Reusing Latent Speech Representations for Query-Conditioned Topic Localization in Transcripts

Excerpt

arXiv:2609.21844v1 Announce Type: new 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 separat