The integration of external personalized context information into
docume...
We improve on the popular conformer architecture by replacing the depthw...
We report on aggressive quantization strategies that greatly accelerate
...
We introduce two techniques, length perturbation and n-best based label
...
Large-scale distributed training of deep acoustic models plays an import...
We investigate the impact of aggressive low-precision representations of...
When recurrent neural network transducers (RNNTs) are trained using the
...
In the recent advances of natural language processing, the scale of the
...
Large-scale distributed training of Deep Neural Networks (DNNs) on
state...
Data privacy and protection is a crucial issue for any automatic speech
...
In Speech Emotion Recognition (SER), emotional characteristics often app...
Accurately and globally mapping human infrastructure is an important and...
The past decade has witnessed great progress in Automatic Speech Recogni...
Decentralized Parallel SGD (D-PSGD) and its asynchronous variant Asynchr...
Adaptive gradient algorithms perform gradient-based updates using the hi...
Generative Adversarial Networks (GANs) are powerful class of generative
...
There has been huge progress in speech recognition over the last several...
In automatic speech recognition (ASR), wideband (WB) and narrowband (NB)...
Evolutionary stochastic gradient descent (ESGD) was proposed as a
popula...
Modern Automatic Speech Recognition (ASR) systems rely on distributed de...
In this paper, we propose and investigate a variety of distributed deep
...
We propose a population-based Evolutionary Stochastic Gradient Descent (...
An embedding-based speaker adaptive training (SAT) approach is proposed ...
Learning with recurrent neural networks (RNNs) on long sequences is a
no...
One of the most difficult speech recognition tasks is accurate recogniti...