Designing a Zero Textbook Machine Learning Course

This course was designed as a Zero Textbook Course (ZTC) with the goal of making machine learning education more accessible, transparent, and equitable. Commercial textbooks in computer science—especially in rapidly evolving fields like machine learning—can be expensive, quickly outdated, and inaccessible to many students. By relying entirely on Open Educational Resources (OER), this course removes cost barriers while allowing students to engage with current, high-quality materials.

Machine learning is particularly well-suited to open pedagogy. Many of the most influential textbooks, research papers, software libraries, and datasets in the field are already openly available. This course takes advantage of that ecosystem by using open textbooks such as An Introduction to Statistical Learning and Dive Into Deep Learning, along with open-source tools like Python, Scikit-learn, and PyTorch. Students work directly with real, publicly available datasets rather than curated or proprietary examples.

Designing the course as ZTC also supports active and reproducible learning. Instead of passively reading a single textbook, students learn by experimenting with code, documenting their work in Jupyter notebooks, and sharing reproducible projects. Open resources allow students to revisit materials after the course ends, reuse code responsibly, and continue learning beyond the classroom.

This site was created on the CUNY Academic Commons as part of the Open Knowledge Fellowship (Winter 2026). By sharing the course publicly under an open license, the aim is not only to support students in this class, but also to contribute to a broader community of educators interested in open, accessible approaches to teaching machine learning.