The 41st In Silico Megabank Research Seminar will be held on Tuesday, December 10.

This Time, we will be welcoming Dr. Yu Nishiyama, The Institute of Statistical Mathematics as our lecturer, and he will be speaking on “Topic of Bayesian Inference Using Positive Definite Kernel as the Most Recent Kernel Method.”

・Date/Time: December 10 (Tue.) 17:00‐18:30
・Venue: Middle Conference Room (2nd Floor), Building #1, Tohoku University School of Medicine
・Title: Topic of Bayesian Inference Using Positive Definite Kernel as the Most Recent Kernel Method
・Lecturer: Yu Nishiyama ( The Institute of Statistical Mathematics)

＊This lecture is transferable as a class in the medical research-related lecture course.

・Abstract: In recent studies regarding Kernel Method, Bayesian inference using positive definite Kernel (Kernel Bayes’ Rule) is examined. This methodology reasons with a kernel mean setting forth a consistency rate instead of probability distribution. The kernel mean is defined with an average in the feature space rather than regular average of probability distribution.  In this study, we will introduce kernelized version of both “Sum rule” used for marginalization and “Bayes’ rule” producing Bayesian Posterior Distribution known as “Kernel Sum Rule” and “Kernel Bayes’ Rules.” Combination of these two rules results in filtering of state special model and reinforcement learning algorithm. Kernel Bayes’ Rule which is the combined form of probability models performed by the presenters recently is also introduced.

・Organizer : Kaname Kojima, Masao Nagasaki