Wuhan University of Technology Speeds Training

Wuhan University of Technology accelerates the cultivation and training of practical “new engineering” talents.

At a glance:

  • The Wuhan University of Technology is at the forefront of the integration of industry and education initiatives.

  • An upgraded data center powered by Intel edge-to-cloud technology is helping the university improve the training roadmap for integrating industry and education creates a more optimized and practical experimental environment for the university, and facilitates the dynamic adjustment of teaching and training programs to adapt to the ever-changing market demand for talent.

author-image

By

The idea of “new engineering” has emerged driven by a new wave of the industrial revolution. Compared with traditional engineering, the concept of new engineering is reflected in the educational philosophy of an interdisciplinary system: integration of research and education with collaborative education, which pays more attention to the close combination of AI, information and communication, electronic control, software design, and other new technologies merged with traditional industrial technologies. By building a talent training system that meets the requirements of the developing times and industry upgrades, this curriculum aims to cultivate a high-quality mix of talent who understand theories, strengthen the practice of engineering, and focus on innovation.

As the research base for New Engineering at the Ministry of Education, Wuhan University of Technology has been at the forefront of the integration of industry and education initiatives, which continue to build a project-based and challenging curriculum through an innovative infrastructure. By increasing the proportion and depth of experimental practice, the university is committed to laying a solid foundation for the development and growth of new engineering talent. In particular, after the launch of the School of Computer and Artificial Intelligence in 2019, a new curriculum of Artificial Intelligence+ was introduced to facilitate the cultivation of innovative and practical talents and further drive the integration of industry and education in the field of artificial intelligence.

"We hope to realize the combination of engineering, software, and hardware in teaching and achieve in-depth integration through the amassment of academia at the university with the practice of industrial applications so that the academic training can better meet the requirements of real-world scenarios and future industrial needs. The cultivation of AI+ talents need to match the development of the industry. Only by integrating with the needs of businesses can the far-reaching competitiveness of talents in the job market be improved. At the same time, they will create greater value for the innovation and development of the industry."—Wuhan University of Technology Director of Intelligent Science and Technology, Teaching and Research Department

Challenge: Under the Dual Pressures of Practical Training and Research, Infrastructure Needs an Upgrade

The continuous increase of teaching and research tasks at the School of Artificial Intelligence and the School of Computer Science and Engineering, both at the Wuhan University of Technology, combined with the continuous upgrading of industrial demands, places new requirements on the schools for the quantity of research equipment, software and hardware capabilities, and case resources. In the process of teaching experiments, the school also carries out practical training and application development of enterprise-level AI projects and concludes a number of research achievements, ranging from important monographs, articles, and patents to software copyrights. Particularly:

  1. The demand for industrial talent is strong and the number of students continues to grow, which increases the demand for software and hardware facilities.
  2. With a number of research projects undertaken by the school entering a crucial stage, the demand for terminal performance and cloud computing power is increasing, and there is also a need for the flexible allocation of digital resources, including the power of the CPU. Therefore, an integrated solution is particularly critical for the development of the school at this time.
  3. The new training model adopted by the school of encouraging learning by competition and combining competition with learning poses new challenges to the multi-scenario transformation of AI innovation laboratories and the scalability of terminals.
  4. When driving the development of AI education, it is only by integrating experimental training with industrial needs that we can truly cultivate well-rounded AI professionals with practical know-how.

With this background in mind, Intel has partnered with Wuhan InnoVison Information Technology Co., Ltd., a typical new-generation information technology manufacturer, to set up new AI innovation laboratories and training units for the Wuhan University of Technology. Thanks to the upgraded data center, the proportion of teaching tasks covered by the AI laboratory will increase from the current 50% to 70%, helping the school improve the training roadmap to integrate industry and education. In addition, the integrated solution of Terminal-Edge-Cloud keeps pace with the times, creates a more optimized and practical experimental environment for the school, and facilitates the dynamic adjustment of teaching and training programs to adapt to the ever-changing market demand for talent. The final deliverables include:

Artificial Intelligence Innovation Laboratories: The Terminal-Edge-Cloud collaborative solution based on Intel® Core™ and Intel® Xeon® Scalable processors not only provides powerful CPUs for the Laboratories but also supports AI application development by faculty members and students in different scenarios through intelligent scheduling and supply of computing power resources, such as AI reasoning, real-life training tasks, and heterogeneous integration. In addition, AINUC as a terminal computing power platform has created a series of built-in computing learning and development kits including a number of professional core courses in the fields of AI and big data; such as autonomous driving, industrial visual defect detection, and natural language processing.

