This paper introduces a novel Token-and-Duration Transducer (TDT)
architecture for sequence-to-sequence tasks. TDT extends conventional
RNN-Transducer architectures by jointly predicting both a token and its
duration, i.e. the number of input frames covered by the emitted token. This is
achieved by using a joint network with two outputs which are independently
normalized to generate distributions over tokens and durations. During
inference, TDT models can skip input frames guided by the predicted duration
output, which makes them significantly faster than conventional Transducers
which process the encoder output frame by frame. TDT models achieve both better
accuracy and significantly faster inference than conventional Transducers on
different sequence transduction tasks. TDT models for Speech Recognition
achieve better accuracy and up to 2.82X faster inference than RNN-Transducers.
TDT models for Speech Translation achieve an absolute gain of over 1 BLEU on
the MUST-C test compared with conventional Transducers, and its inference is
2.27X faster. In Speech Intent Classification and Slot Filling tasks, TDT
models improve the intent accuracy up to over 1% (absolute) over conventional
Transducers, while running up to 1.28X faster.

介绍了一种用于序列到序列任务的新型 Token-and-Duration Transducer (TDT) 框架，该框架通过联合预测令牌及其持续时间来扩展传统的 RNN-Transducer 架构，可以提高序列传导任务上的准确性和推断速度。

联合预测符号和持续时间实现高效的序列转换

Efficient Sequence Transduction by Jointly Predicting Tokens and  Durations

Edit-based approaches have recently shown promising results on multiple
monolingual sequence transduction tasks. In contrast to conventional
sequence-to-sequence (Seq2Seq) models, which learn to generate text from
scratch as they are trained on parallel corpora, these methods have proven to
be much more effective since they are able to learn to make fast and accurate
transformations while leveraging powerful pre-trained language models. Inspired
by these ideas, we present TST, a simple and efficient Text Simplification
system based on sequence Tagging, leveraging pre-trained Transformer-based
encoders. Our system makes simplistic data augmentations and tweaks in training
and inference on a pre-existing system, which makes it less reliant on large
amounts of parallel training data, provides more control over the outputs and
enables faster inference speeds. Our best model achieves near state-of-the-art
performance on benchmark test datasets for the task. Since it is fully
non-autoregressive, it achieves faster inference speeds by over 11 times than
the current state-of-the-art text simplification system.

本论文介绍了一种基于序列标记的简单，高效的文本简化系统 (TST), 使用预训练的 Transformer-based 编码器，在现有系统的训练和推断中进行简单的数据增强和调整，使其不太依赖于大量的平行训练数据，提供了更多对输出的控制并实现了更快的推断速度，目前取得了基准测试数据集任务的几乎最新性能。