TY - JOUR AB - In this paper, we derive an uncertainty decoding rule for automatic speech recognition (ASR), which accounts for both corrupted observations and inter-frame correlation. The conditional independence assumption, prevalent in hidden Markov model-based ASR, is relaxed to obtain a clean speech posterior that is conditioned on the complete observed feature vector sequence. This is a more informative posterior than one conditioned only on the current observation. The novel decoding is used to obtain a transmission-error robust remote ASR system, where the speech capturing unit is connected to the decoder via an error-prone communication network. We show how the clean speech posterior can be computed for communication links being characterized by either bit errors or packet loss. Recognition results are presented for both distributed and network speech recognition, where in the latter case common voice-over-IP codecs are employed. AU - Ion, Valentin AU - Haeb-Umbach, Reinhold ID - 11820 IS - 5 JF - IEEE Transactions on Audio, Speech, and Language Processing KW - automatic speech recognition KW - bit errors KW - codecs KW - communication links KW - corrupted observations KW - decoding KW - distributed speech recognition KW - error-prone communication network KW - feature vector sequence KW - hidden Markov model-based ASR KW - hidden Markov models KW - inter-frame correlation KW - Internet telephony KW - network speech recognition KW - packet loss KW - speech posterior KW - speech recognition KW - transmission error robust speech recognition KW - uncertainty decoding KW - voice-over-IP codecs TI - A Novel Uncertainty Decoding Rule With Applications to Transmission Error Robust Speech Recognition VL - 16 ER -