In this class, we discuss basic concepts in algorithmics such as NP-hardness, approximation algorithms, and online algorithms. Then, we give examples how to apply them to practical research. At the end of this class, students who are not familiar with these theoretical concepts are expected to learn their importance in practical point of views. On the other hands, students who are familiar with them are expected to gain more experiences on applying the concepts to practical settings. 本講義では、NP困難、近似アルゴリズム、オンラインアルゴリズムなどアルゴリズムを解析する理論の概要を説明し、機械学習や データベースなど応用分野に適用する事例を挙げる。 アルゴリズム論を勉強したことがない学生には理論的な解析の重要性を実感させ、勉強したことがある学生にはアルゴリズム論の応用を経験させることを目的としている。

Statistics and biometrics have emerged as crucial disciplines not only in agricultural sciences but also in various other fields. This significance is primarily attributed to three factors: Firstly, advancements in data measurement techniques have facilitated the collection of extensive and diverse biological and agronomic data that were previously unattainable. Secondly, the evolution of data science methodologies has enabled the integration and modeling of such collected data. Thirdly, the enhancement of computational capabilities has empowered the utilization of these methodologies. These advancements have rendered statistical and biometric methods indispensable for extracting insights from the vast and varied biological and agronomic datasets. Throughout this lecture series, a diverse array of biological and agronomic datasets will serve as illustrative examples to demonstrate various analytical methods. Delivered in a hands-on format, utilizing R, Python, and Matlab, the aim is to equip students with practical analysis skills. The initial portion of the course, spanning the first one-third, will focus primarily on techniques for summarizing, visualizing, and modeling relationships within multivariate datasets. In the subsequent one-third, students will delve into linear models, linear mixed models, local regression, and nonlinear models. Finally, in the last segment, students will explore image analysis, machine learning, and deep learning methods. While the course will cover a broad spectrum of methods, ranging from introductory to advanced levels, the emphasis will be on developing the capability to independently conduct analyses rather than on elaborating on the theoretical underpinnings of the methods.

広く産学官にわたってグローバルに活躍するために必要な「俯瞰力」を養成することを目指す。物質科学の各分野について最先端の知識を修得し、自分の専門分野と周辺分野がどのように関連するか、あるいはし得るか、について深く考察するために、第一線で活躍する講師の方々にその分野の最前線を概観していただく。さらに、それらの講義を通して異分野間のコミュニケーションを円滑に進めるための具体的方法論を学ぶ。 This survey course is designed to enable students to develop the broad perspective that is required of global leaders working in and across industry, academia, and government. Students will gain knowledge and insight on advancements in each field of materials science research, given by leading researchers working on the frontline in those fields. This will allow students to consider how peripheral fields are related to their own area of expertise, and to consider the potential for forging bridges between related fields in the future. In addition, students will learn specific methodologies designed to facilitate smooth communication among different disciplines.

In production and consumption of materials, structural materials are dominant. Technological development in materials industry affects significant positive impacts in broad end uses. A newly developed high-performance materials fulfill higher design requirement, which makes more advanced product design possible. The first half of this class introduces production process of major materials and life cycle perspective on materials, especially structural materials. We discuss interactions amongst materials, finished products, economy, and environment in order to understand roles of materials in industrial metabolism. Global material flows and resource strategy are also introduced to give you a global vision on resources and materials. The latter half of this class introduces high-temperature materials for jet engine or gas turbine in the power plant. The processing and the necessary properties of high-temperature materials will be introduced to understand how they are designed.

In this course, students will systematically learn state-of-the-art machine learning and artificial intelligence techniques, including Large Language Models (LLMs), Graph Neural Networks (GNNs), and recommendation systems. Invited researchers from industry will provide insights into how these methods are applied to real-world problems. The course will also cover high-performance computing (HPC) and advanced system infrastructures underlying these techniques, enabling students to gain both theoretical understanding and practical experience. 本講義では、LLM（大規模言語モデル）、GNN（グラフニューラルネットワーク）、推薦システムなど、機械学習・人工知能の最先端技術を体系的に学ぶ。産業界で活躍する研究者をゲストとして招き、実世界の課題に対してこれらの技術がどのように応用されているのかを議論する。また、これらの高度な手法を支える高性能計算（HPC）や最先端のシステム基盤についても学び、理論と実践の両面から理解を深める。

