+- MKLab (https://mklab.gr)
+-- Forum: [INDEX] (https://mklab.gr/forumdisplay.php?fid=1)
+--- Forum: ARTFICIAL INTELLIGENCE (AI) (https://mklab.gr/forumdisplay.php?fid=5)
+---- Forum: COURSES-LECTURES (https://mklab.gr/forumdisplay.php?fid=38)
+---- Thread: 18.408 Theoretical Foundations for Deep Learning (/showthread.php?tid=387)
18.408 Theoretical Foundations for Deep Learning - mklabgr - 06-15-2026
Summary
The page describes MIT course 18.408 (Algorithmic Aspects of Machine Learning) by Ankur Moitra, which focuses on building a rigorous algorithmic theory for machine learning. The course studies why modern ML methods work by designing algorithms with provable guarantees for core problems such as nonnegative matrix factorization, tensor decomposition, sparse coding, learning mixture models, matrix completion, graphical models, and robust statistics.
A central theme is that many of these problems are computationally hard in worst-case settings, so the course emphasizes choosing realistic models (often based on probability, geometry, or algebra) where efficient algorithms can be analyzed and justified. Students typically engage with advanced problem sets and a research-oriented final project connecting theory with modern ML applications.
COURSE PAGE