TL;DR本文提出的一种新方法使用统一的 EM 框架共同优化 AR 和 NAR 模型,以有效地引导系统消除输出序列中的多模态,评估结果表明,该方法在机器翻译任务上实现了具有竞争力的性能,同时显著减少了推理延迟。
Abstract
Autoregressive (AR) models have been the dominating approach to conditional
sequence generation, but are suffering from the issue of high inference
latency. Non-autoregressive (NAR) models have been recently proposed to reduce
the latency by generating all output tokens in parallel but