Metraipolis › archive › dc_a95078d77e59

CORA: A Protocol for Diagnosing Boundary Robustness in Text-to-Audio Retrieval under Query Reformulations — arXiv:2609.33433

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33433
License (as archived)
CC0-1.0
Archived
2026-10-07T21:54:03.551Z
SHA-256 of the archived text
cd979ddbf3448ca6077e00f0b52d8af897bfa8b5dc6b9accea754383d19786d7
Stable URL of this page
https://metraipolis.com/documento/dc_a95078d77e59

How to cite

CORA: A Protocol for Diagnosing Boundary Robustness in Text-to-Audio Retrieval under Query Reformulations — arXiv:2609.33433 — arXiv (en), https://arxiv.org/abs/2609.33433 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a95078d77e59 (Metraipolis archive dc_a95078d77e59, sha256 cd979ddbf3448ca6077e00f0b52d8af897bfa8b5dc6b9accea754383d19786d7)
BibTeX
@misc{dca95078d77e59,
  title = {CORA: A Protocol for Diagnosing Boundary Robustness in Text-to-Audio Retrieval under Query Reformulations — arXiv:2609.33433},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33433},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a95078d77e59; sha256 cd979ddbf3448ca6077e00f0b52d8af897bfa8b5dc6b9accea754383d19786d7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33433. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Text-to-Audio (T2A) retrievers are typically evaluated with caption style queries, but the same user intent can be expressed in many forms. We introduce CORA (Caption-Offset Retrieval for Audio), a caption anchored diagnostic protocol that rewrites each source caption into five intent preserving forms (Command, Question, Indirect, Key phrase, and Statement) while fixing the target audio. By tracking the same target across query forms, CORA defines RankDrop, a metric revealing failures hidden by Recall@k. Using Pearson's correlation coefficient r, we find that RankDrop is weakly associated with raw text space movement (r=0.084), but strongly associated with Target Alignment Loss and Target Boundary Margin Degradation (r=0.508 and r=0.615). The same pattern appears in OEA retrievers, where RankDrop is better explained by boundary degradation (r=0.472/0.478) than by query movement (r=0.084/0.046). Overall, these results suggest that robust T2A retrieval requires preserving the target's boundary advantage over competing audio under reformulation.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.CL).
Categorie arXiv: Sound (cs.SD) ; Computation and Language (cs.CL); Audio and Speech Processing (eess.AS).
Pagina del lavoro: https://arxiv.org/abs/2609.33433 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T21:27:08.769Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).