BriefGPT.xyz
May, 2020
神经机器翻译推理校准
On the Inference Calibration of Neural Machine Translation
HTML
PDF
Shuo Wang, Zhaopeng Tu, Shuming Shi, Yang Liu
TL;DR
本文对神经机器翻译中置信度校准的问题进行研究，发现在推断时训练数据与推断数据的不同导致了校准不准确的问题。作者提出了一种新的 graduated label smoothing 方法来提高推断校准和翻译性能。
Abstract
confidence calibration
, which aims to make model predictions equal to the true correctness measures, is important for
neural machine translation
(NMT) because it is able to offer useful indicators of translation
→