Research Interests
I am interested in machine learning, especially the situation where the information is imperfect which often occurs in the real world. Until now, I have researched on Domain Adaptation, Crowdsourcing and corrupted labels. I am also interested in decision making and Reinforcement learning.
Experiences
- Research and develop machine learning theories and algorithms.
- Belong to Imperfect Information Learning Team.
Projects
Fashion and A.I. Projects with Emarie @ RIKEN AIP.
- Collaboration of machine learning method (especially deep learning) with professional fashion designer EMarie. The fashion show was held during Amazon fashion week spring season in Tokyo.
Machine Learning in Planetary science.
- On going.
Publications
Learning Only from Relevant Keywords and Unlabeled Documents
EMNLP 2019
On Symmetric Losses for Learning from Corrupted Labels
ICML 2019
Domain Discrepancy Measure for Complex Models in Unsupervised Domain Adaptation
Preprint in arXiv.