您当前的位置是:  首页 > technology > Corporate communications > Abstracts >
当前位置:首页 > technology > Corporate communications > Abstracts > Huawei Zheng Yelai: From "Smart" to "Effective," Unleashing Unbounded Potential

Huawei Zheng Yelai: From "Smart" to "Effective," Unleashing Unbounded Potential

2018-05-11 11:20:20 Author: Source:CTI ForumComment:0  Click:


[China, Suzhou, May 10, 2018]
Huawei's vice president and president of Huawei Cloud BU Zheng Yelai shared Huawei's exploration and practice of artificial intelligence at the first global artificial intelligence product application expo. Let Xiao Bian take you back to review this self-proclaimed president who is neither interesting nor gentle:
The term artificial intelligence is already hot and hot. Just like big data six years ago, concepts have become more and more generalized. Huawei's layout in the field of artificial intelligence can be traced back to June 2012. The Noah's Ark Lab was established to conduct research on artificial intelligence algorithms, manage AI technical cooperation, identify AI main application scenarios, and demand management. Another very important task is to identify the entire artificial intelligence technology (enterprise) application scenario. Everyone knows that Huawei has invested in big data and artificial intelligence for many years. It is also what Mr. Ren repeatedly told us about. The key to artificial intelligence is application.
The key to artificial intelligence is application
Huawei itself is an electronic manufacturing company with a history of 30 years. We conducted Huawei's typical manufacturing, supply, manufacturing, warehousing, logistics, customs declaration, e-commerce marketing recommendation, real-time wind control and other manufacturing processes. practice.
We mainly apply in two scenarios:
The first scenario is to solve the automation problems in enterprise management that require a large amount of repetitive labor, reduce the overall management costs, avoid some repetitive work and low-level errors.
The second is to try to use AI to solve unknown or unsolved problems.
Industrial "+ Smart" Challenge
At the 2017 HC (Huawei Full Connect Conference), Huawei Cloud released EI Enterprise Intelligence, which is the scenario where artificial intelligence is applied to enterprises.
In the past years of practice, we have found that industrial "+ intelligence" has great challenges. After the appearance of artificial intelligence in the 50's, although there are no major changes in the nature of the algorithm, there are still two changes:
The first point is that over the past 20 years or so, due to the industry's digitization brought about by a large number of data training models, the model can check its algorithm to correct. The second change is due to the advancement of chip technology and the huge computing power of cloud computing.
According to the forecast of GIV2025, by 2025, the total number of global connections will reach 100 billion, and the vast amount of data generated by extensive connections will be widely distributed on the edges of people, objects, and organizations. The fuzzy training and application of artificial intelligence algorithms based on massive data, especially the smart industry in the industrial field will face many challenges.
  1. Mass data is widely distributed at the edges and terminals. It is very unrealistic to send all data back.
  2. Real-time processing, when everything is transmitted to the cloud for processing, it is difficult to guarantee real-time performance;
  3. Intelligent decision making, industrial-level intelligent decision making. Artificial intelligence will also present a lot of security problems, or it will be a problem of user privacy protection.
Huawei "cloud, edge, end, connection" full stack smart architecture
Huawei Cloud provides "cloud, edge, end, and connection" solutions. It has full stack capabilities from chips, algorithms, AI common services, and EI application development platform on the cloud. It can provide full-life-cycle management of cloud-edge models. Management collaboration, data collaboration, task collaboration, and security collaboration to achieve the purpose of cloud edge coordination and integration decision-making. Huawei is a connection originator and has a very deep accumulation in the field of Internet of Things. It has the ability to connect 800+ categories of 1.8 million industrial devices and is also the initiator of the ECC organization of the Edge Computing Alliance. IOT gateways, protocols, 5G, and enterprise business governance experience are the foundation and foundation for our cloud-side capabilities. To subvert the traditional IoT, Huawei Cloud provides intelligence and enablement capabilities.
Huawei Cloud's Smart Edge Platform
Intelligent Edge Fabric
