Deploying an automatic speech recognition system with reasonable performance requires
expensive and time-consuming in-domain transcription. Previous work demonstrated that non-…

We measure the effects of a weak language model, estimated from as little as 100k words of
text, on unsupervised acoustic model training and then explore the best method of using …

Engineered proteins generally must possess a stable structure in order to achieve their
designed function. Stable designs, however, are astronomically rare within the space of all …

Previous work on self-training of acoustic models using unlabeled data reported significant
reductions in WER assuming a large phonetic dictionary was available. We now assume only …

Mechanical Turk is useful for generating complex speech resources like conversational
speech transcription. In this work, we explore the next step of eliciting narrations of Wikipedia …

We propose a framework to modularize the training of neural language models that use diverse
forms of sentence-external context (including metadata) by eliminating the need to jointly …

End-to-end automatic speech recognition (ASR) systems, such as recurrent neural network
transducer (RNN-T), have become popular, but rare word remains a challenge. In this paper, …

Language modeling (LM) for automatic speech recognition (ASR) does not usually incorporate
utterance level contextual information. For some domains like voice assistants, however, …

The acoustic emanations from keyboards provide a side-channel attack from which an
attacker can recover sensitive user information, such as passwords and personally identifiable …

Useful training data for automatic speech recognition systems of colloquial speech is usually
limited to expensive in-domain transcription. Broadcast news is an appealing source of …