Metraipolis › archive › dc_e242d2d4ddff
Rotated Manifold Optimization for Low-Rank Adaptation — arXiv:2609.34264
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.34264
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T22:38:07.041Z
- SHA-256 of the archived text
a274fec8bf2d00af6024f57015ffb148bc4d5d45ff3d3c51b518375d81e5e429- Stable URL of this page
https://metraipolis.com/documento/dc_e242d2d4ddff
How to cite
Rotated Manifold Optimization for Low-Rank Adaptation — arXiv:2609.34264 — arXiv (en), https://arxiv.org/abs/2609.34264 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e242d2d4ddff (Metraipolis archive dc_e242d2d4ddff, sha256 a274fec8bf2d00af6024f57015ffb148bc4d5d45ff3d3c51b518375d81e5e429)
BibTeX
@misc{dce242d2d4ddff,
title = {Rotated Manifold Optimization for Low-Rank Adaptation — arXiv:2609.34264},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.34264},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e242d2d4ddff; sha256 a274fec8bf2d00af6024f57015ffb148bc4d5d45ff3d3c51b518375d81e5e429},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34264. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We propose a novel optimizer for low-rank adaptation (LoRA) that explicitly incorporates the gauge symmetry of low-rank factorization. Our optimizer extends recent matrix optimizers for full-parameter training to the manifold of fixed-rank matrices by interpreting them as normalization under a rotated basis. We show how rotation and normalization can be integrated with the fixed-rank manifold efficiently. Our optimizer converges faster to lower held-out loss and achieves better or comparable downstream performance on both supervised finetuning and reinforcement learning tasks. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.LG). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.34264 — 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 2026-10-07T22:14:28.144Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).