Knock on the blackboard! The AI teaching assistant calls you to "copy your homework".

Text | China Science Daily Intern Nie Yidan Reporter Chen Bin

"Digital Teacher" launched the "Power System Basics" course. Photo courtesy of the interviewee

"Hello everyone, I am the digital assistant of Professor Liu Yanli of Tianjin University, and your learning buddy!"

On April 2, in the first class of the course "Fundamentals of Power System", Pan Linglan, a 2022 undergraduate student of the School of Electrical Automation and Information Engineering of Tianjin University, looked curiously at another "Teacher Liu" on the screen, and excitedly exchanged heads with his classmates.

This "digital teacher" is the "AI (artificial intelligence) teaching assistant of the whole process and all links" led by Liu Yanli, a professor at the School of Electrical Automation and Information Engineering of Tianjin University. This 1:1 simulation of Liu Yanli's image of the "digital teacher" has many powerful functions.

"The whole process and the whole process of AI teaching assistants can cover all stages of the course and realize all-weather online services. During the course, the AI teaching assistant's Q&A system can answer students' course questions at any time, providing students with a timely and convenient Q&A channel; After class, AI teaching assistants are equipped with test questions for the knowledge points of each chapter, which is convenient for students to consolidate their knowledge, and shows the overall knowledge system of the course through a mind map, which is convenient for students to check and fill in the gaps." ”

And these are just the "tip of the iceberg" of this teacher's "versatile powers".

Let "teaching students according to their aptitude" become a reality

In the past, teachers had to record MOOC courses themselves, which took a lot of time and effort to prepare for the lessons. Liu Yanli told China Science Daily, "'Digital Teacher' can generate MOOCs with one click according to the courseware materials imported by teachers, generate and update MOOC resources anytime, anywhere, efficiently and quickly, and consolidate and expand students' learning content in a targeted manner." ”

After taking the course for a while, Pan Linglan also got to know this teacher to a certain extent. "After a class, if I don't have a deep understanding of the key points and difficulties in the class, I will watch the online class explanation in the AI teaching assistant system, and then do test questions to deepen my mastery. Pan Linglan told China Science Daily that now, AI teaching assistants have become inseparable learning assistants for students.

Another major system of AI teaching assistants, the "Q&A Model", goes beyond classroom teaching, allowing interested students to focus on high-level professional learning.

Liu Yanli introduced that the "Q&A model" is a purely domestic human-computer dialogue model system that serves scientific research and learning, "similar to ChatGPT abroad".

On the one hand, the model can provide 24-hour Q&A service to solve students' problems and questions in terms of professional knowledge; on the other hand, similar to the online shopping platform to recommend personalized products to users, it can recommend different professional questions to different students according to the interaction with each student, make up for the shortcomings of students' professional knowledge, and "realize real teaching according to aptitude".

In addition to the basic Q&A, the "Q&A Model" can also provide students with sufficient auxiliary support in terms of programming skills. "The discipline of power systems involves a lot of analysis and calculations, which need to be programmed. Large models help students write and debug programs, saving time spent on training programming skills. In this way, students have more energy to focus on higher-level professional tasks. Liu Yanli said.

This process is not only a process of professional practice, but also a process of cultivating students' digital literacy. Liu Yanli proudly said that students only need to have basic programming thinking, input instructions into the large model, and get the corresponding code. "Now, the students are very proficient in the application of large models, and they have truly achieved AI as a 'productivity'. ”

Inspired by virtual streamers

"The integrated application of AI technology in the field of education is the trend of the times. In Liu Yanli's view, with the rapid improvement of data storage and computing capabilities, as well as the continuous breakthrough of core algorithms, intelligent teaching supported by artificial intelligence technology will further expand the teaching interaction space and promote the creative application of teaching methods.

However, many large models on the market still have risks in terms of usage rights. In this context, it has always been Liu Yanli's wish to integrate and develop a purely domestic teaching tool to serve students and create a domestic general scientific research and teaching platform.

The inspiration for the "digital teacher" originally came from the virtual anchor of the network.

Liu Yanli told China Science Daily that in the first two years on the Douyin platform, she could often see some virtual anchors, so she thought that if this technology was applied to the teaching scene of higher education, it would inevitably greatly reduce the time and energy required by teachers to record courses, and provide online courses for students more efficiently.

In addition to recording lessons, personalized Q&A is also a much-needed feature in the learning process for students. Based on years of teaching experience, Liu Yanli thought that in the process of knowledge learning, students' questions can basically be summarized in a scope. "Well, it's perfectly possible to train a 'doppelganger' of me to be online 24 hours a day to answer these questions. When human-machine dialogue technology came along, I immediately had the idea of applying it to subject teaching. ”

In the field of higher education, AI teaching assistants must have deep professional knowledge reserves in order to replace real teachers in the classroom, which puts forward higher requirements for the research and development of large models. "In order to create a large-scale model with high safety for the power and energy discipline, we found a third-party company to cooperate and jointly develop it. ”

She said: "The research and development of large-scale models of general and professional integration is becoming increasingly important. On the basis of the general model, we have strengthened the professional knowledge of power and energy, and built a professional model of power energy, which can serve the scientific research of artificial intelligence-enabled new power systems, and become a good teaching assistant to accompany and guide students in learning. ”

Looking forward to AI teaching assistants "blooming everywhere"

As a demonstration course of Tianjin University, the course "Fundamentals of Power System" has become the stage for the first appearance of this "digital teacher". In the near future, this "AI teaching assistant in the whole process and all links" will be fully introduced into a more suitable professional course - "smart grid and artificial intelligence".

"This course focuses on the intersection of power energy and artificial intelligence, and we will explore the use of AI technology to solve the core key technology research and development of smart grids. Liu Yanli plans to transform the application of AI technology in the new power system into a student project, and invite international teachers to carry out the construction of MOOC resources, so that students can use AI teaching assistants to learn theoretical knowledge and practical skills more efficiently.

Allowing AI teaching assistants to serve the cultivation of interdisciplinary talents is the core goal of Liu Yanli in the future.

"According to the previous training model, students who want to enter the interdisciplinary field may need to spend a long time minoring in a second major. "However, with the assistance of AI teaching assistants, students can directly focus on the research topics in the field of interdisciplinary engineering, which greatly shortens the time spent on learning courses in other fields, and at the same time, the school can complete the short-term, high-quality, and personalized cross-talent training tasks." ”

In terms of interdisciplinary teaching teams, AI teaching assistants can also help build an online cross-tutor team. "Virtual teachers from different disciplines can work together to form a cross-mentor group in a short period of time to guide students to carry out cross-cutting research. At the same time, the cooperation of multi-professional virtual teachers can also explore more interdisciplinary points, which is conducive to cultivating multi-field, inter-professional and comprehensive talents. ”

The creation of multidisciplinary "digital teachers" is inseparable from the participation of more universities and teachers. When asked about the promotion of "AI teaching assistants in the whole process and all links", Liu Yanli was very excited, "After the release of our AI teaching assistants, many teachers came to me to 'learn from their experience', and I shared my research and development process with them one by one. ”

She said: "At present, with the deepening development of artificial intelligence technology, the courses of pure knowledge can be completely replaced by 'digital teachers'. We must embrace cutting-edge technologies, transform higher education, and nurture high-caliber and innovative talent. ”

In Liu Yanli's view, AI teaching assistants have strong scalability, and any major can try to create their own "digital teachers". "I'm thinking of making a MOOC of my experience in developing AI teaching assistants, and my 'digital teacher' will share it with you!