To build a high-quality undergraduate academic exchange platform, establish a full-chain innovative education system, stimulate faculty and student academic enthusiasm, and fully enhance their research and practical innovation capabilities, Zhejiang University has launched a special student research project application for its "Qizhen Wenxue" Innovation Platform. Following comprehensive assessment, 15 national innovation projects and 30 general ones have been approved for 2025.
The project "Research on Driven Phonon Band Prediction for High-Throughput Screening of Material Thermal Properties" applied by the team from ZJUI undergraduates, Zhao Hongshuo, class of 2028 in Electrical Engineering, Li Yang and Zhang Shihan, class of 2028 in Mechanical Engineering, has been successfully approved as a national innovation project. The research will be carried out under the guidance of Associate Professor Wee-Liat Ong from ZJUI.
Phonon band calculation is a crucial step in understanding and studying the thermal properties of materials. However, traditional methods such as density functional theory have limitations of low computational efficiency and high resource consumption, making them difficult to apply to complex systems or high-throughput screening. To address this issue, the team proposed to predict the phonon band structure using diffusion model technology. This approach aims to further reduce the computational cost of phonon band prediction, decrease reliance on supercomputing resources, and thereby accelerate the research and development cycle of thermal functional materials. This is the first time that the diffusion model has been introduced into phonon band research, which holds significant value for promoting high-throughput material screening and industrial applications.
Since their freshman year, this team of students from diverse programs has shown strong interest in research practice. Upon learning Zhejiang University launched the "Qizhen Wenxue" Innovation Platform, they submitted their application out of an exploratory passion. Following a rigorous selection, the team successfully joined the research group guided by ZJUI Associate Professor Wee-Liat Ong—also an instructor of the "Qizhen Wenxue" Innovation Platform.
The team members have fully leveraged the advantages of interdisciplinary integration in fields such as computer science and mechanical engineering, innovatively integrating multi-disciplinary theories and methods to systematically carry out research on driven phonon band prediction for high-throughput screening of material thermal properties. With a solid research foundation and innovative research methods, the team has successfully secured the approval of the national innovation project.
▲Zhao Hongshuo, Li Yang, and Zhang Shihan attended the presentation and review meeting
Zhao Hongshuo’s team, then freshmen, acknowledged facing significant challenges during the project. “First, our grasp of professional knowledge was insufficient. Given the project covers material physics, solid-state physics and other fields, many concepts were entirely new,” they noted. To overcome this bottleneck, the team adopted a systematic learning approach: they not only studied professional textbooks and cutting-edge papers, but also leveraged online course resources for self-directed learning. They also set up a regular discussion mechanism—through in-depth exchanges and ideological collisions, they deepened their understanding and effectively consolidated newly acquired knowledge.
During the technical research process, Zhao Hongshuo faced difficulties due to lack of experience in aspects such as model framework selection, large-scale database construction, and hyperparameter optimization, which once made the project hard to proceed. Fortunately, Zhao Hongshuo's team received meticulous guidance from Assoc. Prof. Wee-Liat Ong and his doctoral students Li Yuchen and Huang Kailin.
"The supervisors not only answered our questions in professional fields, but also encouraged us to be bold in innovation with an open and inclusive attitude. What is particularly worth mentioning is that they specially allocated high-performance computing servers for the project, providing solid computing power support for its implementation. From the demonstration of the research topic and the design of the plan to the debugging of experiments and the analysis of results, the supervisor team has always offered patient and meticulous guidance. Even for basic questions, they would listen carefully and explain them in a way that is easy to understand, making profound concepts accessible.
Such careful and warm guidance has greatly enhanced the professional capabilities and confidence of Zhao Hongshuo's team. In the end, through repeated experimental optimization and iterations, they not only successfully overcame the technical difficulties, but also made significant progress in programming practice and algorithm understanding."
This project was a rare growth opportunity. In applying computer technology to materials science research, we not only systematically mastered learning and research processes, but also gained real insight into the innovative appeal of interdisciplinary integration. Through training in areas like literature research, experimental design, and data analysis, we developed rigorous research thinking—and notably, honed the valuable ability to learn from failures and refine plans.
In addition, they also expressed special gratitude to ZJUI for providing such an excellent international engineering platform, interdisciplinary training programs, and abundant opportunities for research practice. Such systematic practical experience and opportunities for multidisciplinary exploration can not only help students clarify their future learning directions at an early stage, but also provide a solid support for them to continuously gain motivation for exploration.
As the team's supervisor, Associate Professor Wee-Liat Ong was not surprised by their successful project approval. He stated, "Zhao Hongshuo's team has demonstrated outstanding learning and practical abilities in their research. Not only did they systematically master the theoretical foundations of phonon physics, but they also conducted interdisciplinary research by integrating computer technology, showcasing solid professional competence and innovative capabilities. This research model, which combines professional knowledge, interdisciplinary research skills, and efficient teamwork, fully reflects the team's excellent comprehensive abilities and growth potential."
ZJUI has long emphasized integrating theory with practice and focused on fostering students’ active learning and innovative exploration abilities via diverse channels—including curricula, competitions, internships and hands-on activities. From theory to practice, unified tasks to open-ended design goals, and course assignments to research and innovation training, students gradually build problem-driven inquiry skills and teamwork capabilities—skills they have fully showcased in various academic competitions and research platforms.
Launched in November 2022, Zhejiang University’s "Qizhen Wenxue" Innovation Platform centers on cultivating top-tier innovative academic scholars with the aim to uphold the traditional values of "respecting morality and pursuing learning". Innovative instructors are selected from scholars of professional departments, and students from undergraduates. Based on research interests, faculty and students undergo two-way selection; once matched, personalized long-term training is delivered via "inquiry groups".
