Dipl.-Ing. Vincent Latzko

PhD Researcher

Vincent Latzko received his Diploma degree in Electrical Engineering and Information Technology from Technische Universität Darmstadt, with a focus on the fundamental theory of electromagnetic fields (www.temf.de) and the numerical solution of Maxwell’s equations using a multitude of mathematical and algorithmic methods (such as DG-FEM, FDTD or BEM/Integral Equation solvers). He worked extensively in fundamental research at one of Europe’s largest particle accelerator facilities, the GSI Helmholtzcentre for Heavy Ion Research (www.gsi.de).

In his diploma thesis “Robust Time of Flight Denoising using Regression Tree Fields” with Prof. PhD Carsten Rother at the computer vision lab dresden he looked at a modern machine learning approach to restore distorted depth maps using non-parametric Gaussian Conditional Random Fields.
He afterwards focused on analytical approaches for regression problems as well as on data driven, deep learning models.

Contact

Phone: +49 351 463-35337
Email: vincent.latzko@tu-dresden.de

Research Interests

• Machine Learning
• Compression
• Software Defined Networks
• Network Slicing
• Network Function Virtualisation

Teaching

WS 18/19

• Problem Based Learning
• Oberseminar
• Mentoring / Grundlagen ET

SS 18

• Traffic Theory
• Electrical and Magnetical Fields

WS 17/18

• Kommunikationsnetze 3/Problem Based Learning
• Mentoring / Grundlagen ET

Supervision

• Theme Improving Next-Generation Video Codec Segmentation using Deep Learning Techniques (Cheng Chiang Huang)
• Learning Optimal Congestion Control Algorithms (Christian Vielhaus)
• Optimising and Learning Coding Parameters (Christian Vielhaus)
• Deep Reinforcement Learning for Traffic Control (Johannes Busch)
• Telepresence Demonstrator – Remote Controlled Car via Augmented Reality (Renbing Zhang, joint supervision with Alexander Kropp)
• Exploring resilience in-car using multipath SDN (Florian Kemser)
• Learning Coding Schemes (Alexander Sarmanow)

Publications