Figure 1: View of laboratory.

Artificial Intelligence Training Units: The Autonomous Driving Training Program and the Industrial Defect Inspection Training Unit, which are the key foci in this solution, use Intel Core processors as the computing base to provide excellent data processing capability for innovative AI CPUs after front-end data collection. Specifically, the Industrial Defect Inspection Training Unit can detect industrial product defects via image data collection as well as identify defect training models and products. The Autonomous Driving Training Program enables comprehensive training practices such as automatic pedestrian recognition, automatic sign recognition, automatic lane recognition, obstacle avoidance functions, and automated road driving tests.

Figure 2: Workspace for practical training.

Solution: To Create an AI Teaching and Training System with an Integrated Terminal-Edge-Cloud Solution That Explores a New Ecosystem of Intelligent Education

Faced with the demand of building innovative AI laboratories put forward by Wuhan University of Technology, Intel, and its partner InnoVison efficiently connect the four elements of AI development (namely: computing power, algorithm, data, and scenario) based on hardware terminals, thereby creating a full-stack solution integrating Terminal-Edge-Cloud. Relying on the super computing power brought on by Intel Core processors, it overcomes the challenges of data collection, algorithm optimization, and scenario application in order to improve the learning skills and practices of students and upgrade the quality of learning and the application efficacy with more practical and scenario-based platform technology.

Figure 3: Intel integrated end-to-end cloud solution.

First of all, the clusters in the CPU of the AI development experimental platform deployed in the data center of the school can quickly store, interact with, and manage experimental data as well as provide a scientific research interface. Moreover, since the software of the learning experiment platform is based on container clusters and the infrastructure adopts a distributed storage solution, the application layer, as well as the infrastructure layer of the learning experiment platform, is capable of higher scalability. In other words, with higher scalability, should the school have further demand for higher computing power in the future, it can expand seamlessly without overthrowing the existing architecture.

Figure 4: Diagram of the on-campus private Terminal-Cloud-Edge solution.

Second, the full-stack deployment task enables the Terminal-Edge-Cloud solution to cover the whole learning process. Whether it is the preparation of lessons, teaching evaluations, research tasks of the teaching faculty, or the experimental/learning process of the students, the AIReady platform can provide technical support throughout the process while monitoring the students’ operations, analyzing learning situations, and evaluating students’ knowledge levels in real-time through a complete system operation model of teaching, learning, practice, development, and evaluation, thus providing strong support for both the skill training of students and the daily work of the faculty members.

Figure 5: AIReady management platform.

Technical Analysis: Combining Both the Software and Hardware, Different Product Levels Help Integrate Industry and Education

At the product level, the integrated Terminal-Edge-Cloud solution jointly launched by Intel and Wuhan InnoVison includes the AI research and development platform, management node platform, computing node platform, AI learning kit, autonomous driving training, industrial defect detection experimental equipment, etc. All of which can improve the development of software and hardware infrastructure such as the curriculum system, learning platform, and experimental training environment in a well-rounded way to help students master the basics of AI knowledge, enhance their practical and innovative capabilities, and build a closed loop of innovative AI talent training processes.

Empowering the CPU: The New Intel® Core™ Processor, the AI Education Rally Amplified by Super Computing Force

The power of the CPU is the key to further commercialization and smarter AI. It is also the core competence of Intel. Particularly, under the support of Intel Core Processor, the value dimension of the CPU has been further enhanced from the core driving force of the software and hardware ecosystem to the basic productivity of digital intelligence education. Therefore, in the field of AI education, in order to improve students’ professional skills and comprehensive application abilities, it is necessary to break away from the basic upgrade of hardware systems.

First of all, the 11th generation Intel Core processor, which integrates super CPU power and cutting-edge manufacturing technologies, is the “brain” of the innovative AI laboratories. In the future, the progressive compatibility of the Intel® x86 architecture (capable of extending to the latest 12th generation Intel Core series) will enhance the CPU of the AI innovation laboratories at the Wuhan University of Technology, which always strives to keep up with the latest technologies. This derives from Intel’s continuous breakthroughs and innovations in core processor technology. Taking the new 12th generation Intel Core processor as an example, its revolutionary hybrid architecture can ensure the energy-efficient collaboration of P-Core and E-Core, which are competent for multi-task parallelism, adapt to various application requirements, and fully meet the learning and training needs of faculty members and students in the AI world.

