Metraipolis › archive › dc_22d50a641cd7

Sparse Text Diffusion under Dual Budgets: Context Connectivity and Expert Activation — arXiv:2610.41487

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41487
License (as archived): CC0-1.0
Archived: 2026-11-02T07:21:01.348Z
SHA-256 of the archived text: aaa113bef1518e47f88929cf55f9a2d1d990e753f28aef47188deb0e6bf741cc
Stable URL of this page: https://metraipolis.com/documento/dc_22d50a641cd7

How to cite

Sparse Text Diffusion under Dual Budgets: Context Connectivity and Expert Activation — arXiv:2610.41487 — arXiv (en), https://arxiv.org/abs/2610.41487 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_22d50a641cd7 (Metraipolis archive dc_22d50a641cd7, sha256 aaa113bef1518e47f88929cf55f9a2d1d990e753f28aef47188deb0e6bf741cc)

BibTeX

@misc{dc22d50a641cd7,
  title = {Sparse Text Diffusion under Dual Budgets: Context Connectivity and Expert Activation — arXiv:2610.41487},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41487},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_22d50a641cd7; sha256 aaa113bef1518e47f88929cf55f9a2d1d990e753f28aef47188deb0e6bf741cc},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41487. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-context diffusion repeatedly pays for both token interaction and feed-forward computation. We formulate these costs as two separate budgets: the edges available to an attention graph and the experts activated for each token. This view connects local, dilated, and global attention with conditional expert computation without assuming that sparsity alone preserves document-level information. We specify normalized masked attention, distinguish denoising depth from sampling iterations, and derive graph and compute bounds. A reported sparse-expert configuration, denoted DualBudget, is analyzed across scientific summarization, multi-hop question answering, dialogue, and sentence-pair prediction. All supplied table measurements are retained. The length study shows gains over the listed baselines through 16K, together with measurable degradation as length increases. Ablations reveal mixed changes across the three reported scores, while the cost record indicates a 20% reduction relative to DiffuSeq under an unspecified timing convention. The analysis separates these observations from unmeasured claims about training speed, absorbing states, global coherence, and deployment memory.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.41487 — 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 2026-11-02T07:16:57.276Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).