Master
2024/2025
Bayesian Methods for Data Analysis
Type:
Compulsory course (Financial Technologies and Data Analysis)
Area of studies:
Applied Mathematics and Informatics
Delivered by:
Joint Department with Sberbank ‘Financial Technologies and Data Analysis’
Where:
Faculty of Computer Science
When:
2 year, 1 module
Mode of studies:
offline
Open to:
students of one campus
Master’s programme:
Финансовые технологии и анализ данных
Language:
English
ECTS credits:
3
Course Syllabus
Abstract
The Bayesian probability allows us to effectively take into account different user preferences when constructing decisive prediction rules. In addition, it allows solving the problems of selecting structural parameters of the model. In particular, it allows solving the problems of feature selection, dimensionality of the reduced space at dimensionality reduction, and values of regularization coefficients without combinatorial search. An important part of the course is the study of modern generative models that use the Bayesian approach.
Learning Objectives
- Forming knowledge, skills and abilities in working with complex probabilistic models that take into account the machine learning application structure, to derive the necessary formulas for solving learning and inference problems within the framework of the constructed probabilistic models, as well as to effectively implement them.
Expected Learning Outcomes
- to be able to operate complex probabilistic models
- to master generative modeling modern approaches
Course Contents
- Bayesian probability
- Bayesian probability model-building
- Sampling methods and generative models
Assessment Elements
- Practical assignment 1
- Practical assignment 2
- ExamStudents are permitted to use any materials during preparation for their answer. Students are not allowed to use anything during their answer (including their written down preparation notes). During the exam students have one hour to prepare their answer to the questions in the exam papers, after which they are to answer the examiner questions from the theoretical minimum, answer to the questions in the exam papers and additional questions about the course and solve the problems.
Interim Assessment
- 2024/2025 1st module0.25 * Exam + 0.375 * Practical assignment 1 + 0.375 * Practical assignment 2
Bibliography
Recommended Core Bibliography
- Barber, D. (2012). Bayesian Reasoning and Machine Learning. Cambridge: Cambridge eText. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=edsebk&AN=432721
- Deep learning, Goodfellow, I., 2016
- Pattern recognition and machine learning, Bishop, C. M., 2006
Recommended Additional Bibliography
- Tipping, M. E. (2001). Sparse Bayesian Learning and the Relevance Vector Machine. Journal of Machine Learning Research, 1(3), 211–244. https://doi.org/10.1162/15324430152748236