In recent years, students and faculty from the College of Information at National Yang Ming Chiao Tung University (NYCU) have consistently excelled on the global stage. Their research has gained recognition at top-tier international conferences in AI, embedded systems, human-computer interaction, and sports technology. Below are highlights of recent student achievements and reflections from major international conferences:

Title: CoachAI+ Badminton Environment: Realistic Badminton Game Simulator for Enhancing Player Performance
Authors: Peng, Wen-Chih; Wang, Kuang-Da; Chien, Yen-Che; Xie, Bo-Zhou; Chen, Yu-An; Tsai, Cheng-Shiuan; Doong, Shao-Jyun; Hung, Jun-Chen
Advisor: Prof. Wen-Chih Peng
Conference: MIT Sloan Sports Analytics Conference (SSAC)

Significance:
The MIT Sloan Sports Analytics Conference (SSAC) is one of the world’s most influential forums in sports analytics, co-hosted by MIT and 42 Analytics, with ESPN as a partner. It brings together data scientists, coaches, and industry experts from organizations such as the NBA, NFL, and MLB to explore innovations in sports technology and AI applications. Being selected for the SSAC Research Paper Competition finals indicates both practical impact and an opportunity to engage with global experts.

The experience of Kuang-Da Wang:
We were honored to present our paper at SSAC 2025, held at the Hynes Convention Center in Boston. This marked our first time presenting our badminton AI research—previously featured at academic venues like KDD and AAAI—at a more application-focused conference deeply connected with industry.
CoachAI+ is a system designed to simulate realistic badminton singles matches and analyze player performance and tactics. At its core is RallyNetv2, an upgraded imitation learning model that enables natural shot exchanges. By integrating the ShuttleSet dataset with real-world constraints (such as shot types, landing points, and movement behavior), the simulation closely mirrors real matches. Our tactical analysis module includes indicators such as dominance patterns, stamina consumption, shot distribution, and shot impact to support coach-player decision-making.
SSAC offered a unique intersection of sports, technology, and business. Professionals from the NBA, MLB, and startups provided valuable feedback on potential applications, sparking ideas on real-time prediction and practical support tools. During the poster session, I had insightful exchanges with researchers from the UK and US on integrating large language models for better generalization and expanding into doubles simulation. I realized our work is actively bridging research and real-world application.
Thanks to the Precision Sports Project and our dedicated team, this experience not only enhanced my ability to communicate the technical value of our work in English but also deepened my understanding of research as a way to bring AI onto the court and help athletes reach their full potential.

Title: Pinning, Sorting, and Categorizing Notifications: A Mixed-methods Usage and Experience Study of Mobile Notification-management Features
Authors: Yong-Han Lin, Li-Ting Su, Uei-Dar Chen, Yi-Chi Lee, Peng-Jui Wang, Yung-Ju Chang
Advisor: Prof. Yung-Ju Chang
Conference: ACM International Joint Conference on Pervasive and Ubiquitous Computing (UbiComp 2024)

Significance:
UbiComp is a premier conference in the fields of ubiquitous computing and human-computer interaction. It covers cutting-edge areas such as interaction design, mobile computing, wearable technologies, and sensing systems. As a global hub for researchers, designers, and developers, UbiComp not only showcases state-of-the-art research but also delves into its societal impacts, continuously redefining technology’s role in everyday life.

The experience of Yong-Han Lin:
I am deeply grateful to Prof. Chang for his guidance and encouragement, and to my fellow team members for their collaboration, which gave me the opportunity to share our work on an international stage. Our research developed a new mobile notification system with user-centric management features to help people better handle notifications. I’m honored that this work was accepted at UbiComp, which validates our efforts and inspires me to continue pursuing research. The conference provided valuable insights through interaction with leading scholars worldwide and offered feedback that will shape my future research direction.

Title: iFKVS: Lightweight Key-Value Store for Flash-Based Intermittently Computing Devices
Authors: Yen-Hsun Chen, Ting-En Liao, Li-Pin Chang
Advisor: Prof. Li-Pin Chang
Conference: EMSOFT: International Conference on Embedded Software (EMSOFT 2024)

Significance:
EMSOFT is a flagship conference under ACM SIGBED, advancing the science and engineering of embedded software and systems. Since 2001, it has served as a leading venue for presenting cutting-edge research on software interacting with the physical world, especially in cyber-physical systems. EMSOFT is a core part of ESWEEK, alongside CASES and CODES+ISSS, covering the full spectrum of hardware-software co-design.

The experience of Ting-En Liao:
It was a great honor to participate in EMSOFT 2024, one of the top international conferences in embedded systems. I am sincerely thankful to Prof. Chang for his mentorship, which enabled me to bring our research to this global platform. What stood out to me was the in-depth discussions with other scholars during the conference. Their perspectives opened new angles for improving our work, and several researchers expressed strong interest in our method. These exchanges broadened my horizons and inspired further research ideas.

Title: NaRCan: Natural Refined Canonical Image with Integration of Diffusion Prior for Video Editing
Authors: Ting-Hsuan Chen, Jiewen Chan, Hau-Shiang Shiu, Shih-Han Yen, Chang-Han Yeh, Yu-Lun Liu
Advisor: Prof. Yu-Lun Liu
Conference: Conference on Neural Information Processing Systems (NeurIPS 2024)

Significance:
NeurIPS is one of the most prestigious conferences in AI and machine learning, drawing top researchers globally with a highly competitive acceptance rate. It spans from theoretical advancements to applied innovations, serving as a key forum for leading-edge developments in AI.

The experience of Ting-Hsuan Chen:
Life is like diffusion—full of noise at the start. Fortunately, I met Prof. Liu, who helped denoise my journey. With his guidance and teamwork, we successfully integrated diffusion priors into a video editing framework, overcoming limitations of existing canonical-based methods for complex video dynamics. Our novel design combining deformation fields and diffusion priors improved the naturalness and quality of canonical images while reducing training time by 14x through optimized noise scheduling.
Attending NeurIPS was unforgettable. Engaging with researchers from around the world sparked new ideas and motivation for future work.