Metraipolis › archive › dc_fb674a39d105
Precision As You Need: Stochastic Computing Is a Dense Adaptive Quantizer — arXiv:2609.32922
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.32922
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T20:56:16.793Z
- SHA-256 of the archived text
0b5339093682df003edf2c3eac9e50995f17d497fc80711eb49e6bc0f08406a8- Stable URL of this page
https://metraipolis.com/documento/dc_fb674a39d105
How to cite
Precision As You Need: Stochastic Computing Is a Dense Adaptive Quantizer — arXiv:2609.32922 — arXiv (en), https://arxiv.org/abs/2609.32922 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fb674a39d105 (Metraipolis archive dc_fb674a39d105, sha256 0b5339093682df003edf2c3eac9e50995f17d497fc80711eb49e6bc0f08406a8)
BibTeX
@misc{dcfb674a39d105,
title = {Precision As You Need: Stochastic Computing Is a Dense Adaptive Quantizer — arXiv:2609.32922},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.32922},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fb674a39d105; sha256 0b5339093682df003edf2c3eac9e50995f17d497fc80711eb49e6bc0f08406a8},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32922. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Matrix multiplications dominate the inference cost of modern transformer-based vision models, yet existing efficiency techniques such as post-training quantization and mixed-precision inference are largely limited to the small set of fixed-width formats (INT4, INT8, BF16, and FP16) supported by conventional accelerators. We revisit stochastic computing (SC) as a way to lift this constraint: viewed as a dense adaptive quantizer, SC controls precision by bit-stream length L rather than a fixed datapath, while each multiplication reduces to a single AND/XNOR gate. We build a GPU library that emulates SC matrix multiplication at scale, exposes stream lengths as first-class kernel arguments, and evaluates SC end-to-end on image classification, object detection and instance segmentation, class-conditional image generation, and visual world-model planning. On top of this substrate, we develop a dynamic per-row mixed-precision policy that assigns stream length per token or group at matched average budget, requires no retraining, and uses the same SC hardware across schedules. Across tasks, SC remains competitive with fixed-format INT quantization at matched bit budgets, while per-row mixed precision helps maintain accuracy at lower average stream lengths. These results provide software-level feasibility evidence that SC can serve as a dense-precision substrate for fine-grained mixed-precision inference on modern vision transformers. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.32922 — 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 2026-10-07T20:29:01.035Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).