Wednesday , December 8 2021

Artificial Intelligence "lower limit" to promote innovation. Experts: Be careful to security risks


Lot & # 39; open platforms & # 39; Artificial intelligence is greatly reduces the technical threshold, allowing artificial intelligence moves towards SMEs, to m & # 39; longer a "luxury" that can & # 39; only a person financially & # 39; robust . Experts believe that open platform is a new trend in the development of artificial intelligence technology, to promote the development of artificial intelligence at the top, but at the same time inevitably bring many security risks and to avoid.

The artificial intelligence "lower limit" to promote innovation

unmanned minibus & # 39; Baidu and Jinlong Bus newly made recently conducted public test test in & # 39; Beijing Haidian Park. This unmanned vehicle is equipped with & # 39; Apollo system Baidu, you can & # 39; based on maps & # 39; high precision and intelligent technology & # 39; sensing. Optimal Planning.

More than just Apollon, f & # 39; more than a year of opening the platform Baidu autopilots Apollo, there were 133 green developer partners and more than 11,000 worldwide.

"The main giants in the chain & # 39; upstream of the automotive industry and downstream including Daimler, Ford, Volvo, Intel, Microsoft, BYD, BAIC, Chery and others joined & # 39; ecology & # 39; Apollo. the development of artificial intelligence technology accumulated Baidu over the years accelerated the whole ecology innovation through & # 39; methods & # 39; open source, "said Zhang Yaqin, chairman a & # 39; Baidu.

Some time ago 5 World Conference on the Internet held in & # 39; Wuzhen, Zhejiang, Baidu Apollo AutoPilot Open Platform and Amazon Machine Learning Full Platform Managed, Li Aiming Platform & # 39; New Artificial Intelligence "Zhou Yi" , Sopet Industrial Industry platform, Intelligent Supply Chain Technology Jingdong number & # 39; & # 39 platforms; artificial intelligence, such as a service platform and Xiaomin Smart Home Open Platform, were released at the conference as an important component of & # 39; 15 'First Scientific Games Chief Technology and the Internet ".

As director of the Chinese Committee of Scientific Achievements Recommendations Technology and Internet World, the Hezhen He Academician of the Chinese Academy of Sciences spoke highly of the results of artificial intelligence platform. "Platform" has become the key technology achieving "the main event". "He Hezhen believes that similar results will create a platform for ecological environment Manage other SMEs to use the platform and technology for further development.

On AutoPilot Baidu open platform, developers can easily learn without using related technologies. The platform & # 39; new artificial intelligence "Zhou Yi" in China can & # 39; reduce cycle & # 39; & # 39 development; a & # 39; artificial intelligence equipment from three to five times and reduce the cost of & # 39; development of computer power five to ten times. On the Teaching of Amazon Machine Full Platform Management, trained developers can complete projects that require 10 to 20 PhD expert to complete in & # 39; 9 months, in just three to four & # 39; weeks.

"Almost all media & # 39; artificial intelligence today are aimed at & # 39 scenarios, specific application, vertical integration, low industrial synergy costs, of & # 39; development often reach several hundred & # 39; million yuan and c will cycle over several years. our industry needs b & # 39; urgent suitability as open and interoperable work. Platform & # 39; development & # 39; hardware and intelligent software to reduce development costs, " An said Wu Xiong, executive chairman and CEO & # 39; Anmou China.

"Everyone knows the importance of & # 39; artificial intelligence, but other than companies such as Google, Tencent, Baidu, m & # 39; there are many companies that can really develop advanced technology & # 39; artificial intelligence. So we open capabilities technology accumulated in the past. " Chairman and CEO Li Yanhong said.

"Wang Endong, academic of the Chinese Academy of Engineering and a leading scientist of Inspur Group, said the speed of innovation is the development of artificial intelligence. Basic Competitiveness.

The risks & # 39; multiple security are & # 39; caution

Many experts and industry entrepreneurs said that the development of artificial intelligence esperjeb many missing, and the current round of development is accelerating. There is consensus in the industry and should be fully acknowledged the problems & # 39; security brought about by artificial intelligence will develop well. "Evolution dynamics."

