Metraipolis › archive › dc_6f09d503beda
Neural Networks with Differentiable Belief-Function Layers — arXiv:2610.30173
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.30173
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-25T21:33:18.550Z
- SHA-256 of the archived text
11a24fe19a162efb1a32f42b315d05d60d8f5849474a25b62f15f59681b3d310- Stable URL of this page
https://metraipolis.com/documento/dc_6f09d503beda
How to cite
Neural Networks with Differentiable Belief-Function Layers — arXiv:2610.30173 — arXiv (en), https://arxiv.org/abs/2610.30173 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6f09d503beda (Metraipolis archive dc_6f09d503beda, sha256 11a24fe19a162efb1a32f42b315d05d60d8f5849474a25b62f15f59681b3d310)
BibTeX
@misc{dc6f09d503beda,
title = {Neural Networks with Differentiable Belief-Function Layers — arXiv:2610.30173},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.30173},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6f09d503beda; sha256 11a24fe19a162efb1a32f42b315d05d60d8f5849474a25b62f15f59681b3d310},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.30173. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Yager's 1987 rule of combination $\boxplus$ was proposed on epistemic grounds: under high conflict, mass should be redistributed to ignorance rather than renormalised onto the class simplex. We observe that backpropagation needs the same rule: Dempster's normaliser $(1{-}\kappa)^{-1}$, where $\kappa$ is the conflict mass, puts a $(1{-}\kappa)^{-2}$ term in the gradient that diverges exactly when evidence sources disagree, while Yager's $\boxplus$ is bilinear and has no such term. Building on this, we construct a deep evidential pipeline in which every mass in the chain from heads to output is a valid Dempster--Shafer mass function, combined by Yager's rule. The output mass on the ignorance set, $m(\Theta)$, serves as the network's ``I do not know'' signal. The network is fit to hard labels and never sees an annotator distribution, yet $m(\Theta)$ orders human disagreement in two modalities, visual and linguistic. Dempster's rises too and ranks contested inputs about as well, but its normaliser divides the conflict out: Yager's vacuity is $9\times$ larger on the most contested CIFAR-10H images, so only Yager's is readable as an amount of ignorance without recalibration. For out-of-distribution detection it is competitive with strong single-pass baselines, while the clamped Dempster variant that scores higher becomes gradient-unstable once the clamp is removed.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.30173 — 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 2026-10-25T21:25:18.066Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).