Swiss German Speech-to-Text

Student: N.N.

Supervisor: Tanja Samardžić

Introduction

At this point, we have installed the most widely used open-source system for speech-to-text, Kaldi, and have performed its initial evaluation. The goal of this thesis is to evaluate the existing system using a larger data set and then improve the performance by modifying at least one component.

Requirements

Familiarity with speech processing, vector representations and deep learning.