Machine learning model weights and activations are represented in
full-p...
Attention-based contextual biasing approaches have shown significant
imp...
We present dual-attention neural biasing, an architecture designed to bo...
Personalization in multi-turn dialogs has been a long standing challenge...
For on-device automatic speech recognition (ASR), quantization aware tra...
The recurrent neural network transducer (RNN-T) is a prominent streaming...
We present a streaming, Transformer-based end-to-end automatic speech
re...
Conversational agents commonly utilize keyword spotting (KWS) to initiat...
Personal rare word recognition in end-to-end Automatic Speech Recognitio...
Dialogue act classification (DAC) is a critical task for spoken language...
End-to-end Spoken Language Understanding (E2E SLU) has attracted increas...
Recent years have seen significant advances in end-to-end (E2E) spoken
l...
We present Bifocal RNN-T, a new variant of the Recurrent Neural Network
...
As more speech processing applications execute locally on edge devices, ...
We introduce Amortized Neural Networks (AmNets), a compute cost- and
lat...
We propose a simple yet effective method to compress an RNN-Transducer
(...
End-to-end spoken language understanding (SLU) models are a class of mod...
In this paper we investigate statistical model compression applied to na...