Metraipolis › archive › dc_5c482bd409f4
Softmax Attention with Any Temperature in Fixed Dimension: Matching the Best Known Maximum-Inner-Product Bounds — arXiv:2611.06055
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.06055
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-05T18:22:39.373Z
- SHA-256 of the archived text
a71f08dfd8633193812dc5ab3da8b9b8325074dc951d63a6bf0a05971420ee18- Stable URL of this page
https://metraipolis.com/documento/dc_5c482bd409f4
How to cite
Softmax Attention with Any Temperature in Fixed Dimension: Matching the Best Known Maximum-Inner-Product Bounds — arXiv:2611.06055 — arXiv (en), https://arxiv.org/abs/2611.06055 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5c482bd409f4 (Metraipolis archive dc_5c482bd409f4, sha256 a71f08dfd8633193812dc5ab3da8b9b8325074dc951d63a6bf0a05971420ee18)
BibTeX
@misc{dc5c482bd409f4,
title = {Softmax Attention with Any Temperature in Fixed Dimension: Matching the Best Known Maximum-Inner-Product Bounds — arXiv:2611.06055},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.06055},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5c482bd409f4; sha256 a71f08dfd8633193812dc5ab3da8b9b8325074dc951d63a6bf0a05971420ee18},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.06055. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Attention is central to modern AI models for language, vision, and scientific modeling. However, direct evaluation compares every query with every key, requiring quadratic time in the context length, which limits its scalability to long contexts. Consequently, characterizing the complexity bounds of attention is of fundamental importance. For every fixed dimension $d$, we give an algorithm for $m$ queries, $n$ key-value pairs, and additive error $\varepsilon$, with expected time $ \widetilde{\mathrm{O}}\bigl((mn)^{k/(k+1)}+m+n\bigr),\qquad k=\lfloor d/2\rfloor. $ For fixed dimension and a fixed number of value channels, soft-O hides factors of the form $\operatorname{poly}_d\bigl(b,\log(m+n),\log B,\log(1/\varepsilon)\bigr),$ where $b$ bounds input encoding lengths and $B$ bounds coordinate and value magnitudes. In particular, the dependence on $B$ and $1/\varepsilon$ is polylogarithmic, which allows arbitrary positive temperature. When $m=n$ and $d\ge2$, the bound is $\widetilde{\mathrm{O}}(n^{2-2/(\lfloor d/2\rfloor+1)})$: near-linear in dimensions two and three, with exponent $4/3$ in dimensions four and five and $3/2$ in dimensions six and seven. It improves the previous $\widetilde{\mathrm{O}}(n^{2-1/d})$ upper bound and matches the best known batch maximum-inner-product bound in each fixed dimension. Additionally, we reduce integer batch maximum inner product computation to one attention call with the same dimension and input sizes $m,n$.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Data Structures and Algorithms (cs.DS) ; Computational Complexity (cs.CC); Computational Geometry (cs.CG); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.06055 — 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 2026-11-05T18:15:49.037Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).