23: Grundlagen der Automatischen Spracherkennung, Vorlesung, WS 2017/18, 07.02.2018
Sebastian Stüker · 2018
23 | 0:00:00 Starten 0:00:52 Inkrementelle Adaption 0:03:12 Verschiedenes zu VTLN 0:09:00 Maximum Likelihood Linear Regression (MLLR) 0:18:25 Welche Adaption unter welchen Bedingungen? 0:21:27 Erinnerung MLLR 0:28:39 Bottleneck Features 0:33:48 Acoustic Model: Neural networks 0:41:07 Time-delay neural networks (TDNN) 0:47:31 Speaker adaptive neural networks 0:55:34 Neural network training: Potential problems? 1:05:53 Connectionist temporal classification (CTC) 1:23:21 Reaching "human parity"