Skip to main content

Module description - Efficient Machine Learning Algorithms
(Effiziente Machine Learning Algorithmen)

ECTS 3.0
Specification Analyze and implement efficient ML data structures and algorithms
Level Advanced
Content Machine learning algorithms convert mathematical models into code and thus enable the extraction of patterns from data, which can be used for predictions and other purposes. Computational efficiency is of great importance, since finding optimal model coefficients and hyperparameters often requires a large number of computational operations.

The implementation of machine learning algorithms is in essence the efficient use of numerical operations on suitable data structures. On the one hand, theoretical considerations regarding the number of operations and runtime can be made. On the other hand, practical framework conditions, such as the physical distribution of data across different storage media, the transport costs of data and computational instructions must be taken into account.
Learning outcomes Fundamentals of efficient numerical computation
Students understand theoretical runtime calculations and can use them as a basis for decision-making when choosing algorithms. They can analyse algorithms for their runtime. They are familiar with computer infrastructure and architectures. They know how computational operations can be optimally exploited and how this affects runtimes in practice.

Efficient algorithms and data structures
Students are familiar with methods and data structures of efficient programming:

Dynamic programming, greedy algorithms, memoisation, metaheuristics (genetic, ant colony, particle swarm optimisation, ... ), linear programming, and binary search trees.

Efficient implementation of ML algorithms and their analysis
Students can perform efficient implementations of standard ML algorithms: random forests, MLP, kMeans, kNN, ... and estimate and measure their computational efficiency / runtime.
Evaluation Mark
Built on the following competences Foundation in Machine Learning, Foundation in Programming, Programming Algorithms, Software Construction
Modultype Portfolio Module
Diese Seite teilen: