Bright pictures/vision china

  Artificial intelligence personnel training is the core of change.

  Artificial intelligence, genetic engineering and nano-science are the three cutting-edge technologies in the 21st century, which are the core of industrial revolution 4.0. Among them, artificial intelligence involves a wide range of knowledge fields, including the mathematical foundation, technical foundation, machine learning method and problem domain in the technical system, as well as the Internet, intelligent writing, machine translation, intelligent transportation, intelligent agriculture, intelligent finance, intelligent medical care, robots, auxiliary education and intelligent manufacturing in the application fields. It can be said that the scientific research innovation and personnel training of artificial intelligence determine the advantages and disadvantages of a country in international competition.

  At present, the distribution of talents in the field of artificial intelligence in the world is extremely uneven, with the United States accounting for nearly half. Although China is located in the second echelon, the gap is still large. Not only is the talent reserve small, but top talents are even scarcer. Internationally, artificial intelligence research in Britain, America and other countries started earlier and developed well. Since the 1950s, the American artificial intelligence major has formed an initial boundary, and a number of graduate students have participated in research laboratories. In 1980s, the artificial intelligence graduate program was established, and gradually equipped with interdisciplinary software and hardware support, which greatly promoted the cultivation of professional talents, technology transformation and application. In recent years, the British government has also paid great attention to the research and application in the field of artificial intelligence, and invested a lot of money to support the artificial intelligence industry and start-up companies; Encourage universities to update knowledge, transform property rights and cultivate talents through policy tools, and plan to add 450 doctoral programs related to artificial intelligence from 2017; It also supported the establishment of alan turing Institute to strengthen the research in key algorithm fields.

  Massachusetts Institute of Technology (MIT)

  Massachusetts Institute of Technology has a profound accumulation in computer science and artificial intelligence, which has gone through more than half a century from early theory to later practice. At present, there is no artificial intelligence major in this school, and the personnel training is mainly concentrated in Schwarczman Computer College, which cooperates with the five colleges, namely, Computing Science and Engineering Center, Electronic Engineering and Computer Science Department, Institute of Data System and Society, and Computing Research Center, to carry out collaborative training, research and innovation. The Center for Computing Science and Engineering offers a master’s program in computing science and engineering. The courses cover knowledge from aerospace to nanotechnology, from Internet protocol to telecommunication system design, focusing on advanced computing methods and applications. The doctoral program of computational science and engineering has jointly set up doctoral courses with eight departments, focusing on the development of new computing methods related to science and engineering disciplines. Optional courses are offered by eight departments, including civil and environmental engineering, mechanical engineering, material science and engineering, chemical engineering, earth, atmosphere and planetary science, aerospace, mathematics, nuclear science and engineering. The courses of the Department of Electronic Engineering and Computer Science are mainly composed of electronic engineering, computer science, artificial intelligence and decision-making. The Department offers a number of degree programs for different levels and needs of students, including Master of Computing and Cognitive Engineering (MEng), Doctor of Computer Science (PhD), Doctor of Computer Science and Engineering (PhD), Master of Computer Science and Molecular Biology Engineering (MEng), Doctor of Electrical Engineering (PhD), and Master of Electrical Engineering and Computer Science Engineering.(MEng)/ Master of Science (SM)/ Doctor (PhD), etc. In addition to attending degree courses, students are required to attend three industry seminars, participate in industry internships, complete industry joint projects/patents/joint publications, receive industry invitations for academic sharing, practice in the government, practice in academic institutions, participate in academic training, and submit entrepreneurial projects. In the programs awarded by Schwarczman Computer College, PhD and ScD are regarded as interchangeable. Computer Science and Artificial Intelligence Laboratory (CSAIL) is the largest laboratory in MIT and the most important information technology research and development center in the world. CSAIL’s research fields involve algorithms and theories, artificial intelligence and machine learning, computational biology, computer architecture, graphics and vision, human-computer interaction, programming language and software engineering, robotics, security and cryptography, systems and networks, and it is a platform for cross-cultivation and research of the school’s characteristics.

