Fast Neural Noise Stabilizes Associative Memory against Slow Synaptic Noise
Abstract
Neural systems operate under multi-timescale stochastic dynamics, yet reliably maintain stable functions. We investigate how fast neural fluctuations and slow synaptic variability asymmetrically influence the structural robustness of associative memories. While synaptic noise degrades memory formation, neural noise stabilizes the network against structural perturbations. For sequential learning of multiple memories, the average recall performance is maximized at an intermediate neural-noise amplitude, whereas increasing synaptic noise rapidly degrades memory performance. Using an adiabatic mean-field framework, we derive an analytical stationary distribution of stability factors. Our theory reveals that fast neural fluctuations push the system away from marginal stability, serving as an indispensable prerequisite for systemic stability.
Related
- Associative Memory for Quantum Entangled States
- Free energy landscape of Dense Associative Memory
- Alignment of the amplitudes of stored states for solving the associative memory problem using quantum annealing
- Sequential Retrieval in Dense Associative Memory: Asymptotic Dynamics and Storage Capacity
- Dynamical mean-field theory for associative memory with block-correlated patterns
BibTeX
@misc{kurikawa2026fast,
title = {{Fast Neural Noise Stabilizes Associative Memory against Slow Synaptic Noise}},
author = {Tomoki Kurikawa and Kunihiko Kaneko},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.00516},
eprint = {2611.00516},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.00516},
note = {CC BY-NC-SA 4.0},
}
RIS
TY - UNPB
TI - Fast Neural Noise Stabilizes Associative Memory against Slow Synaptic Noise
AU - Tomoki Kurikawa
AU - Kunihiko Kaneko
PY - 2026
DA - 2026/11/01
PB - arXiv
DO - 10.48550/arxiv.2611.00516
UR - https://arxiv.org/abs/2611.00516
AB - Neural systems operate under multi-timescale stochastic dynamics, yet reliably maintain stable functions. We investigate how fast neural fluctuations and slow synaptic variability asymmetrically influence the structural robustness of associative memories. While synaptic noise degrades memory formation, neural noise stabilizes the network against structural perturbations. For sequential learning of multiple memories, the average recall performance is maximized at an intermediate neural-noise amplitude, whereas increasing synaptic noise rapidly degrades memory performance. Using an adiabatic mean-field framework, we derive an analytical stationary distribution of stability factors. Our theory reveals that fast neural fluctuations push the system away from marginal stability, serving as an indispensable prerequisite for systemic stability.
ER -