シラバス詳細

コースナンバリング
科目名	Advanced Study of Machine Learning（機械
開講学期	後期
開講時期	３クォータ
曜日・校時	月４
単位数	2
授業担当教員	皆本　晃弥

講義形式

Lecture in person

講義概要

This lecture course focuses on fundamental aspects of machine learning and provides its mathematical basis such as probability and information theories, linear regression and classification problems. One of the goals is to understand such basis of neural networks.

開講意図

This lecture course intends to provide mathematical basis of machine learning such as probability and information theories, linear regression and classification problems, to understand such basis of neural networks.

到達目標

Our goals in this course are:
1. understand basics of probability and information theories for machine learning,
2. learn how to mathematically describe various functions and models in the probability and information theories, and in machine learning.
3. understand basic mathematical knowledge on neural networks as a part of machine learning.
4. touch latest topics in the field and be able to discuss about mathematical details and prospects in the field.

授業計画

回	内容	授業以外の学習本科目は、単位数×45時間の学修が必要な内容で構成されています。授業として実施する学修の他に、授業の内容を深めるために以下の事前・事後学修が必要です。
1	Brief introduction to machine learning	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
2	Preliminaries: curve fitting, probability theory	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
3	Preliminaries: models, decision/information theories	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
4	Probability distribution I	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
5	Probability distribution II	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
6	Probability distribution III	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
7	Regression and Bayesian methods I	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
8	Regression and Bayesian methods II	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
9	Linear models for classification I	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
10	Linear models for classification II	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
11	Neural networks I	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
12	Neural networks II	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
13	Neural networks III	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
14	Neural networks IV	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.
15	Latest research topics	Work on the prescribed assignments. -- For the first half of the course, students are urged to learn more on linear algebra and multivariate calculus. -- For the latter half of the course, students are urged to learn more on vector and functional analysis, mathematiacl models, geometry, and some numerics. -- Throughout the course, students are expected to solve some excersises and study some related textbooks for graduate, a list of which will be provided by the lecturer in due course.

成績評価の方法と基準

Problem solving (30%) and an essay or a presentation in English (70%) will be requested.

開示する成績評価の根拠資料等

Assignment, Aim of the assignment

開示方法

Students who want to request disclosure should contact the lecturer.

教科書

資料名		版
著者名	発行所名・発行者名	出版年
備考（巻冊：上下等）		ISBN
Pattern recognition and machine learning
Christopher M. Bishop	Springer	2006
		9780387310732

オフィスアワー

As needed. Reservations required by e-mail.

アクティブラーニング導入状況

アクティブラーニング導入状況
カテゴリー4	カテゴリー3	カテゴリー2	カテゴリー1	カテゴリー0
学生が自ら主体となって、学習の方向性を定め、問題解決に導くための時間です。PROBLEM BASED LEARNING	グループや個人で行った能動的学習の成果を、教室内外で発表し、その評価を受けたり、質問に対応したりすることにより、学修した内容を深化させるための時間です。OUTPUT	学生自らが自由に発言し、グループやペアでの協働活動により課題に取り組み、何らかの帰結に到達するための能動的学習の時間です。INTERACTION	学生からの自由な発言機会はないものの、授業時間中に得られた知識や技能を自ら運用して、問題を解いたり、課題に取り組んだり、授業の振り返りをしたりする能動的学習を行う時間です。ACTION	基本的に学生は着席のまま、講義を聞き、ノートをとり、知識や技能を習得に努める時間です。INPUT
0	0	10	30	60