  Carnegie Mellon University

  Carnegie Mellon University is a leader in artificial intelligence research, with an artificial intelligence major in its undergraduate program. The training in the research stage is distributed in many departments, mainly concentrated in computer college, software engineering research institute, robotics research institute, human-computer interaction research institute, language technology research institute and machine learning department. CMU teachers have diverse backgrounds. Nearly 200 faculty members come from 11 departments, and their research scope covers many fields related to artificial intelligence, from mathematics and physics to computers, art to economics and management. The doctoral program of "Robotics" was carried out in Computer College. Each student must complete core courses and specialized courses. Core courses need to be selected from four core areas. Specialized courses need to be selected by students in a specific core area and complete 48 credits in this direction. Usually, there are four graduate courses. The four core areas are: perception (vision, image sensor, distance data interpretation, tactile and force sensor, inertial guidance sensor, etc.), cognition (robot’s artificial intelligence, knowledge, representation, planning, task scheduling and learning), movement (kinematics, dynamics, control, manipulation and movement) and mathematical foundation (optimal estimation, differential geometry, computational geometry and operational research). The Master of Automation Science (MSAS) is the first master program of automation science in the world. It mainly provides training in three aspects: first, practical training in using scientific experimental robot instruments; second, data analysis and modeling using machine learning and related methods; and third, selecting experiments using artificial intelligence.The professional courses required by this project include four modules, namely, background knowledge, modeling and analysis, automation science, practice and career discussion. The program provides students with professional and research options. Before the start of the second year, students can choose further course modules according to their career development intention to enter the labor market or scientific research market in the future, and the students who choose research will match a research tutor. Master of Human-Computer Interaction (MHCI) is the first professional degree program in the world devoted to human-computer interaction, user experience design and user-centered research. The core of the course is interdisciplinary, and students come from different backgrounds such as design, social science, business and computer science. In the first semester, students need to learn core methodology and technology in the classroom, take elective courses in the second and third semesters, and complete an important industry project with external customers. CMU also promotes the research and application of artificial intelligence by establishing laboratories and centers in cross-fields. Robotics Research Institute was established in 1979, and regularly cooperates with government, industry and non-profit organizations in sponsoring research and education.

  University of Illinois

  The University of Illinois at Urbana-Champaign has not set up an artificial intelligence major alone, and its training focuses on computer science department, electronic computer engineering department and information science department, and has set the direction of artificial intelligence. The computer science department provides training programs for master, master, general and direct students. The course of the project covers ten major directions, such as compilation, AI and bioinformatics. Students need to choose three major directions in the basic course sequence first. Secondly, you need to choose at least one advanced course in one of the selected directions of basic courses; In the elective part, at least three courses in ten major directions can be selected to ensure the breadth of students’ basic knowledge of planning. In recent years, in the research field, various departments in the school, off-campus enterprise associations and government agencies have jointly carried out various AI research cooperation, and established a number of cross-research central laboratories. Founded in 2015, the Intelligent Robot Laboratory (IRL) is highly integrated with seven departments of engineering schools such as aviation engineering and electronic engineering, and the College of Agricultural Consumer and Environmental Science, and cooperates with the Coordination Science Laboratory (CS) to build intelligent robots. In November, 2019, the Artificial Intelligence Innovation Research Center (CAII) was built on the basis of the National Supercomputing Center (NCSA). It is dedicated to promoting AI research, providing AI employment opportunities for students, and coping with major challenges in the industry through cooperative scientific research and innovation, and has become a link between academia and industry on campus. In August, 2020, the National Science Foundation (NSF) and the National Institute of Food and Agriculture (NIFA) gave the UIUC Digital Agriculture Center $20 million to establish the Intelligent Farm Research Institute.In May, 2021, Illinois and IBM planned to carry out a ten-year strategic cooperation with UIUC Institute of Technology, spending 200 million US dollars, mainly focusing on fast-growing areas such as AI and hybrid cloud.

