Yaser Said Abu-Mostafa (Arabic: ياسر سعيد أبو مصطفى) is a professor of electrical engineering and computer science at the California Institute of Technology (Caltech). He is known for his research and educational contributions in Machine learning, including cofounding the Conference on Neural Information Processing Systems (NeurIPS) and authoring the textbook Learning From Data.
Abu-Mostafa was born in Egypt and pursued his undergraduate studies at Cairo University, earning a BSc degree in 1979. He then moved to the United States for graduate work, receiving an MS degree from the Georgia Institute of Technology in 1981 and a PhD from Caltech in 1983. He joined the Caltech faculty immediately after completing his doctorate and has remained there since, teaching courses on machine learning, computational complexity, and the theory of neural networks.
In 1987, Abu-Mostafa cofounded NeurIPS, which has grown into one of the most influential conferences in artificial intelligence and machine learning, attracting thousands of researchers annually. His own research interests have spanned computational learning theory, pattern recognition, and the application of machine learning to fields such as finance and bioinformatics.
Academic and Administrative Roles
Beyond his professorship, Abu-Mostafa serves as chairman of Paraconic Technologies Ltd and Machine Learning Consultants LLC, companies that commercialize machine learning solutions. These roles reflect his commitment to bridging academic research and industrial practice, a theme he emphasizes in his teaching and writing.
Textbook and Online Course
Abu-Mostafa is the author of Learning From Data, a widely used textbook that introduces the fundamental concepts of machine learning, including the bias-variance tradeoff, regularization, and model selection. The book is accompanied by a popular online course of the same name, which he developed and teaches through Caltech's online platform. The course has reached a global audience and is often cited as a starting point for students entering the field.
Contributions to Machine Learning
Abu-Mostafa's research has focused on the theoretical foundations of learning algorithms and their practical applications. He has published extensively on topics such as the VC dimension, learning with hints, and the use of computational learning theory to analyze neural networks. His work has influenced later developments in Deep learning and Neural network architectures, though he himself has remained more closely associated with the classical theory of learning from data.
Recognition and Influence
As a cofounder of NeurIPS, Abu-Mostafa helped establish a platform that later facilitated major breakthroughs in Artificial intelligence, including the rise of large-scale models. His textbook and online course have trained many practitioners and researchers who went on to contribute to the modern AI landscape. While he has not been as prominent in the recent Generative AI boom, his pedagogical materials continue to be used in university curricula worldwide.
Selected Works
- Learning From Data (with Malik Magdon-Ismail and Hsuan-Tien Lin), AMLBook, 2012.
- Numerous papers on computational learning theory and neural networks published in journals such as IEEE Transactions on Neural Networks and Neural Computation.
See Also
References
This article is based on publicly available biographical information from Yaser Abu-Mostafa's professional home page and conference records.
External Links
- Yaser S. Abu-Mostafa professional home page at Caltech.