BriefGPT.xyz
Jan, 2020
利用ASR N-best假设来提高口语理解
Improving Spoken Language Understanding By Exploiting ASR N-best Hypotheses
HTML
PDF
Mingda Li, Weitong Ruan, Xinyue Liu, Luca Soldaini, Wael Hamza...
TL;DR
本文介绍了一些简单而有效的模型，以从自动语音识别模块获取的N（N为一个任意数）-best解释中共同挖掘语音信号的语义信息，提高现代口语理解系统的性能。
Abstract
In a modern spoken language understanding (SLU) system, the
natural language understanding
(NLU) module takes
interpretations
of a speech from the automatic
→