  University College London

  University College London has a good academic reputation in the field of artificial intelligence, with a huge system of departments and strong cross-training characteristics. In particular, its reinforcement learning and neurocomputing are in an advantageous position in the UK, and its research and application involve data science, information science, electronics and electricity, biomedicine, education, architecture, brain science, finance and other fields. UCL joined the European Research Laboratory for Learning and Intelligent Systems (ELLIS) in 2020, and jointly established the UCL ELLIS Unit in the school, which focuses on the research of basic science, technological innovation and social impact. The major of artificial intelligence in this school is rooted in the Computer Department of the College of Engineering Science, and there is an artificial intelligence research center under the Computer Department. The center offers doctoral programs in basic artificial intelligence, master programs in machine learning, master programs in computational statistics and machine learning, master programs in data science and machine learning, and master programs in computer vision. The doctoral program focuses on the research and training of basic and interdisciplinary subjects, as well as elective courses in neuroscience, entrepreneurship, artificial intelligence ethics, etc. It also requires comprehensive ability training such as speech and communication skills and writing skills, and has the opportunity to participate in professional seminars offered by Turing Center in Allen and internships provided by cooperative organizations. Master’s courses usually consist of compulsory courses, core elective courses, general elective courses and graduation projects. Students have a high degree of freedom in course selection, and the course selection of students from different majors can be similar.The course doesn’t set too high demands on students’ mathematical foundations, such as linear algebra and calculus, but pays more attention to the algorithm realization of artificial intelligence and its integration with other disciplines, and its development at the application level is more prominent. In recent years, with the deepening cooperation between UCL and Deepmind in teaching and research, the two sides jointly opened a "Deep Learning Lecture", which was taught by several Deepmind staff and UCL professors, including 12 topics such as convolutional neural networks for image recognition and variational judgment. It is a very popular postgraduate course. In addition, UCL has established a set of academic career framework for the professional development of teachers and researchers. The framework has set four levels, and the promotion is measured by scientific research/teaching activities, core competence, professional expertise and influence. For artificial intelligence-related majors, UCL has deep cooperation with off-campus enterprises, and hired industry veterans as professors in the field of machine learning to promote the intercommunication between theory and practice.

  Characterized by interdisciplinary multi-intersection.

  In terms of specialty setting, the field of artificial intelligence in Britain and the United States appeared early and developed rapidly. As a specialty, artificial intelligence in case colleges has appeared in all stages of this master’s degree, but more training and research are carried out in a cross-disciplinary and interdisciplinary way. Many departments offer a variety of degrees, and students can choose master’s or doctoral level, academic degree programs or professional degree programs according to their own abilities and career plans.

  In terms of curriculum structure, although the names of module structures in different schools are different, they basically classify courses according to fields and directions, allowing students to choose a number of "wide-field" courses and then advance to "specific-field" courses. This can not only ensure the professionalism of training, complete the hard requirements of credits, but also meet the practical needs of students’ personalized and closer to career development as much as possible. Comparatively speaking, American courses pay more attention to students’ mathematical foundation, while British courses pay more attention to algorithm realization and its integration with other disciplines.

  In terms of training programs, compared with academic degrees, professional degree programs require higher practical output and more flexible academic achievements. In some programs, students not only need to complete course credits and link credits (papers, practice, reports, etc.) and reach the minimum required cumulative GPA, but also need to complete specific projects (such as "professional perspective requirements of MIT" and "top industry projects" required by CMU) in order to obtain a degree.

  In terms of characteristic courses, the trend of "AI+X" is obvious, and there are increasingly detailed and diverse interdisciplinary courses, from computer, machinery, electronics and robotics to aerospace, biology, medicine, language, society, economics and management, architecture, education, art, philosophy, humanities, civil engineering, transportation, agriculture and other fields. Some courses pay special attention to the cultivation of students’ practical ability, and also require students to complete specific projects in the real environment in combination with the needs of the community and industry. At the same time, courses on humanistic quality, law, ethics, expression and communication, writing and so on are also set up to cultivate comprehensive ability.

  In the aspect of cross-training, case universities have established a cross-platform that runs through Industry-University-Research, and actively combined with academic resources of multi-disciplines in the school, various industrial resources in the community and government project resources to provide good support for talent training and scientific research innovation.

  In terms of teaching staff, colleges and universities continue to expand excellent scientific research teams, including professors from multiple disciplines and key fields, or hired as tenured professors or awarded honorary/chair professors; It also attracts many senior people who know the industry as part-time professors/business directors. Some schools create new posts (such as teaching posts, cross-research posts, or off-campus participation posts) by setting up new cross-research platforms, which helps solve the teaching/research status of young scholars to some extent.

  At the same time, the discipline and training of artificial intelligence also reflect some international trends. First of all, the cooperation between departments and the establishment of cross-training platform is an important way to promote the breadth and depth development of "AI+X". Secondly, close cooperation with off-campus companies, governments and research institutions can absorb external superior resources and strength and build a "Industry-University-Research integration" discipline ecology. Thirdly, the number of practical, project and application courses has increased, and the course content is more interdisciplinary and closer to the needs of the industry. At the same time, it pays attention to the cultivation of engineering students’ humanistic quality and comprehensive quality. The design of training scheme increasingly embodies the unity of professionalism and flexibility. Finally, the continuous expansion of teachers with diverse backgrounds is an important foundation for promoting the development of artificial intelligence.

  (Author: Li Fengliang Pang Yaran, associate professor and doctoral student of Tsinghua University Education Research Institute)