Towards Multi-Level Transcript Segmentation: LoRA Fine-Tuning for Table-of-Contents Generation

  • Segmenting speech transcripts into thematic sections benefits both downstream processing and users who depend on written text for accessibility. We introduce a novel approach to hierarchical topic segmentation in transcripts, generating multi-level tables of contents that capture both topic and subtopic boundaries. We compare zero-shot prompting and LoRA fine-tuning on large language models, while also exploring the integration of high-level speech pause features. Evaluations on English meeting recordings and multilingual lecture transcripts (Portuguese, German) show significant improvements over established topic segmentation baselines. Additionally, we adapt a common evaluation measure for multi-level segmentation, taking into account all hierarchical levels within one metric.

Export metadata

Additional Services

Search Google Scholar
Metadaten
Author:Steffen Freisinger, Philipp Seeberger, Thomas Ranzenberger, Tobias BockletORCiD, Korbinian Riedhammer
DOI:https://doi.org/10.21437/Interspeech.2025-2792
ISSN:2958-1796
Parent Title (English):Interspeech 2025
Publisher:ISCA
Place of publication:ISCA
Document Type:conference proceeding (article)
Language:English
Reviewed:Begutachtet/Reviewed
Release Date:2025/11/04
Tag:hierarchical segmentation; spoken content segmentation; table of contents generation; topic segmentation
Pagenumber:5
First Page:276
Last Page:280
institutes:Fakultät Informatik
Zentrum für Künstliche Intelligenz (KIZ)
Research Themes:Digitalisierung & Künstliche Intelligenz
Verstanden ✔
Diese Webseite verwendet technisch erforderliche Session-Cookies. Durch die weitere Nutzung der Webseite stimmen Sie diesem zu. Unsere Datenschutzerklärung finden Sie hier.