Project description/goals
Investigate statistical tensor signal processing method for multi-dimensional data.

Importance/impact, challenges/pain points
Conventional methods need costly hyper-parameter tuning. The proposed method try to mitigate this pain point, and achieves automatic model order selection. 

Solution description
Propose intelligent tensor signal processing method that is based on statistical learning.

Key contribution/commercial implication
Realize automatic hyper-parameter learning for tensor models in data science.