Master's Thesis (June - December 2023)
Institute: Max Planck Institute of Animal Behavior (Konstanz, Germany)
Project Supervisor: Dr. Julian Schäfer-Zimmermann and Dr Ariana Strandburg-Peshkin
Bioacoustic research is critical for understanding biodiversity and animal communication, yet obtaining high-quality datasets is hindered by environmental variability and limited access. This study applies Prototypical Networks, a deep learning approach, to analyze meerkat vocalizations—a sparse and noisy dataset totaling nearly 200 hours. Building on the DCASE 2023 challenge framework, this research assesses the efficacy of Prototypical Networks in detecting bioacoustic events under challenging conditionsa and limited labels.


