About

The course covers the main tasks of teaching by use cases: classification, clustering, regression, dimension reduction. The methods of their solution, both classical and new, created over the past 10-15 years, are being studied. The emphasis is on a deep understanding of the mathematical foundations, relationships, advantages and limitations of the methods under consideration. Theorems are mostly given without proofs.

Syllabus

• Linear classifier and stochastic gradient.
• Metric methods of classification and regression.
• Support vector machine.
• Linear regression. Nonlinear regression.
• Model selection criteria and feature selection methods.
• Logical classification methods.
• Linear ensembles. Advanced methods of ensembling.
• Density estimation and Bayesian classification.
• Clustering and partial learning.
• Deep neural networks.
• Deep neural networks for unsupervised learning.
• Vector representations of texts and graphs.
• Attention models and transformers.
• Topic modeling.
• Learning to rank.
• Recommendation systems.
• Search for associative rules.
• Adaptive forecasting methods.
• Incremental and online learning.
• Reinforcement learning.
• Active learning.

Labworks

No.

Grading

At the end of the course, students take an oral exam on all topics of the course.

Prerequisites

Probability theory, statistics, optimization methods, linear algebra.