On October 21st, Dr. Christos Merkatas from Aalto University in Finland gave a talk at the iTHEMS Biology Seminar. His talk was entitled, “Bayesian Nonparametric Estimation of Random Dynamical Systems”. Suppose that the observed time series is small, and the noise process is non-Gaussian. How can we reconstruct and predict the behavior of the system? Dr. Merkatas showed that the proposed Bayesian approach enables us to reconstruct and predict the system by inferring the number of unknown components and their variances. According to him, the method can be applied to the problems in physics, biology, and economy. Since some of the audiences including me were not familiar with his method, we asked many basic questions. The talk by Dr. Merkatas was great and also educative. Gen Kurosawa