CS-E4740 - Federated Learning D, Lecture, 24.2.2025-28.5.2025

Guest access
Please note! Course description is confirmed for two academic years, which means that in general, e.g. Learning outcomes, assessment methods and key content stays unchanged. However, via course syllabus, it is possible to specify or change the course execution in each realization of the course, such as how the contact sessions are organized, assessment methods weighted or materials used.

LEARNING OUTCOMES

After completing the course, students 

* can model federated learning (FL) applications using graphs

* are familiar with total variation (TV) minimization as a flexible design principle for FL algorithms

* can design federated learning algorithms by applying optimizatoin methods to TV minimization

* can analyze computational and statistical properties of FL algorithms

Credits: 5 - 10

Schedule: 24.02.2025 - 28.05.2025

Teacher in charge (valid for whole curriculum period):

Teacher in charge (applies in this implementation): Alex Jung

Contact information for the course (applies in this implementation):

CEFR level (valid for whole curriculum period):

Language of instruction and studies (applies in this implementation):

Teaching language: English. Languages of study attainment: English

CONTENT, ASSESSMENT AND WORKLOAD

Content

  • valid for whole curriculum period:

    * machine learning basics: data, model, loss, regularization, multi-task learning, semi-supervised learning, transfer learning

    * network basics: graphs and their matrices, community/cluster structure  

    * TV minimization as a flexible design principle for FL

    * main flavors of FL (centralized, clustered, personalized) as special cases of TV minimization

    * distributed optimization: models for distributed computation, fixed-point iterations, gradient-based methods

Assessment Methods and Criteria

  • valid for whole curriculum period:

    Assignments; Project-Reports; Peer-Grading

Workload

  • valid for whole curriculum period:

    Independent Study; Assignments; Project-Work; Peer-Grading

DETAILS

Study Material

  • valid for whole curriculum period:

    * A. Jung, "Machine Learning: The Basic", Springer, 2022 http://mlbook.cs.aalto.fi 

    * A. Jung, "Introductory Lectures on Federated Learning,", Lecture Notes, 2023, https://github.com/alexjungaalto/FederatedLearning/blob/main/material/FL_LectureNotes.pdf

    * Youtube Channel: https://www.youtube.com/channel/UC_tW4Z_GfJ2WCnKDtwMuDUA

Substitutes for Courses
Prerequisites
SDG: Sustainable Development Goals

    1 No Poverty

    2 Zero Hunger

    3 Good Health and Well-being

    5 Gender Equality

    6 Clean Water and Sanitation

    7 Affordable and Clean Energy

    12 Responsible Production and Consumption

    13 Climate Action

    14 Life Below Water

    15 Life on Land

    16 Peace and Justice Strong Institutions

    17 Partnerships for the Goals

FURTHER INFORMATION

Further Information

  • valid for whole curriculum period:

    Teaching Language: English

    Teaching Period: 2024-2025 Spring IV - V
    2025-2026 Spring IV - V

    Registration:

    You must have completed a Bsc level course on

    * Linear Algebra such as MS-A0001 - Matrix Algebra (or equivalent)

    * Machine Learning such as CS-C3240 - Machine Learning (or equivalent)

    * Data Analysis with Python (see the MOOC by University of Helsinki) 

Rules of Conduct - Students´ Rights and Responsibilities (in force from 1 August 2020)

Aalto University Code of Academic Integrity and Handling Violations Thereof