Metraipolis › archive › dc_77b45e1cb4e0
Spherical Interpolation for Backward-Compatible Multimodal Representations — arXiv:2609.39836
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.39836
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T13:47:47.341Z
- SHA-256 of the archived text
d3ab71520544930a82c58ac2d49c244bf770b8cc7cb9627dcff25b84631e68b4- Stable URL of this page
https://metraipolis.com/documento/dc_77b45e1cb4e0
How to cite
Spherical Interpolation for Backward-Compatible Multimodal Representations — arXiv:2609.39836 — arXiv (en), https://arxiv.org/abs/2609.39836 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_77b45e1cb4e0 (Metraipolis archive dc_77b45e1cb4e0, sha256 d3ab71520544930a82c58ac2d49c244bf770b8cc7cb9627dcff25b84631e68b4)
BibTeX
@misc{dc77b45e1cb4e0,
title = {Spherical Interpolation for Backward-Compatible Multimodal Representations — arXiv:2609.39836},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.39836},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_77b45e1cb4e0; sha256 d3ab71520544930a82c58ac2d49c244bf770b8cc7cb9627dcff25b84631e68b4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39836. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Contrastive vision-language models map visual and textual representations into a shared normalized embedding space, making cosine similarity the natural metric for cross-modal retrieval. A practical challenge arises during model upgrades: independently trained models generally produce incompatible representation spaces, so replacing a deployed model typically requires recomputing embeddings for the entire gallery, which is prohibitively expensive at scale. Orthogonal post-hoc alignment can partially mitigate this problem by mapping new-model queries into the old-model gallery space. However, because independently trained models can differ in fine-grained representation structure, the orthogonal alignment remains approximate, leaving a residual angular discrepancy between the old-model query and the aligned new-model query. We study whether interpolation along the spherical geodesic between these two normalized query representations can improve retrieval without re-indexing the gallery. We characterize when this path contains an interior query direction closer to an idealized retrieval-optimal direction than either endpoint, and connect this characterization to Recall@$K$ through a local margin-based certification result. Experiments across multiple benchmarks and model families show that post-alignment spherical interpolation improves over orthogonal alignment alone, recovering backward-compatibility in most evaluated settings. Consistent with our geometric characterization, per-query oracle analysis shows that retrieval-favorable interior points occur frequently in practice. Code is available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.LG). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.39836 — 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 2026-10-07T13:18:10.675Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).