To train a machine while preserving privacy

We adapt privacy-preserving techniques such as differential privacy and homomorphic encryption to ML algorithms.

To learn a global model without data sharing in a situation where data is distributed

We target a large number of distributed data or data containing sensitive information.

To extract useful representations from graphs for high-performed prediction

We tackle forecasting, classification, and anomaly detection in time-series data.

To help medical decision-making via data analytics

Ongoing research includes the development of a drug recommendation system, disease prediction models, and biological age.

To improve processes or support decision-making in technology management contexts such as technology forecasting and strategy

We cover macro societal issues such as ESG and COVID-19 here. 

To solve the problems in manufacturing such as predictive maintenance, quality control, and causality

We worked with POSCO, SK Lubricants, LG Electronics, and some small business.