| （英） |
Hopfield Networks, a well-known model of associative memory, defines energy function and time evolution.Recently, a new model, Modern Hopfield Networks, has been proposed to redefine the energy function and time evolution of Hopfield Networks.This model improves the memory capacity of Hopfield Networks and clarifies the relationship between Hopfield Networks and deep learning models.Furthermore, Quantum Hopfield Networks has been proposed that uses the quantum master equation describing the dynamics of open quantum system for the time evolution of Hopfield Networks.In this study, we propose an extension of Quantum Hopfield Networks to a discrete Modern Hopfield network and analyze the phase diagram and stability of the solution. |