Oct 24, 2020 · Abstract. This paper presents an approach to long-context end-to-end automatic speech recognition. (ASR) using Transformers, aiming at ...

This paper presents an approach to long-context end-to-end automatic speech recognition (ASR) using Transformers, aiming at improving ASR accuracy for long ...

A Transformer-based architecture that accepts multiple consecutive utterances at the same time and predicts an output sequence for the last utterance is ...

Most end-to-end systems are basically designed to recognize independent utterances. • However, contextual information over multiple utterances, ...

Apr 19, 2021 · This paper addresses end-to-end automatic speech recognition (ASR) for long audio recordings such as lecture and conversational speeches.

May 18, 2022 · In this work, Transformer models and an end-to-end model based on connectionist temporal classification were considered to build a system for automatic ...

We present a novel large-context end-to-end automatic speech recognition (E2E-ASR) model and its effective training method based on knowledge distillation.

First, to the best of our knowledge, this is the first work to efficiently model attention pooling compressed lowdimensional cross utterance contexts in ...

Jul 2, 2022 · Transformer-based models have demonstrated their effective- ness in automatic speech recognition (ASR) tasks and even.

It is demonstrated that the extended Transformer provides state-of-the-art end-to-end ASR performance, and the new decoding method reduces decoding time by ...