This article is produced by NetEase Smart Studio (public number smartman 163). Focus on AI and read the next big era!
Netease Smart News March 18 News In 2017, artificial intelligence gained $12 billion in venture capital. And we are only just beginning to discover the usefulness of AI applications. Amazon recently launched a physical grocery store and successfully replaced cashiers and cashiers with technologies such as computer vision, sensors, and deep learning. In the area of ​​investment, news reports and major innovations, "artificial intelligence (AI)" has become a hot topic. But so far, does it really exist?
At the World Economic Forum, Dr. Li Kaifu said: "I think every entrepreneur wants to package his or her company into an artificial intelligence company. Every venture investor also wants to describe himself as an 'AI investor.'" He also observed that some of these AI bubbles may be shattered by the end of 2018, specifically referring to "those start-up companies that created a beautiful story that could not be realized, and those who were deceived by venture capital companies because they did not understand specific situation."
However, Kaifu Lee firmly believes that artificial intelligence will continue to make progress and will take away many employment opportunities from the workers. So what is the difference between legitimate, meritorious and deficient artificial intelligence and those fictional stories?
If you only parse through some stories related to artificial intelligence, then you will soon find that there is a big difference between different people's definition of it, and the boundaries between simulation intelligence and machine learning applications are also blurred. I communicated with experts in the field of artificial intelligence, trying to find a consensus, but this question has brought more problems. For example, when must the original definition of a term be followed? When is the definition of accuracy not so important? This is not obvious, and the big details about artificial intelligence do not care about these details. In addition, there is now a vested interest in such propaganda - precisely, $12 billion.
World-renowned thought leaders have openly debated the dangers of artificial intelligence. Facebook CEO Mark Zuckerberg suggested that opponents’ attempts to “promote these doomsday scenarios†are negative and irresponsible. On Twitter, business giant and OpenAI co-founder Elon Musk retorted that Zuckerberg's understanding of this issue is one-sided. In February of this year, Elon Musk once again had a similar exchange with Harvard professor Steven Pinker. Elon Musk stated on Twitter that Steven Pinker did not understand the difference between functional AI (narrow AI) and ordinary AI.
Given the concerns surrounding this technology, the public must clearly understand the differences between different levels of artificial intelligence so that they can truly assess potential threats and benefits.
Smart like people?
Erik Cambria, a specialist in natural language processing, told me: “Today no one actually does AI, but everyone says they are doing AI because it’s a cool and sexy buzzword, with a few years ago The data is the same."
Cambria mentioned that the term artificial intelligence came from human intelligence. "And even today there is nothing that is as intelligent as the stupid humans on earth. So, strictly speaking, nobody is doing AI because we don't know the simple fact that the human brain works," he said. .
He added that the term "artificial intelligence" usually refers to powerful data classification tools. These tools are impressive, but they are totally different from human cognition. In addition, Cambria has noticed that people claim that neural networks are part of the new wave of artificial intelligence. He thinks this is strange because this technology has existed fifty years ago.
However, technicians now do not need manual feature extraction. They also gained more powerful computing power. All these advances have been welcomed, but the statement that "machines mimic human complex cognitive processes" may not be accurate.
“Companies are only looking for skills to create what looks like intelligence, but this is not real intelligence. It is just a mirror of intelligence. These expert systems may be very good in a particular area and very stupid in other areas,†he said. Say.
This imitation of intelligence has inspired the imagination of the public, and systems in specific areas have already achieved their value in many industries. But these do not change the true definition of artificial intelligence.
Assisted intelligence, enhanced intelligence and autonomous intelligence
When it comes to the issue of scientific integrity, accurate definitions are important. In the 1974 opening ceremony of the California Institute of Technology, Richard Feynman said a word and later spread it widely: "The first principle is that you cannot deceive yourself - and you are the one who is most easily deceived." In the same article In his speech, Feynman also said: "You shouldn't deceive the layman when you speak as a scientist." He believes that scientists should bend back to show that they may be wrong. "If you think of yourself as a scientist, then you should explain what you are doing to the outside world - if they don't want to support you in this case, then this is their decision."
