Abstract: In deep learning-based speaker verification frameworks, extraction of a speaker embedding vector plays a key role. In this contribution, we propose a hybrid neural network that employs a ...