Metraipolis › archive › dc_6564dbca6778

Analog-Friendly Predictive Coding without Activation Derivatives — arXiv:2609.32350

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.32350
License (as archived)
CC0-1.0
Archived
2026-10-07T21:00:06.016Z
SHA-256 of the archived text
43964544d24697b47446c9b890e7f15f63acfc6568f256af266864bea72b21e7
Stable URL of this page
https://metraipolis.com/documento/dc_6564dbca6778

How to cite

Analog-Friendly Predictive Coding without Activation Derivatives — arXiv:2609.32350 — arXiv (en), https://arxiv.org/abs/2609.32350 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6564dbca6778 (Metraipolis archive dc_6564dbca6778, sha256 43964544d24697b47446c9b890e7f15f63acfc6568f256af266864bea72b21e7)
BibTeX
@misc{dc6564dbca6778,
  title = {Analog-Friendly Predictive Coding without Activation Derivatives — arXiv:2609.32350},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.32350},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6564dbca6778; sha256 43964544d24697b47446c9b890e7f15f63acfc6568f256af266864bea72b21e7},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.32350. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Predictive coding (PC) is a local, energy-based alternative to backpropagation (BP) whose iterative inference dynamics make it attractive for implementation on analog hardware. However, standard nonlinear PC requires evaluating the derivative of the activation function during both inference and learning, which can be difficult to realise physically. Here, we introduce \textit{activation-matched Bregman PC}, replacing standard squared-error energies with Bregman divergences matched to the activation function. This formulation eliminates activation derivatives and, when combined with inference via mirror descent, yields local inference and learning rules requiring only weighted sums, local prediction errors, state integration, and the activation function. In digital experiments, Bregman PC performs comparably to standard PC and BP on classification and generative tasks, while preserving characteristic learning dynamics of PC and its convergence to BP under stable large-model parameterisations. These results provide a more analog-friendly formulation of nonlinear PC while retaining its key computational properties.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE).
Pagina del lavoro: https://arxiv.org/abs/2609.32350 — 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 2026-10-07T20:29:02.300Z 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).