Efficient Extraction of Noise-Robust Discrete Units from Self-Supervised Speech Models | IEEE Conference Publication | IEEE Xplore