It is worth mentioning that Intel has also expanded the cache of the 12th generation Intel Core processor. In addition to increasing the L2 cache per core of E-Core and P-Core, Intel has also enhanced and expanded the shared L3 smart cache. With the different numbers of cores, the next-generation computing unit based on the 12th generation Intel Core processor will increase the L3 smart cache to a maximum of 24 MB, which will effectively upgrade the volume of memory data and reduce latency.

In addition, the 12th generation Intel Core Processor also integrates a dedicated AI computing unit with more powerful performance. Thanks to Intel® Arc™ graphics, it creates the 3I platform, establishes the Deeplink mode, coordinates with Intel® Iris® Xe graphics, and Intel Arc graphics so that faculty members and students will enjoy better performance, and speeds up AI reasoning when running applications containing AI algorithms without worrying about whether they are incompatible with the existing IT environment of the school. It can also be expandable and compatible with future system updates.

Hardware Boost: Intel® Next Unit of Computing (NUC), a Small, Flexible, and Powerful Computing Power Box

If Intel® Core™ mobile processor is the heart of unleashing computing power at the edge, then Intel® NUC mini PCs turbocharge the mission of computing power box. It not only integrates a variety of heterogeneous computing platforms such as the general-purpose processors, GPU, and VPU but also provides a large number of pretrained model libraries and rich sample codes and deploys Intel NUC mini PCs in training units to carry out data training in various scenarios. For example, if the Intel NUC mini PC is deployed on a self-driving model car, a series of computing power boost such as data collection can be carried out through a set of built-in learning and AI development kits so that students can build self-driving projects in a faster way.

Figure 6: Intel’s next-generation CPU.

More importantly, the Intel NUC mini PC is small in size and is largely portable so it does not take up a lot of desktop space and can even be attached to the back of a monitor, striking the perfect balance of practicality. In addition to providing reliable and stable computing support, the Intel NUC mini PC can greatly enhance the presentation experience for both the teaching and the faculty-student interaction such that the teaching quality of all computerized courses will be significantly improved.

Figure 7: Model of an autonomous driving car.

As an epoch-making edge computing product, the localized deployment of the Intel NUC mini PC can effectively eliminate system failure and unnecessary hidden dangers caused by network delay or crash while efficiently balancing cloud and local CPU resources. Remote management can also be carried out through the Intel NUC mini PC powered by the Intel vPro® platform, thus reducing maintenance costs. The highly customizable features of the Intel NUC mini PC also enable it to meet most of the customized needs of faculty members and students and can be easily deployed. According to different teaching scenarios, customized performance expansion can also be fulfilled, and more devices can be connected to achieve the best teaching experience.

 

  • Integrated AI practice – Integrating architectural design, it provides comprehensive training on mechanically controlled sorting, AI depth vision algorithm, communication, and signal acquisition technologies. Students can fully experience the entire process of data acquisition, training, and reasoning.
  • In-depth integration – Integrating cutting-edge industrial detection algorithms as well as comprehensive applications such as PLC technology, mechanical technology, and automation technology.
  • Integrated design – The integrated design simplifies the complexities of training, integrates PLC, intelligent edge terminal AINUC, and high-speed industrial cameras.
  • Practical training courses – Includes industrial defect detection training courses.

Figure 8: Training platform for detecting industrial defects.

The Power of Cloud: How Intel Xeon Scalable Processors Help the School of Data Collection Harness the Value of AI

At present, the teachings of AI at the Wuhan University of Technology can be carried out not only through local terminals but also through the remote interactive environment of cloud + edge, powered by the Intel Xeon processor. Specifically, through excellent microarchitecture design, more cores, and larger memory support, the Intel Xeon processor enables higher vocational faculty members and students to experience consistent training and reasoning in AI teachings and scientific research.

Intel’s scalable processors are optimized for many workload types and performance levels and can be equipped with any of Intel’s open architectures, while the built-in advanced security function and AI acceleration can release the most of cloud computing power, especially the feature of in-depth learning acceleration, which can significantly improve the reasoning performance and optimize the in-depth learning workloads.

The Power of the Ecosystem: The Open-Source Ecosystem of Digital Technologies Brings AI Teaching Within Reach

In order to drive AI education from laboratory to industry, Intel not only provides a large amount of hardware computing infrastructure needed by edge networks but also offers completely definable software to better meet teaching needs with a complete solution that integrates software and hardware requirements. Among them, the aid of the OpenVINO™ toolkit and mainstream in-depth learning framework that is optimized for Intel’s architecture greatly improves the efficiency of AI training and reasoning for faculty members and students.

