor market, to curb the development of artificial intelligence, is that this is the last reminder to Trump before the Obama administration left.
In addition to the White House, many governments have also released relevant development strategies and plans this year. For example, Japan's "Rehabilitation Strategy 2016" listed artificial intelligence development as the top of the top ten revitalization strategy. The Chinese government's "Internet +" artificial intelligence three-year action implementation plan proposed to build a 100 billion yuan artificial intelligence market by 2018. .
NO.3 Five major technology giants Super AI Alliance
Attention: ★★★★
On September 21, 16th, the five technology giants of Alphabet (Google, DeepMind), IBM, Facebook, Amazon and Microsoft announced the formation of the Artificial Intelligence Alliance "Partnership on AI" in order to establish a benign competition and cooperation in the field of scientific research.
The AI ​​Alliance, with the slogan “Benefiting People and Society for Benefitsâ€, stated that its core mission is to conduct industry standards and normative discussions among giants, including exploring ways in which artificial intelligence can promote change in society. In addition, the development of ethics, inclusiveness and privacy of artificial intelligence is also an important research topic of this alliance.
DeepMind co-founder Mustafa Suleiman hopes that the alliance will have a strong agglomeration effect: "We hope that all parties affected by artificial intelligence can participate in this alliance."
At present, other giant companies in the field of artificial intelligence such as Apple, Twitter, Intel and Baidu have not yet joined the alliance. Despite news that Apple is "full of enthusiasm" for the project, it has not seen any concrete action taken by Apple.
NO.4 Tesla unmanned driving the owner to the hospital
Attention: ★★★★
According to media reports, an American man had a sudden pulmonary embolism on his way to Tesla, and finally arrived safely with the help of Model X's Autopilot autopilot function, which was driven by Tesla's driverless function. Saved a life.
It is reported that the American man lives in Missouri and is named Neely. He is 37 years old. In a 20-mile (32-kilometer) freeway drive to the hospital, Autopilot completed a fully automated drive and arrived at a freeway crossing near the hospital. Nellie then drove to the hospital by car.
In fact, today's cars have applied a lot of automation technology. When these automatic sensing technologies are combined with other technologies, they become a popular driverless mode. Therefore, automatic technology is not mysterious, and sometimes even plays a key role.
NO.5 The language understanding ability of Microsoft artificial intelligence devices beats humans
Attention: ★★★★
Artificial intelligence once again defeated humans in the vertical field. The theme of this time is discourse recognition. Although artificial intelligence assistants have begun to use speech recognition technology extensively, they are still not ideal in the face of more complex real-world applications.
But in October of the 16th, scientists from the Microsoft Research Institute of Artificial Intelligence published an academic report announcing that the shorthand error rate of its speech recognition system ASR was lower than that of a human professional typist for the first time, with an error rate of only 5.9%.
The Microsoft team used CNTK, an open source localized deep learning system that has undergone years of technology accumulation to train ASR. The research team used industry-standard Switchboard voice testing. At the same time, Microsoft also used speech recognition / shorthand test in other environments such as CallHome. In these tests, ASR's speech recognition results are also better than professional stenographers.
NO.6 Stanford University released a hundred years report on artificial intelligence research
Attention: ★★★★
In the fall of 2014, Stanford University launched a century-old research project on artificial intelligence. Two years later - in early September 2016 - a number of experts and scholars from Stanford University, including technology giants such as Google and Microsoft, famous universities such as Harvard University and MIT, and the Allied Artificial Intelligence Research Institute. The committee finally released their first report on the 100-year project, Artificial Intelligence and Life in 2030 (AI and Life in 2030), which includes recent developments in artificial intelligence and employment, environment, transportation, public safety, Medical, community involvement and the potential impact of the government.
After the report was released, it received great attention and, to a certain extent, played a guiding role in the development of artificial intelligence in the near future. The report writes: "The primary goal of the "Hundred Years Research" regular expert review is to provide a collection of collected and connected artificial intelligence and its impact as the field of artificial intelligence develops. These studies hope to provide direction and comprehensive assessment of professional inferences in research, development, and system design in the field of artificial intelligence, as well as in projects and policies that help ensure that systems are broadly beneficial to individuals and society. "
In addition to this report from Stanford University, there are many other research institutions and technology media that have published their own surveys or forecast reports. Here, of course, Amway has a series of monthly reports on the AI00: 100 Most Interested Companies in the World, which are released by the heart of the machine. The report is organized according to ten topics: voice and natural language, computer vision, chip and hardware. The 100 companies that are most worthy of attention in the field of artificial intelligence. The report will continue to maintain the frequency of monthly updates, and the latest version of the report can be viewed here.
