On the morning of April 15th, the launch ceremony of the second phase of the cooperation between the Department of Computer Science, Tsinghua University, and the TAL Intelligent Education Information Technology Joint Research Center (hereinafter referred to as the "Joint Research Center") was held at Tsinghua University. The two parties will deepen the frontier exploration of intelligent education based on the achievements of the first phase of the Joint Research Center.

Liu Yiqun, Party Secretary of the Department of Computer Science, Tsinghua University, Zhou Lizhu, former head of the Department of Computer Science, Tsinghua University, Feng Jianhua, Deputy Head of the Department of Computer Science, Tsinghua University, Li Guoliang, professor of the Department of Computer Science, Tsinghua University and director of the Joint Research Center, Tian Mi, CTO of TAL Education Group, Wu Zhongqin, Vice President of Technology of TAL Education Group and head of the AI ​​Institute, Li Fei, Vice President of Technology of TAL Education Group, and Liu Xiao, deputy director of the Joint Research Center and head of the Vision and Algorithm Platform of the AI ​​Institute of TAL Education Group attended the signing and launch ceremony.

Liu Yiqun, Party Secretary of the Department of Computer Science, Tsinghua University, and Tian Mi, CTO of TAL Education Group start the second phase of cooperation

Phase II of the Joint Research Center will focus on addressing real-world issues in teaching practices across TAL's educational brands including Xueersi Peiyou, Xueersi Online School, and Xiaohou Enlightenment. The collaboration will explore the construction of online teaching environments, design of instructional interaction mechanisms, and intelligent assessment of learning progress across diverse teaching models, academic stages, and subject areas, aiming to build more comprehensive digital education ecosystems, facilitate more scientifically effective teacher-student interactions, and develop personalized student evaluation systems.Meanwhile, the collaboration will strive to achieve technological breakthroughs in knowledge graph construction and reasoning based on authentic educational data, intelligent question generation, and automatic problem-solving capabilities. Both parties aim to promote innovation in core areas of educational technology through this partnership.

At the meeting, Secretary Liu Yiqun affirmed the achievements of the first phase of the Joint Research Center's cooperation and expressed high expectations for the second phase of cooperation. He said, "In the exploration of smart education, TAL Education Group has always been at the forefront of the industry. We are happy to jointly promote the research on educational intelligence with them. I hope that the two parties can fully leverage their joint strengths, and, based on the real scenarios and needs of TAL Education Group, carry out various attempts to build intelligent education, and apply more scientific research results."

Tian Mi said, "We hope to fully utilize Tsinghua University's front-edge technology research experience in big data and artificial intelligence, combine it with TAL Education Group's application advantages in the field of education, continuously explore new models, new technologies, and new methods of combining technology and education, and open the new generation of artificial intelligence innovation platform for smart education to the whole industry, jointly promote the digital transformation of the education industry."

Representatives of both parties had in-depth exchanges at the scene

The expectations and confidence of Tsinghua University and TAL Education Group on the second phase of cooperation are based on the considerable achievements of their previous efforts. It is reported that the cooperation between the two parties began in 2018. While focusing on the research of mining and analyzing education big data, exploring the basic theory and methods of intelligent education, and conducting joint research on multiple topics such as automatic problem solving, oral expression evaluation, and personalized speech recognition and evaluation. At present, the jointly developed technological achievements have been successively applied to a variety of products of TAL Education Group, achieving integrated production, learning, research, and application.

One of the systems developed under the "Personalized Emotional Voice Synthesis for Teachers" project enable rapid construction of speaker models using minimal speech data, empowering educators to achieve personalized functions such as customized roll calls and feedback comments. This research achievement was not only presented at Interspeech 2019, a top-tier international conference in speech technology, but also successfully implemented in practical applications. Taking the large-class live teaching scenario of Xueersi Online School as an example, the system employs personalized emotional voice synthesis to generate unique voice timbres for each lecturer in real-time. This innovation allows every student to receive tailored feedback and encouragement, significantly enhancing their sense of engagement in learning.

It is well known that the richness of a teacher's emotions during class is one of the influencing factors of the classroom effect. The "Teacher Emotion Analysis" project jointly researched by both parties mainly calculates the emotion of the teacher's teaching audio and quantifies the emotional state of the teacher's lecture. This research result has been published at the top international conference ACM Multimedia 2019 in the multimedia field. The system developed by the project can intelligently detect and provide feedback on the teacher's pre-class video to ensure the quality of the preparation. In online classrooms, the system can analyze teaching behavior in real time, assist teachers in making immediate feedback in class, and help teachers discover deficiencies in teaching behavior and make targeted improvements.

Meanwhile, the Joint Research Center has also made achievements in cultivating students' learning habits. Focusing on oral explanation - one of TAL's "Five Key Habits" - TAL Education Group organizes annual speech competitions and oral explanation practice activities. The research outcomes from the Joint Research Center have been successfully implemented in TAL's Chinese and English oral expression assessment solutions, addressing issues such as long feedback cycles for manual scoring and subjective scoring criteria in oral explanation evaluations. These solutions enable real-time, objective, and quantifiable multi-dimensional evaluation of students' oral communication skills based on specific questions and their verbal responses.

The smart education situation has already opened up, but the practical difficulties in the deep integration of artificial intelligence and education still exist. The second phase of the Joint Research Center will, under its mission of "Love and Technology Make Education Better," work with various sectors of industry, academia, and research to explore the "no-man's land" of intelligent education using cutting-edge technology and push the industry steadily forward.