　多体問題は、銀河のダイナミクスから量子ビット間のエンタングルメントまで、自然科学研究の様々な場面に現れ、計算機を用いた計算科学の主な研究対象となっている。本講義では、物質科学を具体例として多体問題に現れる、モンテカルロ法や、ランチョス法などのクリロフ部分空間法など、多体問題が内包する巨大な自由度を有限の計算機資源で扱うためのアルゴリズムを学ぶ。 　また、実際にこれらのアルゴリズムを実装したアプリケーションを使った計算機シミュレーションをレポート課題を通して習得する。 From dynamics of galaxy to entanglement among quantum bits, many-body problems have been studied in the wide range of science by computational science approaches. In this lecture, numerical algorithms for many-body problems in condensed matter physics, such as Monte Carlo samplings and Krylov subspace methods represented by the Lanczos method, are introduced with emphasis on treatment of the large degrees of freedom inherent in the many-body problems. Through exercises by using these algorithms implemented in open source software for many-body problems, students will learn how to utilize them. Papers on application of the algorithms and software to selected many-body problems are assigned as requirements for a credit.

High Performance Computer (HPC)のハードウエア的、ソフトウエア的構成を理解する。HPCの性能を引き出す計算科学向けアルゴリズムを理解する。 To understand the structures of High Performance Computers(HPCs) from the both viewpoints of hardware and software. To understand algorithms for computational science to bring out the performance of HPCs. This lecture will be done in Japanese.

　多体問題は、銀河のダイナミクスから量子ビット間のエンタングルメントまで、自然科学研究の様々な場面に現れ、計算機を用いた計算科学の主な研究対象となっている。本講義では、物質科学を具体例として多体問題に現れる、モンテカルロ法や、ランチョス法などのクリロフ部分空間法など、多体問題が内包する巨大な自由度を有限の計算機資源で扱うためのアルゴリズムを学ぶ。 　また、実際にこれらのアルゴリズムを実装したアプリケーションを使った計算機シミュレーションをレポート課題を通して習得する。 From dynamics of galaxy to entanglement among quantum bits, many-body problems have been studied in the wide range of science by computational science approaches. In this lecture, numerical algorithms for many-body problems in condensed matter physics, such as Monte Carlo samplings and Krylov subspace methods represented by the Lanczos method, are introduced with emphasis on treatment of the large degrees of freedom inherent in the many-body problems. Through exercises by using these algorithms implemented in open source software for many-body problems, students will learn how to utilize them. Papers on application of the algorithms and software to selected many-body problems are assigned as requirements for a credit.

　多体問題は、銀河のダイナミクスから量子ビット間のエンタングルメントまで、自然科学研究の様々な場面に現れ、計算機を用いた計算科学の主な研究対象となっている。本講義では、物質科学を具体例として多体問題に現れる、モンテカルロ法や、ランチョス法などのクリロフ部分空間法など、多体問題が内包する巨大な自由度を有限の計算機資源で扱うためのアルゴリズムを学ぶ。 　また、実際にこれらのアルゴリズムを実装したアプリケーションを使った計算機シミュレーションをレポート課題を通して習得する。 From dynamics of galaxy to entanglement among quantum bits, many-body problems have been studied in the wide range of science by computational science approaches. In this lecture, numerical algorithms for many-body problems in condensed matter physics, such as Monte Carlo samplings and Krylov subspace methods represented by the Lanczos method, are introduced with emphasis on treatment of the large degrees of freedom inherent in the many-body problems. Through exercises by using these algorithms implemented in open source software for many-body problems, students will learn how to utilize them. Papers on application of the algorithms and software to selected many-body problems are assigned as requirements for a credit.

　多体問題は、銀河のダイナミクスから量子ビット間のエンタングルメントまで、自然科学研究の様々な場面に現れ、計算機を用いた計算科学の主な研究対象となっている。本講義では、物質科学を具体例として多体問題に現れる、モンテカルロ法や、ランチョス法などのクリロフ部分空間法など、多体問題が内包する巨大な自由度を有限の計算機資源で扱うためのアルゴリズムを学ぶ。 　また、実際にこれらのアルゴリズムを実装したアプリケーションを使った計算機シミュレーションをレポート課題を通して習得する。 From dynamics of galaxy to entanglement among quantum bits, many-body problems have been studied in the wide range of science by computational science approaches. In this lecture, numerical algorithms for many-body problems in condensed matter physics, such as Monte Carlo samplings and Krylov subspace methods represented by the Lanczos method, are introduced with emphasis on treatment of the large degrees of freedom inherent in the many-body problems. Through exercises by using these algorithms implemented in open source software for many-body problems, students will learn how to utilize them. Papers on application of the algorithms and software to selected many-body problems are assigned as requirements for a credit.