PaddlePaddle Deep Learning in Practice
Since this is framed as a hands-on book, each knowledge point comes with code for reference, which is its biggest practical strength.
The opening sections cover the kind of material commonly found in AI textbooks. If you are not yet familiar with the background of artificial intelligence or the overall technical landscape, those chapters are worth reading as orientation.
The later part shifts into project-based presentations of deep learning tasks. The emphasis is on mathematical code and model structure, and it gives fairly complete explanations of common tasks and the deep learning models associated with them. That makes it useful as a targeted reference when you run into a specific task and want to study it in context.
The limitation is also obvious: there are not that many situations where PaddlePaddle is the default tool. If you really want to get full value from the book, or genuinely learn PaddlePaddle itself, you have to deliberately create opportunities to use it.
The First Book on Self-Driving
This is a very good introductory book. It walks through the hardware and technical components needed for autonomous driving and explains them clearly.
What stands out is that it also includes a lot of practical experience from real autonomous driving system design, which is not something every introductory book gives you. It works especially well for beginners who want to understand the industry as a whole and get a feel for the major unsolved problems that still matter.
ROS Robot Programming and a Guide to SLAM Algorithms
The technology here feels slightly dated. Its role is less about presenting the current frontier and more about showing some of the code and tools used in the field.
Because of that, it works better as a review book, or as a source of supplementary examples alongside other, more specialized books.
ROS Robot Programming Practice
My first reaction to this book was that it is a treasure house—a complete one. It seems to contain everything you might want to learn about ROS.
The important thing is that the material forms a whole. You really need to work through the book in full to get at the essence of ROS. In other words, if you only skim it casually and try to bluff your way through, the book will return the favor and give you only a superficial understanding.
This is the kind of book to attack when you have a lot of time and energy available and can focus hard on it. A reasonable pace would be something like 50 pages a day, while reading the code carefully and, ideally, having an actual robot hardware platform you can use for hands-on practice.
Fourteen Lectures on Visual SLAM
This one is math-intensive, but it explains visual SLAM progressively, from the basics to deeper ideas.
The right way to study it is simple: read it seriously and make sure you truly understand every concept. The goal is not just vague familiarity, but being able to explain each idea back in your own words.
Robot Perception
This feels like the unofficial fifteenth lecture after Fourteen Lectures on Visual SLAM. Its focus is much more concentrated: factor graphs, or more plainly, the role of probability in robotic SLAM tasks.
It is a relatively short book, and it makes sense as extension reading after finishing Fourteen Lectures on Visual SLAM.
Interpretable Machine Learning
This book covers many machine learning concepts, so if your fundamentals are already strong, you can selectively skip some chapters.
Its emphasis is on ideas. It shows how the academic world explains and evaluates interpretability, and it introduces some models that are considered interpretable.
That said, my personal feeling is that this treatment of interpretable learning is a bit too detached from practical reality. It stays at the theoretical level more than I would like.
So this is a book that should be read with questions in mind. It is more useful if you treat it as a way to build your own framework for understanding interpretable machine learning, rather than expecting it to hand you a ready-made practical answer.
Human Anatomy Principles and Drawing Instruction
This is a very new book—published in October 2021. Since it was also my first time reading this kind of book, I do not feel fully confident making a strong judgment on it.
What does seem important while reading is method. You really should use a ruler to compare proportions, observe carefully, and organize your own drawing logic. The exercises on each page also need to be done as you go. At this stage, the point is not to polish finished artwork, but to sketch freely and build a feel for drawing first. Refinement can come later.