arXiv Open Access 2025

Accuracy and capacity of Modern Hopfield networks with synaptic noise

Sharba Bhattacharjee Ivar Martin

Lihat Sumber

Abstrak

We study the retrieval accuracy and capacity of modern Hopfield networks of with two-state (Ising) spins interacting via modified Hebbian $n$-spin interactions. In particular, we consider systems where the interactions deviate from the Hebb rule through additive or multiplicative noise or through clipping or deleting interactions. We find that the capacity scales as $N^{n-1}$ with the number of spins $N$ in all cases, but with a prefactor reduced compared to the Hebbian case. For $n=2$ our results agree with the previously known results for the conventional $n = 2$ Hopfield network.

Topik & Kata Kunci

cond-mat.dis-nn

Penulis (2)

Sharba Bhattacharjee

Ivar Martin

Format Sitasi

APA MLA BibTeX

Bhattacharjee, S., Martin, I. (2025). Accuracy and capacity of Modern Hopfield networks with synaptic noise. https://arxiv.org/abs/2503.00241

Akses Cepat

Lihat di Sumber

Informasi Jurnal

Tahun Terbit: 2025
Bahasa: en
Sumber Database: arXiv
Akses: Open Access ✓