In the future, through the support of software and hardware capabilities and rich product portfolio, Wuhan University of Technology will have the hierarchical and progressive capabilities of an AI system training. AI and its learnings can be fully integrated throughout all the processes. Faculty members and students can build and train within the AI environment locally based on OpenVINO AI development kits preinstalled with AINUC and hundreds of OpenVINO pre-trained models, as well as accelerate model reasoning and inference through teaching experiments combined with OpenVINO as a key teaching environment in autonomous driving and other practical training experiments.

Figure 9: AI framework supported by OpenVINO and Intel system calls.

The Outlook: Intel Joins Hands with Industry Partners to Accelerate the Implementation of Production, Education, and Research, Bringing More Possibilities in the Field of Education Informatization

It is the original intention of Wuhan University of Technology to introduce AI innovation laboratories with innovative educational models to spark students’ talents, cultivate well-rounded standards, and better adapt to and create the future. With the rising demand for the integration of industry and education, Intel will continue to work with more partners to build a complete education ecosystem together with Wuhan University of Technology:

  1. By taking full advantage of enterprises to assist faculty members in applying for national or provincial research projects and support the faculty’s academic research and innovation efforts in the field of AI. At the same time building an incubation platform for innovative and entrepreneurial students, guided by the faculty’s research directions, and facilitating and promoting the transformation of research findings and achievements, whilst enhancing and innovating the students’ employment potentials.
  2. Providing a training platform for undergraduates in the School of Artificial Intelligence as well as internships and employment services for students through the Intelligent Pilot Factory built by Intel and its partner in Jiangsu. After passing the examination, students are awarded internships and employment opportunities in the factory.
  3. Assisting the school to build the curriculum system of AI innovation laboratories and supporting the development of intelligent science and technology majors and artificial intelligence majors at the Wuhan University of Technology.

Figure 10: The Intelligent Pilot Factory built by Intel and its partner in Jiangsu.

Conclusion: Intel Has Always Been Committed to Building a “Bridge” for the Integration of Industry and Education Through Innovative Technologies

With the launch of the AI innovation laboratories, the AI education model at the Wuhan University of Technology will continue to target the industrial end and accelerate deployment of AI technologies in various industries with the help of the integrated Terminal-Edge-Cloud Solution centered at its CPU. As the Director of the Intelligence Science and Technology Department at Wuhan University of Technology explained, the simple platter of hardware systems and software resources cannot fully meet the demands of faculty members and students for AI teachings and the resources for practical training. Only an integrated Terminal-Edge-Cloud solution can effectively connect resources end-to-end to help the school foster practical talent in the field of AI and improve the design of AI-related programs. It can also help students better achieve innovation breakthroughs and transformation of training achievements by meeting their needs for professional skills so that the students truly serve the innovative movement of the industry.

Currently, with the steady advancement of major strategies nationwide such as the 14th Five-Year Digital Economy Development Plan, Made in China 2025 and New Generation Artificial Intelligence Development Plan, the concept and development of new engineering has become an inevitable requirement for the upgrading and evolution of industries in our current world. At the Wuhan University of Technology, it is an important step in the research and practice of new engineering to focus on empirical research and related practices, paying attention to both faculty members and students and implementing the proper professional accreditations and qualifications. As a leader in the innovation of software and hardware systems, Intel is accelerating the in-depth integration of education with science and technology through an enriched portfolio of AI solutions and a strong ecosystem so as to cultivate more engineering and technology talents with the capabilities of engineering innovation and adaptability to future changes.

Figure 11: School of Computer Science and Engineering, Wuhan Engineering University School of Artificial Intelligence.

About Wuhan InnoVision

As part of China’s new generation of information technology enterprises, Wuhan InnoVison Information Technology Co., Ltd. focuses on education informatization, big data, cloud computing, mobile Internet, Internet of Things, information security and other fields. Its businesses cover a range of ICT products and services, application software development, system integration, value-added distribution, and other ICT services. In the context of “Internet,” the company fully promotes the smart campus strategy. Using the Internet platform, with software systems and campus terminals as the entry point, it combines the Internet with the education industry through information and communication technologies to create a new ecosystem in the field of education. At present, the company has an in-depth collaboration with the education informatization authorities of cities and counties in Hubei Province. Based in Wuhan, its system platform covers the entire region of Central China.

Download the PDF ›