About this course

The course covers:

- Recap of linear algebra and analysis

- Matrix decompositions (eigendecomposition, SVD, Cholesky, low-rank approximations, etc.)

- Vector calculus (gradients of vector-valued functions, backpropagation and automatic differentiation, Taylor series, etc.)

- Probability and distributions (probability spaces, Bayes’ theorem, Gaussian distributions, change of variables, exponential family)

-Continuous optimisation (gradient descent, constrained optimisation, convex optimisation)

Learning outcome

This course provides an overview of mathematical methods and tools relevant to data science applications and equips students with the necessary background for advanced courses in machine learning and statistics.

  • Learning activities

    Daily lectures and complementary exercises.

  • Teaching support

    Supervised exercises.

  • Syllabus

    Deisenroth, Marc Peter, A. Aldo Faisal, and Cheng Soon Ong. Mathematics for machine learning. Cambridge University Press, 2020.

  • Prerequisites

    None

  • Recommended prerequisites

    Linear algebra and analysis knowledge, MATH121, MATH122 and MATH123 or equivalent

  • Assessment method

    Portfolio assessment. Active participation and sufficient hand in exercises each week. Attendance at lectures and exercise sessions on at least 10 out of 14 days. Pass/fail.

  • About use of AI

  • Teaching hours

    • Lectures: 30 hours
    • Exercise sessions: 45 hours

  • Preferential right

    Master in Data Science

  • Admission requirements

    REALFAG