Название: Machine Learning for Engineers: Introduction to Physics-Informed, Explainable Learning Methods for AI in Engineering ApplicationsАвтор: Marcus J. Neuer
Издательство: Springer
Год: 2025
Страниц: 253
Язык: английский
Формат: pdf (true), epub
Размер: 79.7 MB
Machine Learning and Artificial Intelligence are ubiquitous terms for improving technical processes. However, practical implementation in real-world problems is often difficult and complex.
This textbook explains learning methods based on analytical concepts in conjunction with complete programming examples in Python, always referring to real technical application scenarios. It demonstrates the use of physics-informed learning strategies, the incorporation of uncertainty into modeling, and the development of explainable, trustworthy Artificial Intelligence with the help of specialized databases.
Therefore, this textbook is aimed at students of engineering, natural science, medicine, and business administration as well as practitioners from industry (especially data scientists), developers of expert databases, and software developers.
We will be working with Python throughout this book. All programs will be provided to you as so-called Jupyter Notebooks. In Appendix A, we suggest a suitable Python distribution and the associated programming environment. Appendix A also contains a compilation of basic knowledge in Python. It can help you get to grips with the language and offers an introduction for those who are completely unfamiliar with Python. However, we do assume some basic knowledge of programming languages and computer science for this book.
There are many programming languages that can be used to implement machine learning. Lua, C#, C++, Java, Matlab, and Julia are just a few examples. For applications in the technical field, it is particularly important that a language allows non-computer scientists to get started. Python has many advantages here:
For one thing, Python is easy to learn. Supported by a large community on the net, it is considered one of the best languages for programming beginners. It also has a wide syntactic range. Code can be written functionally or object-oriented, depending on the requirements of a project.
Python is an interpreted language. This means we can execute it directly without performing a compilation. With the help of Jupyter notebooks, we can directly check changes to the code for their effect, which promotes testing and simply playing with the code.
If you want to use your own learning methods in industry, it is easy to integrate solutions developed in Python into existing infrastructures. Almost every data management system of larger companies, manufacturing companies or even authorities offers interfaces to this programming language. It runs on both large computer centers and microcomputers like the Raspberry Pi.
The book is divided into three larger sections. Chapters 1 to 3 form a kind of introductory. They deal with the handling of data, mathematical tools to describe them, and ultimately methods to adapt them purposefully.
Chapters 4, 5 and 6 deal with Machine Learning, starting with supervised learning methods, moving on to unsupervised methods, and then to the idea of physics-informed learning. Each method is not only presented, but also supported with code examples. This application at the programming level is of great importance, as it deepens understanding and enables future applicability in the first place.
There are approaches that we have deliberately left out. Support Vector Machines (SVM), Kohonen's Self-Organizing Map (SOM), or even Restricted Boltzmann Machines (RBM) would have certainly fit into the context. However, a sufficiently detailed presentation would have exceeded the scope of this introduction and should therefore be revisited elsewhere in the future.
In all chapters, explainability and basic physical understanding are repeatedly addressed as themes. Chapter 7 concludes the book and addresses this aspect from various perspectives. We not only show helpful semantic tools to store context and knowledge references, but we also devote ourselves to the question of which data technologies and strategies support explainability.
Скачать Machine Learning for Engineers: Introduction to Physics-Informed, Explainable Learning Methods for AI