ZJU Undergraduate Course on Artificial Intelligence: Learn and Practice AI in Your Browser

Editor: Yu Liu     Time: 2020-04-28      Number of visits :22


In response to the COVID-19 pandemic, the School of Computer Science launched online lecturing. The new teaching method allows teachers and students from all over the world to gather in the cloud to jointly create a classroom experience different from the past.

Undergraduate Course on AI

Artificial intelligence is an enabling technology like water and electricity. It is multi-disciplinary and cross-intrinsic by nature. Artificial intelligence is not just a course, a first-hand technology, a product or an application, but a comprehensive synergy of profound theory, dynamic technology, and society empowerment.

After nearly five years of reform, the undergraduate artificial intelligence course offered by the School of Computer Science has established a teaching concept centered around knowledge, comprehension, procedures and application. It is a beloved course by students and has formed a core of symbolism for logical reasoning, exploration based on problem solving, machine learning driven by data, reinforcement learning based on behaviorism, and collective intelligence with game theory at the core.

Teaching Team Adjusts for Pandemic

According to the previous plan, the undergraduate course on Artificial Intelligence in spring 2020 was already upgraded. Huawei continues to provide Huawei Cloud ModelArts training capabilities, and students used their AI chip Atlas 200 for face recognition, speech recognition or natural language processing tasks from the Cao Guangbiao West laboratory at Yuquan Campus.

Authorization ceremony for authors of the new generation artificial intelligence serial textbooks at the 2nd National Moganshan Forum on Artificial Intelligence Talents and Technology


However, the teaching schedule was abruptly disturbed because of the pandemic. In the school's arrangement for stopping classes without stopping teaching, the artificial intelligence undergraduate course was tasked to retain the effects of classroom learning in an online environment.


Learning AI in the Browser


A total of 238 students enrolled in the artificial intelligence course on the ZJU course-selection platform. Another 247 students are auditing in the Dingding group, so there are a total of 485 students attending the online lectures. The course was created by five teachers: Wu Fei, Wang Donghui, Li Xi, Zheng Nenggan and Li Yingming. In addition, 22,000 people signed up for the MOOC version of the course on icourse163. The teaching team was pleased that online learning overcame spatial limitations, but they were also concerned about helping students convert the models they learn into code to further understand the principles behind the algorithms.


Teachers also carefully prepared courseware to ensure the quality of teaching. In addition to high-quality live-streams, “Mo” (an online artificial intelligence education platform at https://edu.momodel.cn/) was also established to provide algorithm training for students. The Mo platform is jointly developed by the team members of the Institute of Artificial Intelligence of the School of Computer Science and the School of Public Administration. It is a platform for the training and popularization of artificial intelligence algorithms. The Mo platform provides an environment for writing and learning machine learning algorithms in the browser

Game Algorithm and Image Recognition Assignments


On the Mo platform, students began to tackle their first practical assignment: miniAlphoGo. In this assignment, students were asked to design a game algorithm to compete with each other. The platform automatically performs quality tests on the algorithms submitted by students, and then organizes the tested algorithms to conduct a group game, and finally ranks all the algorithms in turn according to the results of the round-robin game. In order to mobilize everyone's enthusiasm, the students are rewarded according to the ranking. This mechanism stimulates the competitive spirit of the students who are willing to code, and motivates them in the learning of artificial intelligence.


The second assignment is more challenging, as it requires students to be familiar with mainstream image classification models and algorithms based on convolutional neural networks such as ResNet, VGG, Inception V3 through the Mo platform. On this basis, Huawei's artificial intelligence development platform (MindStudio, ModelArts) and development framework MindSpore are used for model training and empowerment.


In this assignment, students use ModelArts, a cloud computing resource provided by Huawei, to perform parameter tuning on complex neural networks for image classification on the local machine. This helps students experience the immense power of deep convolutional neural networks and the charm of artificial intelligence algorithms.  


Artificial Intelligence Course Strategy


The course covers theoretical knowledge such as deep learning, statistical learning, search strategy, logical reasoning, and reinforcement learning. The practical training programs are set up reasonably according to the course content, which is more helpful for students to strengthen their understanding of theoretical knowledge. The hands-on tool platform is provided by Huawei and includes resources such as toolkits and hardware equipment, so that students can quickly implement their own solutions in the training projects. The course can help students understand the field of artificial intelligence and lay a solid foundation for further research in the future; it also provides students with practical project training to help them gain hands-on experience. Furthermore, the course is equipped with multiple teaching assistants and technical assistants, so that the problems encountered by the students can be resolved in a timely and effective manner. At the same time, the course content is fully synchronized on Zhejiang University platforms, Dingding and QQ Groups, providing students with an agile learning experience.


Education after the epidemic will be conducted both offline and online for the foreseeable future. Classroom teaching, online courses, teaching materials, practical training platforms, industry-university cooperation are indispensable.


After the epidemic, the teaching team will continue to take root in the needs of the economy, society, and national security. At the same time, the teaching team is also making great efforts to contribute human resources for artificial intelligence to become a dynamic engine of economic and social development, and is committed to cultivating young talentsthat can connect with the Intelligence X concept.

Back to Top

Copyright © 2018 College of Computer Science and Technology, Zhejiang University