Koki Kawabata
DATA MINING · MACHINE LEARNING
Finding structure in evolving data
I develop scalable methods for understanding patterns, changes, and anomalies in complex data streams.
Assistant Professor · SANKEN, The University of Osaka
Research directions
01
Streaming data
Learning patterns and structural changes from data that arrive continuously.
02
Tensor methods
Modeling multi-aspect data with scalable and interpretable representations.
03
Anomaly detection
Identifying subtle and evolving signals in complex real-world systems.
Research in motion
Current interests include temporal graphs, nonlinear tensor analysis, event streams, and real-time forecasting.