Metraipolis › archive › dc_866382eac3ee

Change the Product, Keep the Parameters: Associative Algebra Layers for Transformers — arXiv:2609.32814

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.32814
License (as archived)
CC0-1.0
Archived
2026-10-07T22:13:25.081Z
SHA-256 of the archived text
62b93d570c48d9931b7e596f1e332fff50420bd5fd760daa09da89fefe8a5d63
Stable URL of this page
https://metraipolis.com/documento/dc_866382eac3ee

How to cite

Change the Product, Keep the Parameters: Associative Algebra Layers for Transformers — arXiv:2609.32814 — arXiv (en), https://arxiv.org/abs/2609.32814 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_866382eac3ee (Metraipolis archive dc_866382eac3ee, sha256 62b93d570c48d9931b7e596f1e332fff50420bd5fd760daa09da89fefe8a5d63)
BibTeX
@misc{dc866382eac3ee,
  title = {Change the Product, Keep the Parameters: Associative Algebra Layers for Transformers — arXiv:2609.32814},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.32814},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_866382eac3ee; sha256 62b93d570c48d9931b7e596f1e332fff50420bd5fd760daa09da89fefe8a5d63},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32814. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Fast matrix multiplication algorithms keep the product fixed and search for a cheaper way to evaluate it. We instead ask whether a Transformer's learned projections can use a different, cheaper product altogether. Building on an associative-algebra construction that replaces ordinary matrix multiplication with a sparser interaction table over the same weight blocks, we construct a family with quadratic arithmetic in the matrix dimension when the physical block size remains fixed, and derive finite-shape constraints for GPU execution. The construction is provably optimal for its bilinear rank by the Alder--Strassen bound and can be realized as row-typed rectangular projections compatible with causal masking and KV-cached decoding. We provide an empirical test of this approach by training two approximately 110M-parameter decoder-only Transformer LMs from the same recipe and 12.3B-token budget, differing only in their feed-forward layer: one uses ordinary dense matrix multiplication and the other uses the associative-algebra product. Across four prompt domains, the algebraic model achieves a 6.2--7.8\% increase in end-to-end generation throughput, while obtaining lower scores on all three reported downstream metrics. We treat these results as a feasibility and trainability check for the proposed approach at small scale, leaving further investigation to future work.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-09-29 in cs.PF).
Categorie arXiv: Machine Learning (cs.LG) ; Performance (cs.PF).
Pagina del lavoro: https://arxiv.org/abs/2609.32814 — 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 2026-10-07T21:54:46.126Z dalla pagina https://arxiv.org/catchup/cs.PF/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).