T E A C H I N G

CS/CNS/EE 156 - Learning Systems

This course covers the theory, algorithms, and applications of Machine Learning and Artificial Intelligence. It is a subject that combines mathematical theory with heuristic techniques, and it is one of the most widely applicable subjects in engineering and scientific research as well as in practical applications from computational finance to recommender systems to medical applications to robotics to language models, among other fields.

The technical topics covered include linear models, theory of generalization, regularization and validation, Occam's razor and data snooping, neural networks, support vector machines, as well as specialized techniques. The second term of the course has focused on projects with huge datasets, e.g., Netflix prize data, COVID-19 infection data, and medical imaging data.

The course has recorded lectures and the material has evolved into a textbook entitled Learning From Data.

This course has more than 3,000 alumni from more than 20 different majors at Caltech, and more than a million online participants.

CS/EE/Ma 129 - Information and Complexity

This novel course covered information theory and computational complexity in a unified way. It developed the subject from first principles, building up from the basic premise of information to Shannon's information theory, and from the basic premise of computation to Turing's theory of computation. The duality between the two theories leads naturally to the theory of Kolmogorov complexity.

The technical topics covered included source coding, channel coding, rate-distortion theory, Turning machines, computability, computational complexity, and algorithmic entropy, as well as specialized topics and projects. These topics inspired a Watson Lecture about "randomness".

The course emphasized the basic understanding of the subject that enables the students to use the notions of information and complexity in their own research work. There are complete notes for the course that were made available to the students.

This course has more than 1,000 alumni.

EE 32/111 - Linear Systems and Signal Processing

This is a standard course on Fourier analysis and its application to linear systems and signal processing. In addition to the standard topics, e.g., Fast Fourier Transform and filtering, I also covered stochastic processes and power spectral density.

Magazine Articles

• Here is a magazine article about Professor Abu-Mostafa's teaching among other Feynman awardees, which appeared in Engineering & Science in the Fall of 2011.
• Here is an older, more comprehensive article in Caltech News.