Metraipolis › archive › dc_eb53ae60fb49
FraQ: Efficient Coordinate-Space Recompression for Federated Low-Rank Adaptation — arXiv:2608.03605
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.03605
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T12:16:28.929Z
- SHA-256 of the archived text
5920bcf3f039fec1777564671172362a376ff60b44db3bb0a7de19ba547f4564- Stable URL of this page
https://metraipolis.com/documento/dc_eb53ae60fb49
How to cite
FraQ: Efficient Coordinate-Space Recompression for Federated Low-Rank Adaptation — arXiv:2608.03605 — arXiv (en), https://arxiv.org/abs/2608.03605 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_eb53ae60fb49 (Metraipolis archive dc_eb53ae60fb49, sha256 5920bcf3f039fec1777564671172362a376ff60b44db3bb0a7de19ba547f4564)
BibTeX
@misc{dceb53ae60fb49,
title = {FraQ: Efficient Coordinate-Space Recompression for Federated Low-Rank Adaptation — arXiv:2608.03605},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.03605},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_eb53ae60fb49; sha256 5920bcf3f039fec1777564671172362a376ff60b44db3bb0a7de19ba547f4564},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.03605. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Federated fine-tuning with Low-Rank Adaptation (LoRA) enables efficient collaborative adaptation of Large Language Models (LLMs) without centralizing private data. However, LoRA's two-factor parameterization creates an aggregation mismatch across clients: naively averaging the factors does not recover the average of their induced updates. This mismatch can be avoided by forming the exact aggregate in the full weight space and then recompressing it, but decomposing the resulting dense matrix is computationally expensive and memory-intensive. We propose FraQ, an efficient coordinate-space recompression method for federated LoRA. Starting from stacked factors that exactly represent the aggregate, FraQ factorizes it into an orthonormal basis and a compact coordinate matrix. It then recovers the singular spectrum from a small Gram matrix, selects the smallest rank satisfying a prescribed energy threshold, and maps the selected coordinate subspace back through the basis to construct the global adapter. Experiments on text classification and commonsense reasoning benchmarks show that FraQ achieves accuracy close to uncompressed baselines while substantially reducing downlink communication with low server-side recompression overhead. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-05 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.03605 — 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 2026-10-31T12:05:52.820Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).