Reusing Latent Speech Representations for Query-Conditioned Topic Localization in Transcripts
Mirrored from arXiv — NLP / Computation & Language for archival readability. Support the source by reading on the original site.
Computer Science > Computation and Language
Title:Reusing Latent Speech Representations for Query-Conditioned Topic Localization in Transcripts
Abstract:Long transcripts are costly inputs for downstream NLP systems and often contain irrelevant context. We study query-conditioned topic localization: predicting the sentence span in a transcript that best addresses a topic-title query. To improve span localization, we reuse ASR encoder states as sentence-level representations and fuse them with textual embeddings. This lets lightweight span locators exploit speech information without running a separate audio encoder. Experiments on two public datasets show consistent gains over text-only baselines, especially under strict boundary-matching criteria. Cross-dataset experiments further indicate that the benefits are strongest for structured or semi-structured speech, while gains on spontaneous speech are limited and mixed.
| Comments: | Accepted at EMNLP 2026 Main Conference |
| Subjects: | Computation and Language (cs.CL); Audio and Speech Processing (eess.AS) |
| ACM classes: | I.2.7; H.3.3 |
| Cite as: | arXiv:2609.21844 [cs.CL] |
| (or arXiv:2609.21844v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2609.21844
arXiv-issued DOI via DataCite (pending registration)
|
Submission history
From: Steffen Freisinger [view email][v1] Fri, 18 Sep 2026 14:41:30 UTC (2,316 KB)
Access Paper:
- View PDF
- HTML (experimental)
- TeX Source
Current browse context:
References & Citations
Bookmark
Bibliographic and Citation Tools
Code, Data and Media Associated with this Article
Demos
Recommenders and Search Tools
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
More from arXiv — NLP / Computation & Language
-
A Mechanistic Study of AI-Text Detection Neurons in Frozen BERT: Sparse Probing and Activation Patching on RAID
Sep 28
-
Manifold Projection and Iterative Autoencoder Refinement for Masked Language Modeling
Sep 28
-
Not All Memories Are Equal: Hierarchical Collaborative Memory for Validity-Aware Retrieval in LLM Agents
Sep 28
-
Auditing and Repairing LLM-as-Judge Failures in a Production Text-to-SQL Pipeline
Sep 28
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.