Metraipolis › archive › dc_ccf598d96bdc
A Nuclear-Norm Lower Bound for Dithered Scalar Quantization of Matrix Products — arXiv:2609.05641
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.05641
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T13:58:40.599Z
- SHA-256 of the archived text
ba3d61c42b65e9a0ef7e885340a039c6fcf593fc6ab399cf386484b1d96c6e57- Stable URL of this page
https://metraipolis.com/documento/dc_ccf598d96bdc
How to cite
A Nuclear-Norm Lower Bound for Dithered Scalar Quantization of Matrix Products — arXiv:2609.05641 — arXiv (en), https://arxiv.org/abs/2609.05641 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ccf598d96bdc (Metraipolis archive dc_ccf598d96bdc, sha256 ba3d61c42b65e9a0ef7e885340a039c6fcf593fc6ab399cf386484b1d96c6e57)
BibTeX
@misc{dcccf598d96bdc,
title = {A Nuclear-Norm Lower Bound for Dithered Scalar Quantization of Matrix Products — arXiv:2609.05641},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.05641},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ccf598d96bdc; sha256 ba3d61c42b65e9a0ef7e885340a039c6fcf593fc6ab399cf386484b1d96c6e57},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.05641. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We consider the problem of minimizing error in quantized matrix multiplication $C=AB$. Scalar quantization of the factors introduces rounding errors whose scale depends on the maximum absolute entries -- the ranges -- of their rows and columns. These ranges determine the quantization grid steps. To reduce the error, we optimize over product-preserving transformations that alter the factor ranges and grid steps without changing $C$. Specifically, we seek the smallest leading expected squared error over invertible inner changes of basis and orthogonal outer rotations. Under independent, zero-mean subtractive dither noise on an unbounded lattice, we prove the output-only bound $E_{\rm lead} \ge (c_A+c_B)/K \Vert AB\Vert_*^2$, where $K$ is the inner dimension, $c_A$ and $c_B$ are normalized noise variances, and $\Vert AB\Vert_*$ is the nuclear norm. The bound is tight: an SVD-aligned Hadamard construction attains the infimum whenever a Hadamard matrix of order $K$ exists, including every power of two, while an SVD-aligned DCT construction is within a factor of two for every $K$. Without outer rotations, Gram-matrix balancing minimizes factorization energy, and finite-set flattening achieves the bound within $C\log(K(m+n))$. For power-of-two $K$, conditional expectations deterministically select the Hadamard signs in $O((m+n)K^2)$ exact-real operations. Synthetic experiments verify both constructions and illustrate the tradeoff between regularization and conditioning. These results characterize the full-gauge optimum and quantify the cost of preserving row and column indices.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-09 in cs.LG).
Categorie arXiv: Information Theory (cs.IT) ; Machine Learning (cs.LG); Numerical Analysis (math.NA).
Pagina del lavoro: https://arxiv.org/abs/2609.05641 — 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 2026-10-16T13:39:43.121Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).