Metraipolis › archive › dc_85f796c4484f
On the Geometry of Multimodal Saturation: Riemannian VICReg — arXiv:2610.06096
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.06096
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T05:27:27.793Z
- SHA-256 of the archived text
41e5ee27220ab3ee6d53fc5dc5d1b72af95f5806e07dc40d37927ab55d882106- Stable URL of this page
https://metraipolis.com/documento/dc_85f796c4484f
How to cite
On the Geometry of Multimodal Saturation: Riemannian VICReg — arXiv:2610.06096 — arXiv, https://arxiv.org/abs/2610.06096 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_85f796c4484f (Metraipolis archive dc_85f796c4484f, sha256 41e5ee27220ab3ee6d53fc5dc5d1b72af95f5806e07dc40d37927ab55d882106)
BibTeX
@misc{dc85f796c4484f,
title = {On the Geometry of Multimodal Saturation: Riemannian VICReg — arXiv:2610.06096},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.06096},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_85f796c4484f; sha256 41e5ee27220ab3ee6d53fc5dc5d1b72af95f5806e07dc40d37927ab55d882106},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06096. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
In self-supervised learning, a third modality should improve, or at least preserve, performance. Across nine image-text-tabular datasets, we show that it instead harms performance: the trimodal model underperforms its own best bimodal subset in 55.6% of paired runs under VICReg. The same failure occurs in 51.1% of paired runs under SimSiam. We call this failure multimodal saturation. We propose that the failure lies in the alignment geometry. Riemannian VICReg (R-VICReg) generalizes classical VICReg: it aligns views by squared geodesic distance on learnable negative-curvature product factors and recovers VICReg exactly as curvature vanishes. Over the same 45 paired runs, R-VICReg raises the probability that the third modality helps from 44.4% to 64.4%, with gains concentrated where VICReg saturates. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.06096 — 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 2026-10-06T05:24:06.122Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).