The artificial intelligence "cast thresholds" spawn more hackers, viruses and more cycles & # 39; shorter attack. Hehe said that in the network security field, the artificial intelligence popularity will help identify risks to network security, and will reduce the hackers threshold, so that more people learn hackers, more and more variants of & # 39; viruses and the cycle will become more and more. Short.

"With the advancement of artificial intelligence, the characteristics of the" sword b & # 39; dual angle "artificial intelligence will intensify." Hehe said.

Zhou Hongyi, chairman and CEO & # 39; 360 Group, said that today's cyber attacks were intelligent, automated and extremely fast. "The network security monitoring often finds some attacks on the network. After that discovered and analyzed the research, the solution has not been implemented.

The sensors, date training and open source software used by artificial intelligence to security risks. The White Paper on Artificial Intelligence released by the China Institute for Information and Communication recently showed that research and development & # 39; products and applications & # 39; Artificial intelligence is mainly based on frames and components & # 39; & # 39 learning; released by artificial intelligence technology giants like Google, Microsoft, Amazon, Facebook, Baidu, etc. The components m & # 39; have strict management test and safety certification, and vulnerabilities and back are vulnerable to malicious exploitation.

"There are way & # 39; ' Data poisoning ", try mixing some abnormal data in the training data, and the system & # 39; artificial intelligence trained make mistakes. If the software is written by someone, there must be gaps, such as 1000 line a & # 39; code. there will be four gaps in it. "Zhou Hongyi said that due to the inevitable gaps, the attack and defense of the current network are highly irregular. "Strictly speaking, the discovery of & # 39; traditional intrusion, a firewall, a & # 39; software became free anti virus technologies power in defense against such attacks."

"At present, the ecology of artificial intelligence has not been fully established, especially in a relatively fragmented state in all aspects of the industry, which means that hackers can take advantage of it." Jiang Wei, senior researcher and expert in artificial intelligence, Blue Shield, The security risks include denial of service, disclose & # 39; information, system hijacking, etc., Which could lead to sentences lost, bad judgments, grabbing system, kidnapping or designing & # 39; smart devices in & # 39; & # 39 tool, zombie attack.

Need to build a system & # 39; & # 39 management and; Risk governance

"The artificial intelligence has multidisciplinary features and very complex. We have a plan and we & # 39; forward and focus on improving originality." Experts like Hejun He said that while further enhancing research and development & # 39; artificial intelligence and the exercise is necessary to prejudice the challenge & # 39; artificial intelligence, and "strengthen and avoid" weaknesses in & # 39; & # 39 terms; technology and law to avoid risks.

Firstly, it is necessary that exceeded the key technologies of artificial intelligence. Endong Wang believes that enterprises & # 39; artificial intelligence and domestic institutions & # 39; Scientific research should have a clear understanding. We still have obvious deficiencies in the basic level, and we should strengthen basic research and contribution. Hezhen He said the application of artificial intelligence has just begun, but is still in its infancy, and currently m & # 39; there is no way to explain the causal relationship. More hope placed on the advancement of artificial intelligence.

Second, data silos should be opened to improve the security protection. "The network security is part of the whole process & # 39; ' let the whole body. "The big problem we face is that every company is" individual fighter ", the data is open and m & # 39; no overall vision or national, and process & # 39; real cyber attacks do not appear." Said Zhou Hongjun.

Third, establish and improve the laws, regulations and institutional systems that guarantee the development b & # 39; his health artificial intelligence. "The development of science and technology surpassed many existing social regulations and ethical standards, and it is extremely urgent to establish order & # 39; new network." Li Cheng, director of the Center of & # 39; John Thornton China & # 39; Brookings Institution of American thinktank, believes that to iç- society, think tanks and enterprises should establish law and rules unified network.

Shen Xiangyang, vice president of Microsoft Global, said that Microsoft proposed "six principles" of & # 39; "justice, tolerance, transparency, accountability, reliability and security, privacy and confidentiality" to guide the development of & # 39; artificial intelligence. Manage and direct the development of & # 39; & # 39 products; artificial intelligence "responsible".

Moreover, we must strengthen international exchanges to explore a & # 39; response to governance strategies. Bruce McConnor, deputy global president of the Eastern Institute and the Western United States, suggested that all parts of the world to prevent the risk of building a wall safe by building rules & # 39; official misconduct and unofficially about the cyberspace. (Reporter Zhu Han)

Source link