大阪大学　数理・データ科学教育研究センター
Center for Mathematical Modeling and Data Science,Osaka University

Introduction to Sequential Monte Carlo Methods and Multilevel Sequential Monte Carlo

Ajay Jasra (National University of Singapore)

1. ホーム
2. 学内向け情報
3. セミナー
4. Introduction to Sequential Monte Carlo Methods and Multilevel Sequential Monte Carlo

2016年度 第1回データ科学 ミニレクチャーシリーズ（CREST JST 共催）

Introduction to Sequential Monte Carlo Methods and Multilevel Sequential Monte Carlo

Ajay Jasra (National University of Singapore)

日程
7/5（火） 13:00~14:30　Mini-Lecture 1/3
7/12（火） 13:00~14:30　Mini-Lecture 2/3
7/19（火） 13:00~14:30　Mini-Lecture 3/3

【Abstract】
This course will provide a tutorial style introduction to sequential Monte Carlo methods, attempting to start from a basic introduction, to a coverage of some of the latest research developments.
Sequential Monte Carlo methods are applied in a wide variety of applications, including engineering, economics and biology. They combine importance sampling and resampling to approximate distributions. The idea is to introduce a sequence of proposal densities and sequentially simulate a collection of N >1 samples, termed particles, in parallel from these proposals. In most scenarios it is not possible to use the distribution of interest as a proposal.
Therefore, one must correct for the discrepancy between proposal and target via importance weights. There are a variety of algorithms and extensions which form some of the most cutting edge methodology for `exact' computations in many academic fields. An extension of the methodology for use in the multilevel method is also described and developed.