No.7 Microsoft speech recognition technology reaches professional transcription level
Attention: ★★★
If human beings are a single, strong and powerful living body, then it must not be good for it today, because artificial intelligence not only makes it suffer from losses in Go, but also surpasses human beings in other fields. In a paper published by Microsoft researchers in October, "Achieving Human Parity in Conversa TIonal Speech RecogniTIon," a team of researchers and engineers in the Artificial Intelligence and Research Department reported that their speech recognition systems were comparable to professional speed recorders. Even lower word error rate (WER), reaching 5.9%! Just a month ago, Microsoft just achieved a historic 6.3%. This achievement and the pace of progress have even made some human transcriptionists feel the risk of unemployment.
The maturity of this technology has also laid the foundation for the development of voice assistants. At the 2016 AWS re:invent conference, Amazon released several development tools that would make it easier for developers to develop their voice assistant Alexa-related applications. According to Rohit Prasad, vice president of Amazon's Alexa division, Alexa now has more than 5,000 skills; Amazon Echo equipped with Alexa is now able to help users complete many daily tasks through voice, including schedule planning, music search. and many more. In addition to Amazon, companies such as Google, Microsoft, Samsung, and Apple have dedicated teams to help build their own voice assistant platform. Maybe in the near future, you no longer have to worry about no one listening to you, because the machine will always hear and understand you.
NO.8 Google WaveNet can synthesize more realistic vocals
Attention: ★★★
Google's DeepMind Research Lab unveiled WaveNet, the latest in computer synthesis, which mimics human voices and produces half the original audio compared to existing and human voices.
DeepMind said in a blog post: "The technology has passed the blind test of American English and Mandarin Chinese. The audience thinks that the voice generated by WaveNet sounds more natural than the voice produced by Google's existing voice technology. But DeppMind said that this is currently The technology has not yet been applied to any of Google's products, and the system requires powerful computing power and will not be applied to real-world scenarios in the near future.
DeepMind said that WaveNet is not yet ready for commercial use because it requires very powerful computing power and requires up to 16,000 acquisitions per second or more. And for each signal acquired, it is based on the sample of the previous signal, and analyzes and predicts what the next sound will look like. DeepMind researchers acknowledge that this is a very big challenge.
Despite this, technology companies are still paying great attention to the progress made by DeepMind in speech synthesis. Human-computer interaction has become an important way for people to communicate with their mobile phones and cars. Amazon, Apple, Microsoft, and Google are all investing heavily in the research and development of personal digital assistance, hoping that one day they will be able to naturally talk to humans and machines. Zhu Zhi, founder of Sesame Technology, told the First Financial News: "The development of smart devices is destined to become more and more important in the way that machines communicate with people. The more fluent it is, the more in-depth it is to communicate with others. It grows up like a child."
NO.9 De developed an artificial nervous system that can sense pain
Attention: ★★★
German scientists are developing an artificial nervous system to teach robots to perceive pain. It is said that this system not only allows the robot to react quickly to potential dangers, protects itself from harm, but also protects human companions working with the robot.
The researchers plan to develop a robotic nervous system based on "the results of human pain research." To test the system, the researchers installed a finger-like sensor on the robot arm to detect pressure and temperature.
Researchers at the University of Leibniz in Hannover, Germany, said that the system they developed allows robots to "detect and identify unexpected objects and disturbances, analyze possible hazards and react appropriately." Just as human neurons can transmit pain, the artificial nervous system can also transmit pain information, and the robot divides it into mild, moderate or severe pain.
NO.10 artificial intelligence predicts US election
Attention: ★★
The results of the US presidential election have surprised many people. However, when most of the nationwide opinion polls show that Hillary will win, the 12-year-old artificial intelligence system "Mogli" has long been convinced that the next president will be Trump. Previously, it also said that the two presidents were elected and the two presidential candidates.
MogAI is an artificial intelligence system developed by Indian Shakira in 2004. The name comes from the child raised by the wolf in Fantasy Forest. The "Wolf Baby" jungle is the entire cyberspace, which quickly handles the vast amount of information on search engines, social media and video sites.
Before the results of the US election were announced, Mogley’s speculation was contrary to most public opinion polls in the United States. It is optimistic about Republican candidate Donald Trump, but the polls mostly show that Democratic candidate Hillary will win. Mogley analyzed the emotions on social media through 20 million social media data points. In this way, you can truly understand the public opinion that traditional polls cannot understand.
High Current Terminal Blocks,Panel Terminal Block,Feed Through Terminal Block,Heavy Power Terminal Block
Sichuan Xinlian electronic science and technology Company , https://www.sztmlchs.com