Towards Robust Generative Speech Enhancement Using Vector Quantisation-Based Neural Audio Codec | ArxivCSExplorer