Metraipolis › archive › dc_b2fa808a249b

BDNet: Fast-slow thinking is all you need — arXiv:2610.41485

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.41485
License (as archived)
CC0-1.0
Archived
2026-11-03T08:21:42.094Z
SHA-256 of the archived text
7bfa6d52239723bb11e917dfd4fcbfa59d4738c8e352e821e6f4472bca763104
Stable URL of this page
https://metraipolis.com/documento/dc_b2fa808a249b

How to cite

BDNet: Fast-slow thinking is all you need — arXiv:2610.41485 — arXiv (en), https://arxiv.org/abs/2610.41485 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b2fa808a249b (Metraipolis archive dc_b2fa808a249b, sha256 7bfa6d52239723bb11e917dfd4fcbfa59d4738c8e352e821e6f4472bca763104)
BibTeX
@misc{dcb2fa808a249b,
  title = {BDNet: Fast-slow thinking is all you need — arXiv:2610.41485},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41485},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b2fa808a249b; sha256 7bfa6d52239723bb11e917dfd4fcbfa59d4738c8e352e821e6f4472bca763104},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41485. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We present BDNet, a hierarchical image-classification backbone built around fast image features and slow statistical references. Its baseline deformation operator (BDO) transforms the deviation of a feature from a stage-specific reference through a low-rank projection and a learned channel scale. Rather than computing pairwise token similarity, the branch operates directly on baseline-relative residuals with complexity linear in the token count. Each of four reference buffers accumulates detached stage statistics during training and remains frozen at inference. A four-stage pyramid with depths $(2,2,6,2)$ and widths $(96,192,384,768)$ yields a 20.70M-parameter classifier at $224\times224$ resolution. We organize an ImageNet-1K study around clean classification, eight controlled ablations, and Gaussian-noise, contrast, and brightness perturbations. The central hypothesis is that a slow reference can improve the robustness of a compact representation while retaining a simple tensor implementation. We distinguish this training-time statistical mechanism from input-dependent attention and make its linear inference behavior explicit.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.41485 — 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 2026-11-03T08:06:38.832Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).