ASR Error Management for Improving Spoken Language Understanding

Abstract

This paper addresses the problem of automatic speech recog- nition (ASR) error detection and their use for improving spo- ken language understanding (SLU) systems. In this study, the SLU task consists in automatically extracting, from ASR tran- scriptions, semantic concepts and concept/values pairs in a e.g touristic information system. An approach is proposed for en- riching the set of semantic labels with error specific labels and by using a recently proposed neural approach based on word embeddings to compute well calibrated ASR confidence mea- sures. Experimental results are reported showing that it is possi- ble to decrease significantly the Concept/Value Error Rate with a state of the art system, outperforming previously published re- sults performance on the same experimental data. It also shown that combining an SLU approach based on conditional random fields with a neural encoder/decoder attention based architec- ture, it is possible to effectively identifying confidence islands and uncertain semantic output segments useful for deciding appropriate error handling actions by the dialogue manager strategy.

Publication
18th Annual Conference of the International Speech (Interspeech 2017)