M.Sc. Francesca Diedolo
Technische Universität München
Lehrstuhl für Nachrichtentechnik (Prof. Kramer)
Postadresse
Postal:
Theresienstr. 90
80333 München
Biografie
- Seit Februar 2021 wissenschaftliche Mitarbeiterin am Lehrstuhl für Nachrichtentechnik der TUM
- M.Sc. in Communications Engineering an der TUM, 2018-2020
- B.Sc. in Elektro- und Informationstechnik an der Università degli Studi di Padova, 2015-2018
Lehre
Machine Learning for Communications(WS 21/22, WS 22/23, WS 23/24)
Abschlussarbeiten
Angebotene Abschlussarbeiten
Laufende Abschlussarbeiten
Probabilistic Principal Component Analysis
Beschreibung
Principal component analysis (PCA) [1] is a famous technique for feature extraction and dimensionality reduction. Despite its widespread use and effectiveness, PCA is constrained by its linear nature.
Probabilistic PCA [2] improves upon classical PCA by linking it with maximum likelihood estimation based on a probability density model of the observed data.
The task of the student is to read about probabilistic PCA. The student should understand the underlying theoretical concepts of the algorithm, its application spectrum, as well as its limitations.
[1] https://arxiv.org/pdf/1404.1100.pdf
[2] https://www.cs.columbia.edu/~blei/seminar/2020-representation/readings/TippingBishop1999.pdf
Voraussetzungen
Information Theory, Linear Algebra
Betreuer:
Publikationen
2024
2022
- Nonlinear Equalization for Optical Communications Based on Entropy-Regularized Mean Square Error. European Conference on Optical Communication (ECOC), 2022 mehr… BibTeX
- Nonlinear Component Equalization: A Comparison of Deep Neural Networks and Volterra Series. European Conference on Optical Communication (ECOC), 2022 mehr… BibTeX
2021
- Label Extension for 32QAM: The Extra Bit for a Better FEC Performance-Complexity Tradeoff. 2020 European Conference on Optical Communications (ECOC), IEEE, 2021 mehr… BibTeX Volltext ( DOI )