… novel automatic speech recognition (ASR) framework called Integrated Source-Channel and
… of traditional systems based on the noisy source-channel model (SC) and end-to-end style …

… applying a model-based source separation mask to the output of a beamformer that combines
the source signals recorded … The source separation algorithm MESSL (Model-based EM …

… a universal speech model from a general corpus of speech and … model to separate speech
from other sound sources. This model is used in lieu of a speech model trained on speaker-…

… In this paper the Bayesian framework of speech recognition was reformulated for the
server side of a distributed system. This resulted in a predictive decision rule which was …

… -based methods, use explicit models of the speech sources to separate them [7]–[10]. The
model for each source in a given frame (a small window of speech of approximately 40 ms) is …

… -separation by predicting multiple speaker and noise masks for each channel. Then a
multi-source neural beamformer is used to spatially separate multiple speaker sources. In the last …

… for channel modeling of an Automatic Speech Recognition (ASR) system. This can have
implications in improving speech recognition … of the source language in the translation model. …

… technique for a distributed speech recognition system over IP. … of the information contained
in the data-source, because, in IP … previous and next received speech vector sequences, we …

… threat” system for speech recognition, where LSTMs drive … array processing, acoustic
modeling, and language modeling. This … source was significantly larger than the other source. For …

… (AM) training, we used approximately 250 hours of retransmitted conversational speech (LDC2011E111
and LDC2011E93). For language model (LM) training, we used various sources…