How AI enters the last mile of traditional industries

The concept of artificial intelligence has moved from theory to reality over the past few years. From software recommendation algorithms to artificial intelligence voice assistants, the seeds of AI are already sprouting in our lives. Today, the power of AI is slowly growing, entering traditional industries and moving towards scale.

AI out of the lab

After AI technology came into our field of vision through electronic products, we continue to explore more possibilities of artificial intelligence.In particular, we will explore the combination of artificial intelligence and traditional industries, hoping that AI can inject new energy into the development of traditional industries.

But the landing of artificial intelligence in traditional industries does not seem to be as easy as people think. Wu Enda, a well-known scholar in the field of artificial intelligence, raised a question in an article last year: Why is artificial intelligence that is “handy” applied in the internet industry “unacceptable” in traditional industries? Why is the application speed and scope of AI technology in traditional industries far less than in industries such as consumer Internet?

How AI enters the last mile of traditional industries

On the one hand, people’s understanding of artificial intelligence is still insufficient. The rational use of AI capabilities in traditional industries requires innovation and constant exploration. Only when a suitable application scenario appears, the related applications of artificial intelligence will emerge as the times require.

Unmanned delivery Robots, disinfection drones, and non-contact temperature detection equipment that have emerged with the outbreak of the epidemic are good examples. When the epidemic appeared, new application scenarios in traditional industries were created, and corresponding artificial intelligence applications came into being.

The potential of artificial intelligence as a novelty in traditional industries still needs to be tapped in practice.

On the other hand, artificial intelligence has been widely used in different scenarios, and traditional industries have the characteristics of large-scale and industrialization.This requires artificial intelligence to complete the “industrialization”, say goodbye to “small fights”, and truly move towards large-scale.At present, this process also faces challenges in many aspects such as talent pool, software and hardware ecology.

Towards large-scale, AI landing “Industrial revolution”

At present, AI still faces a difficult problem in the transformation to real large-scale applications: the current entry threshold for AI is not low, especially for traditional enterprises. At present, although many traditional enterprises are very interested in artificial intelligence, their accumulation of talents, capabilities and experience is relatively weak. This has resulted in the inability of these traditional companies to really implement ideas into practice.

Under such circumstances, although traditional companies are eager to try AI, and AI solution providers also want to expand their business scope as much as possible, there are information barriers on both sides.

On the one hand, there is a gap between AI solution providers and enterprises, and it is impossible to know exactly the needs of enterprises, and it is impossible to design optimization solutions according to specific needs. On the other hand, developers have no way of knowing the latest achievements of providers in development tools and development solutions.

In order to solve this problem, some AI solution providers have begun to work on innovation sharing and ecological openness. The “AI Practice Day” promoted by Intel is an attempt to communicate directly with AI solution providers and developers.

On March 15, 2022, at the online event “Intel Embraces Developers’ Software and Hardware Collaborative Innovation Ecosystem to Accelerate AI Landing”, Xia Lei, Intel’s chief engineer and China’s chief architect of artificial intelligence technology, introduced the relevant situation of the AI ​​practice day .

How AI enters the last mile of traditional industries

Xia Lei is one of the initiators of the AI ​​Practice Day. He said that the original intention of launching the AI ​​Practice Day two years ago was to hope that Intel could pass on the innovations it has invested a lot of effort to developers in an effective way. It is hoped that the distance from Intel to the market and customers can be shortened.

In the process, Intel has successfully promoted some artificial intelligence projects to land in traditional industries.

Intel has cooperated with Jinfeng Huineng in the energy field and used Intel’s AI technology to build a set of models for accurate forecasting of wind energy. Wind power generation has always had the problem of unstable power generation: when the wind is large, the power generation is large, and when the wind is small, the power generation is small.

This fluctuation of wind power output will not only generate a lot of curtailment and waste energy, but also affect the stability of the power grid. Therefore, wind power generation requires a relatively accurate prediction of the power output of the entire wind farm, which is conducive to the integration of wind power generation into the grid.

According to Xia Lei, Intel’s AI solution has made Jin Fenghuineng’s accuracy rate up to 80%, which is 20% higher than before. This means that 120 tons of carbon emissions can be reduced in the process of generating electricity every day. It will reduce the deforestation of 24,000 tons of trees a year.

Coincidentally, Intel’s cooperation with Weining in the medical field has also made artificial intelligence show its power in precision medicine. In the medical field, the dosage should be different according to each person’s physical condition. With the development of modern medicine, bone age testing is helping doctors to achieve precise drug delivery for each individual.

Intel and Weining have cooperated to build an AI solution for bone age detection based on Intel Xeon platform. It can reduce the time to process an image from 11 seconds to 6 seconds, nearly doubling the efficiency.

The nature of AI practice days has quietly changed as Intel and developer exchanges have made more and more achievements.

Google, Amazon, Baidu, Ali and other manufacturers have also joined the ranks of ecological sharing, using Intel’s AI practice day to communicate with the industry. The AI ​​Practice Day has changed from an event where Intel provides practical opportunities for the industry to a platform for ecological sharing in the industry.

Xia Lei also mentioned that Intel’s next step will be to segment the audience of the AI ​​practice day. In the face of developers who focus on different fields, Intel will classify them from the aspects of algorithm innovation and rapid deployment, and more accurately provide information of interest to different groups. At the same time, special sessions will be opened for developers in different fields to provide developers with more accurate solutions.

Today, AI is becoming large-scale and industrialized, and the increasingly perfect ecology and open communication environment are accelerating the implementation of AI. Maybe in the near future, AI can play its role in every corner of life as people expect.

The Links:   3HAC058665-001 JANCD-XEW01-1 IGBT

Published on 12/23/2022