密集振荡联想记忆中的相变与调制
Phase Transitions and Modulation in Dense Oscillatory Associative Memories
- Harbin Institute of Technology（哈尔滨工业大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
本文提出高阶Winfree模型用于振荡联想记忆，通过局部稳定性二分法分析串扰噪声驱动的相变级联，推导相边界并设计辅助场调制方案，电路与模拟验证了理论。
AI中文摘要：
我们引入了用于振荡联想记忆的高阶Winfree模型。我们通过存储与非存储模式之间的局部稳定性二分法来表征记忆性能。利用这一准则，我们表明随着记忆负载的增加，串扰噪声驱动了铁磁、混合和自旋玻璃相变的级联。对于多项式型和指数型Winfree模型，我们推导了相边界，并开发了适应不同相位的辅助场调制方案。我们进一步展示了相应的电路设计和数值模拟，以支持我们的理论预测。
英文摘要：
We introduce high-order Winfree models for oscillatory associative memory. We characterize memory performance through the local-stability dichotomy between stored and unstored patterns. Using this criterion, we show that crosstalk noise drives a cascade of ferromagnetic, mixed, and spin-glass transitions as the memory load increases. For both polynomial and exponential Winfree models, we derive the phase boundaries and develop auxiliary-field modulation schemes adapted to the different phases. We further present corresponding circuit designs and numerical simulations that support our theoretical predictions.
