Investigating the Impact of ASR Errors on Spoken Implicit Discourse Relation Recognition

Linh The Nguyen, Dat Quoc Nguyen

Keywords: ASR Errors, Implicit Discourse Relation Recognition, IDRR, Spoken IDRR

TL;DR: "Investigating the Impact of ASR Errors on Spoken Implicit Discourse Relation Recognition" at the Transcript Understanding Workshop 2022

Abstract: We present an empirical study investigating the influence of automatic speech recognition (ASR) errors on the spoken implicit discourse relation recognition (IDRR) task. We construct a spoken dataset for this task based on the Penn Discourse Treebank 2.0. On this dataset, we conduct "Cascaded" experiments employing state-of-the-art ASR and text-based IDRR models and find that the ASR errors significantly decrease the IDRR performance. In addition, the "Cascaded" approach does remarkably better than an "End-to-End" one that directly predicts a relation label for each input argument speech pair.