Description
This book provides an intuitive, hands-on introduction to PyTorch that connects modern machine learning with classical signal processing. It is accompanied by interactive Jupyter notebooks that tightly interleave explanations, formulas, visualizations, and runnable Python code, so readers can immediately experiment with the presented methods.
The material serves two purposes: it offers a structured path to learning PyTorch and training practical models, while revisiting core concepts such as convolution, filtering, gradient-based optimization, and classification. Instead of being a programming-only course that may quickly become outdated, it emphasizes durable conceptual understanding, explaining why methods work and how to adapt them to new problems. Centered on time-series and sequence data such as audio, it provides a clear alternative to image-centric introductions, using carefully designed toy examples and synthesized datasets that run efficiently on a standard CPU.
Aimed at advanced undergraduate and graduate students, as well as researchers and practitioners, the book supports both self-study and teaching. Each unit includes coding exercises with reference solutions, and the open-source notebooks can be used directly as course material or as a basis for projects and research.
Reviews
There are no reviews yet.