SNU Professor U Kang Leads the world with AI technologies That Analyze and Predict Real-world Data  
Seoul National University's College of Engineering (Dean Kookheon Char) announced on June 7 that Professor U Kang's research team of the Department of Computer Science and Engineering has developed techniques to accurately and effectively perform real-world data analysis and prediction. These techniques were designed for data with temporal characteristics, and were optimized according to the nature and characteristics of each data.
The research results of the research team is a core technology that is universally used for analysis and prediction of various data such as time series data, tensor data, stock price data, and knowledge graph and is expected to be used in various AI applications in the future.
 

• Development of PFT (Partial Fourier Transform), a technology to quickly and accurately obtain certain coefficients of the Fourier Transform, which is widely used in data analysis.
  “Fast and Accurate Partial Fourier Transform for Time Series Data.”, Yong-chan Park, Jun-Gi Jang, and U Kang.

• Development of Zoom-Tucker (Zoomable Tucker decomposition), a technology that efficiently finds specific time-period patterns of high-dimensional tensor data through Tucker decomposition
  “Fast and Memory-Efficient Tucker decomposition for Answering Diverse Time Range Queries”, Jun-Gi Jang and U Kang.

• Development of DTML (Data-Axis Transformer with Multi-Level Contexts), a model that accurately predicts stock price movements by learning the correlation between stocks
  “Accurate Multivariate Stock Movement Prediction via Data-Axis Transformer with Multi-Level Contexts.”, Jaemin Yoo, Yejun Soun, Yong-chan Park, and U Kang.

• Development of T-GAP (Temporal GNN with Attention Propagation), a model that accurately performs knowledge graph inference by learning a graph neural network considering temporal information from the knowledge graph
  “Learning to Walk across Time for Interpretable Temporal Knowledge Graph Completion.”, Jaehun Jung, Jinhong Jung, and U Kang.

The four papers above will be presented at The 27th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2021), the best conference in the field of big data and artificial intelligence, in August. This is a highly exceptional achievement.