Investigation of Whisper ASR Hallucinations Induced by Non-Speech Audio

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Barański, Mateusz, Jasiński, Jan, Bartolewska, Julitta, Kacprzak, Stanisław, Witkowski, Marcin, Kowalczyk, Konrad
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916653215776768
author Barański, Mateusz
Jasiński, Jan
Bartolewska, Julitta
Kacprzak, Stanisław
Witkowski, Marcin
Kowalczyk, Konrad
author_facet Barański, Mateusz
Jasiński, Jan
Bartolewska, Julitta
Kacprzak, Stanisław
Witkowski, Marcin
Kowalczyk, Konrad
contents Hallucinations of deep neural models are amongst key challenges in automatic speech recognition (ASR). In this paper, we investigate hallucinations of the Whisper ASR model induced by non-speech audio segments present during inference. By inducting hallucinations with various types of sounds, we show that there exists a set of hallucinations that appear frequently. We then study hallucinations caused by the augmentation of speech with such sounds. Finally, we describe the creation of a bag of hallucinations (BoH) that allows to remove the effect of hallucinations through the post-processing of text transcriptions. The results of our experiments show that such post-processing is capable of reducing word error rate (WER) and acts as a good safeguard against problematic hallucinations.
format Preprint
id arxiv_https___arxiv_org_abs_2501_11378
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Investigation of Whisper ASR Hallucinations Induced by Non-Speech Audio
Barański, Mateusz
Jasiński, Jan
Bartolewska, Julitta
Kacprzak, Stanisław
Witkowski, Marcin
Kowalczyk, Konrad
Sound
Artificial Intelligence
Audio and Speech Processing
Hallucinations of deep neural models are amongst key challenges in automatic speech recognition (ASR). In this paper, we investigate hallucinations of the Whisper ASR model induced by non-speech audio segments present during inference. By inducting hallucinations with various types of sounds, we show that there exists a set of hallucinations that appear frequently. We then study hallucinations caused by the augmentation of speech with such sounds. Finally, we describe the creation of a bag of hallucinations (BoH) that allows to remove the effect of hallucinations through the post-processing of text transcriptions. The results of our experiments show that such post-processing is capable of reducing word error rate (WER) and acts as a good safeguard against problematic hallucinations.
title Investigation of Whisper ASR Hallucinations Induced by Non-Speech Audio
topic Sound
Artificial Intelligence
Audio and Speech Processing
url https://arxiv.org/abs/2501.11378