Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Sinhamahapatra, Supriti, Niehues, Jan
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2510.13979
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909849588072448
author Sinhamahapatra, Supriti
Niehues, Jan
author_facet Sinhamahapatra, Supriti
Niehues, Jan
contents State-of-the-art (SOTA) Automatic Speech Recognition (ASR) systems primarily rely on acoustic information while disregarding additional multi-modal context. However, visual information are essential in disambiguation and adaptation. While most work focus on speaker images to handle noise conditions, this work also focuses on integrating presentation slides for the use cases of scientific presentation. In a first step, we create a benchmark for multi-modal presentation including an automatic analysis of transcribing domain-specific terminology. Next, we explore methods for augmenting speech models with multi-modal information. We mitigate the lack of datasets with accompanying slides by a suitable approach of data augmentation. Finally, we train a model using the augmented dataset, resulting in a relative reduction in word error rate of approximately 34%, across all words and 35%, for domain-specific terms compared to the baseline model.
format Preprint
id arxiv_https___arxiv_org_abs_2510_13979
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Do Slides Help? Multi-modal Context for Automatic Transcription of Conference Talks
Sinhamahapatra, Supriti
Niehues, Jan
Artificial Intelligence
Computation and Language
State-of-the-art (SOTA) Automatic Speech Recognition (ASR) systems primarily rely on acoustic information while disregarding additional multi-modal context. However, visual information are essential in disambiguation and adaptation. While most work focus on speaker images to handle noise conditions, this work also focuses on integrating presentation slides for the use cases of scientific presentation. In a first step, we create a benchmark for multi-modal presentation including an automatic analysis of transcribing domain-specific terminology. Next, we explore methods for augmenting speech models with multi-modal information. We mitigate the lack of datasets with accompanying slides by a suitable approach of data augmentation. Finally, we train a model using the augmented dataset, resulting in a relative reduction in word error rate of approximately 34%, across all words and 35%, for domain-specific terms compared to the baseline model.
title Do Slides Help? Multi-modal Context for Automatic Transcription of Conference Talks
topic Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2510.13979