In the case of artificial intelligence, this may mean that professional scientists are obliged to clearly state that they are developing extremely powerful, controversial, profitable, and even dangerous tools that do not constitute any familiar or comprehensive sense. Smart.
The term "artificial intelligence" may have been over-hyped up causing confusion, but there have been some efforts to make it clear. A recent report by PricewaterhouseCoopers raised the distinction between "assist intelligence", "enhanced intelligence" and "independent intelligence." The representative of assistive intelligence is the GPS navigation program commonly used in today's automobiles. Enhancing intelligence can “make people and organizations do what they can't do.†Autonomous intelligence is “establishing autonomous machinesâ€, such as self-driving cars.
Roman Yampolskiy is an artificial intelligence security researcher who wrote the book "Artificial Superintelligence: A Futuristic Approach." I asked him if the different meanings of artificial intelligence would make it difficult for lawmakers trying to regulate artificial intelligence. Yampolskiy explained that “(artificial or natural) intelligence is a continuum, so these technologies may have potential problems. We often refer to Artificial Intelligence (AI) as Strong Artificial Intelligence (AGI) to avoid some confusion. In other cases, it is super intelligence. The AIs that we discuss today and are often used for business are very difficult to regulate anything, and the technology is no exception. The problem is not the term but the current system Complexity."
When asked if people should worry about artificial intelligence systems, Dr. Yampolskiy commented: “Since the abilities are continuous, so are the problems associated with each competency level.†He mentioned that there have been incidents that have occurred. Artificial intelligence products, and as the technology develops further, this effect may spread beyond privacy issues or technical unemployment. These concerns about the real-world effects of artificial intelligence may surpass people’s concerns about fixed behavior patterns of artificial intelligence. However, this issue also concerns the issue of honesty and deception.
"Artificial intelligence" has become a buzzword?
Finally, I turned my question to a company that is actively promoting the "AI virtual assistant." "Conversica's chief marketing officer, Carl Landers, acknowledges that there are multiple explanations for what AI is and what it is not.
He said: "My definition of AI is a technological innovation that helps solve business problems. I really don't want to talk theoretically, 'Can we think of machines as humans?' This is a good question, but What I am trying to solve is a practical business problem."
I asked him if "artificial intelligence" is a popular word for promoting and attracting customers. Landers said that it was true three years ago, but these effects have begun to weaken. Many companies now claim AI in their products, so it is not the key to distinguishing these products. However, there is a specific intention behind this word. Landers hopes to convey that what was previously impossible is now possible. "There are some new things you haven't seen before. You haven't heard of them before," he said.
According to EncoreLab founder Brian Decker, machine learning algorithms can only satisfy their pre-existing programming needs, rather than better understanding internal drivers. Therefore, he believes that artificial intelligence is a complete semantic argument.
Brian Decker said: "A marketing executive would claim that photodiode-controlled porch lights use artificial intelligence because 'it knows when the outside goes dark', and a good hardware engineer will point out that the porch lights throughout the process. The work system will only change based on pre-programmed systems."
While everyone must pay attention to specific content and potential significance, artificial intelligence-driven products have driven these debates by creating direct value for humans. Ultimately, humans are more concerned with value than with semantic differences. In an interview with Quartz, Li Kaifu revealed that the algorithmic trading system has allowed his private bank investment to return 8 times. "I no longer deal with humans," he said.
(From: SingularityHub Compilation: Netease Intelligence Engagement: Li Qing)
Pay attention to NetEase smart public number (smartman163), obtain the latest report of artificial intelligence industry.
Usb Docking Station,Usb Hub Docking Station,Docking Station With Hdmi,Usb C Hub Laptop Docking Station
Shenzhen GuanChen Electronics Co., Ltd. , https://www.gcneotech.com