Currently, Huawei's Intelligent Edge Fabric is on the line, and many corporate customers have entered the actual measurement phase. Through Huawei's own proven capabilities, real-time stream processing, intelligent quality inspection, predictive maintenance analysis, and video preprocessing are officially shared with companies. Satisfy customer's demands for remote control, data processing, analysis and decision-making, and intelligentization of edge computing resources, so that the data concerning the production and business activities of the enterprise can be analyzed and processed as locally as possible to provide users with real edge and cloud collaboration. Integrated services to achieve real-time decision-making of convergence.
Here are some typical examples from Huawei:
Intelligent edge platform realizes unattended production line
  • The inspection of PCB defects by means of a worker on duty on four machines, and the manual inspection of the defect rate of the welding PCB, this method not only makes the testing product inefficient, but also prone to error, leading to false detection, missing detection and other issues. According to statistics, the production line of Huawei's Songshan Lake factory takes about 5 minutes each time a worker checks.
  • The intelligent edge platform can achieve image capture, return, or direct model prediction. Unattended industrial production, improved defect discovery efficiency in the manufacturing process, and increased the production rate of the production line from 99.2% to 99.55%.
Using artificial intelligence to tame the "electric tiger"
  • In the automotive, machinery, electronics, power, mining, construction, chemical, textile and textile industries, a large number of machines are called air compressors, which use air kinetic energy to provide energy. The air compressor consumes a lot of energy. It is a well-known “electric tiger”. There are 1.2 million air compressors in the country, and the annual power consumption accounts for 6% of the national energy consumption.
  • Huawei Cloud EI Machine learns the air compressor program and generates an abnormal model of the motor and the model of the pipe loss. At the same time, the air pressure demand at the end and the loss of the pipe are calculated to adjust the optimal output pressure of the air compressor. . The current optimization result for the first phase is that each air compressor saves RMB 400 per day, which means that each unit can save about 2% of energy consumption. It is planned that the second phase can save energy consumption by 4%.
Mobilize tens of millions of Peugeot Citroen cars
With the capability of cloud-side + IoT connection, Peugeot Citroën Group (PSA) can provide cross-industry connectivity through 1000+ data collection points, vehicle speed, braking signals, mileage, and engine status, and will provide remote vehicle faults in the future. New services such as diagnostics, cloud software updates, and car sharing capabilities. Support the integration of car networking services with smart homes and smart cities to achieve integrated services for people, cars, life, and society. This is truly to provide a cloud, edge, and end-to-end collaboration and become a globally consistent service that can guarantee the security of data exchange. It is also Huawei's positive response to the forthcoming GDPR regulations that will take effect on May 25th. It advocates full respect for the integrity, authenticity, confidentiality of data, and the use of personal data throughout the entire human-based change.
Let ordinary camera learn to think
Based on Huawei Cloud EI's smart campus, we can achieve dynamic portrait services that support billions of seconds of second-level retrieval with an accuracy rate of more than 95% based on heterogeneous computing and cloud-side collaboration, and realize intelligent monitoring based on face recognition in the park. Cross-sector joint analysis can be realized. The key is to use an ordinary camera to achieve intelligent monitoring, truly do not change the end of the device. (Smart camera price is about 10 times that of a normal camera)
Huawei Cloud is a “Black Land” for Digital and Intelligent Enterprise Transformation
?
Huawei Cloud focuses on ICT infrastructure. Based on long-term practice in computing, storage, and networking, it has accumulated a large number of algorithms, chips, and hardware/software combination capabilities to create an optimal cloud infrastructure. At the same time, Huawei's ICT infrastructure technology is used to serve customers. In the online model, Huawei Cloud insists on respecting data sovereignty, does not use customer data for commercial realisation, and at the same time builds a win-win ecosystem with its partners. It is a “cloudy partner with technology, future, and trustworthiness”.
[Disclaimer] This article only represents the author's own opinion and has nothing to do with the CTI Forum. The CTI Forum maintains its neutrality in the presentation of statements and opinions, and does not provide any express or implied warranty for the accuracy, reliability or completeness of the contents contained therein. Readers are for reference only, and please bear full responsibility for yourself.
0

Topics