High-Quality Machine Learning solution to translate spoken German dialect into Official German ("Hochdeutsch")

back to overview

Type and Duration

Preproposal PhD-Thesis, since February 2024

Coordinator

Liechtenstein Business School

Main Research

Business Process Management

Description

As part of the research work, a solution based on machine learning and NLP - natural language processing - should be developed that can automatically translate spoken German dialect into Official German. This solution should be able to carry out proven translations with very high quality; the aim is to subsequently obtain certification as a trustworthy AI to allow it to be used in critical environments. The first step of the research work is about finding, analysing and evaluating suitable training data followed by finding, analysing and evaluating suitable models. In addition to identifying the best algorithms and models, it is also necessary to develop criteria and methods for evaluating the training data and the translation results. Only with the appropriate evaluations the quality of the training data and the finished solution can be objectively proven.

PhD-Student

Rene Pilz, MSc

Supervisor

Assoz. Prof. Dr. Johannes Schneider