51CTO Community Thousands of Developers Interacted: Five Challenges Need to Be Unlocked Before AI Can Become a Reality

[51CTO.com original article] Which technology is the hottest in 2018? It must be AI. Recently, 51CTO organized IT elites from the developer community to have a heated discussion on topics such as the value of AI, AI applications, AI implementation pain points, and future development trends of AI. After the discussion of thousands of developers, 51CTO extracted some very representative views, which may be used as a reference for more developers interested in AI and spark more ideas.

[[244374]]

AI is a capability that will cover all industries

[[244375]]

When talking about the changes that AI has brought to existing industries, the vast majority of developers recognize the value of AI. Some developers said that at this stage, the industries that are more satisfied with AI applications are mainly concentrated in e-commerce shopping, medical care, and translation industries. Many developers also believe that artificial intelligence is a capability and a technology that can be applied in all industries. It can be applied to various scenarios and industries. Although the industry innovation is still most active in software and the Internet, it will gradually extend to more industries in the future, such as combining with manufacturing to do intelligent manufacturing, etc.

The views of Huawei Cloud BU President Zheng Yelai are consistent with the latter. At the 2018 China International Big Data Industry Expo, Zheng Yelai pointed out that with the rapid development of the digital economy, AI is increasingly being applied in various fields. As the driving force of a new round of industrial revolution, AI will continue to affect technology, products, industries, formats, and models, and ultimately bring about major changes in the economic structure.

@苏州--小刘—java: It is well applied in big data analysis. I personally feel that the recommendations of Google and Amazon are closer to my personal preferences. Especially for Kindle books, the recommendations are very accurate.

@开发-北京-烨: AI is widely used in the medical industry. Deep learning and image recognition technology can be used to diagnose the condition and provide treatment plans. In the field of pharmaceutical retail, face recognition technology can be used to build smart pharmacies to provide more accurate services.

Python-Karl-Beijing: I think the most common AI application is speech recognition. Direct English conversion solves the hearing problem, but it will take time to accurately match real-person semantics in complex scenarios.

How many steps does it take for AI to become a reality? Huawei's "Inclusive AI" solves the implementation problems

When talking about the pain points of AI in application, netizens believe that five problems are most prominent, namely, the mismatch between algorithm models and industry attributes, the shortage of AI talents, the difficulty of use, the high cost for enterprises to purchase AI equipment themselves, and the inconsistent understanding of AI between senior executives and actual developers.

[[244376]]

51CTO believes that the price of AI is indeed the most criticized. After all, the cost of training face recognition, comprehensive traffic management, autonomous driving, and model training is very high. The difficulty of using AI is also easy to understand. Most companies lack a unified development framework and cannot adapt to a variety of application scenarios from training to reasoning, from public cloud to private cloud, edge, and terminal. The workload of development, tuning, and deployment is huge. Of course, this also includes the problem of talent shortage. In fact, there is another difficulty, that is, the difficulty of purchasing AI. For example, the GPU supply cycle is long and the supply is limited. It is not "buy it if you want to buy it."

[[244377]]

In order to solve the pain points of netizens, Huawei proposed the concept of "Inclusive AI", which is to enable ordinary people to use AI well, afford AI, and use AI with peace of mind.

From the perspective of affordability, Huawei reduces the overall cost through the public cloud, develops the public cloud as a public module, simplifies complex problems, and enables ordinary software engineers to use AI.

From the perspective of user safety, Huawei Cloud adheres to the "Three Nos" principle, namely "no application development, no data access, and no equity investment". In addition, Huawei Cloud is the only cloud service platform in China that has passed the PCI-DSS security certification for all platforms, all nodes, and all services, making it more trustworthy.

From the perspective of good use, Huawei Cloud has strong technical support. Huawei Cloud's "AI Intelligent Platform New Graph Computing Technology (EYWA)" won the 2018 Leading Scientific and Technological Achievement Award - New Technology and Black Technology Awards.

In short, Huawei hopes to provide enterprises with a simple, reliable, and powerful enterprise intelligent ecosystem. The concept of "inclusive AI" has also been widely recognized by thousands of developers, and everyone agrees that such AI services are greatly needed.

@Container-Mario-Shenzhen: If it is not only cheap but also easy to use, it will definitely be popular. AI is just in its infancy and there is still a long way to go. If we start to cultivate brand trust now, we will definitely see Huawei Cloud AI in more fields in the future.

Test-Feihu-Beijing: Well, the threshold for artificial intelligence technology is high now, and there are too few corresponding talents. If Huawei really has such a tool, it will play a big role in the implementation of artificial intelligence.

AI will only be viable if it returns to products and scenarios.

When discussing Huawei's "Inclusive AI", some netizens also asked why Huawei rarely promotes the concept of AI. In fact, this is a misunderstanding. Huawei believes that AI should not be a publicity gimmick, but should empower the implementation of AI applications in enterprises and take on the mission of doubling application efficiency. More importantly, the AI ​​competition between enterprises will ultimately focus on whether it can benefit the people and the economy. Only by returning to the scenarios and products, and implementing reusable technical models and combining them with actual application scenarios and real needs, can people clearly understand and feel the value and significance of AI. Therefore, Huawei has been using the name of Huawei Cloud EI (Enterprise Intelligent Service) for publicity.

In 51CTO's view, Huawei Cloud's "high-cost but not expensive" inclusive AI has allowed artificial intelligence to get out of the "conceptual fog" ahead of time, and has also enabled partners to use AI to complete an efficiency revolution. It has also eliminated customers' concerns about AI applications with a "+AI" attitude, which is also the most natural and effective way to implement AI.

*** Let's insert an advertisement: Since AI is so important and everyone is paying so much attention to AI, why not get up close and personal with it! From October 10th to 12th, Huawei will hold the 2018 Huawei Connect Conference at the Shanghai World Expo Center. The world's top AI experts will be present, with more than 25,000 participants and a 34,000 square meter exhibition hall. All topics related to AI will be discussed in depth! Why not sign up? Click here to sign up: Sign up for the 2018 Huawei Connect Conference

[[244378]]

@Product-Nebula-Anhui: AI will be everywhere in the future, and manufacturing has the greatest potential. However, it will take a long time for AI to truly achieve high intelligence.

@Management-Moon Shadow-Changsha: I am optimistic about the intelligent transformation of industries such as retail and logistics.

@Development-Xunzi-Wuhan: Most useful in game development!

[51CTO original article, please indicate the original author and source as 51CTO.com when reprinting on partner sites]