The "SunPredict" team, comprised of doctoral students Yongchae Kim and Sangkyung Lee, and master's students Bongmo Kim, Jiwon Lee and Minjeong Kim, affiliated with Professor Byungdong Yoon's Laboratory for System Health & Risk Management, won the encouragement award at the 5th Solar Power Generation Prediction Competition, organized by POSTECH Open Innovation Big Data Center, Industrial Artificial Intelligence Program, and H Energy.

 

The competition aimed to open up the closed traditional energy market through technology and innovation, establishing an artificial intelligence platform ecosystem accessible to everyone, developing technology, leading global research, and addressing societal issues. Accordingly, a total of 119 teams participated, competing to develop ensemble techniques to enhance solar prediction performance. The SunPredict team processed 13 real-time weather observation data sets appropriately and developed a high-level model using transformer-based deep learning techniques. Additionally, they introduced a probability-based ensemble technique, showcasing robust performance during the competition period and achieving commendable results.

 

Regarding this award, Professor Byungdong Yoon commented, "Research on the interpretability of AI model results is essential not only for developing AI models to improve the accuracy of industrial data but also for applying them to actual industries."