Exploiting Static Conductance and Dynamic Switching of Memristors for Artificial Intelligence Applications

Memristors are resistive switching devices whose conductance states can be programmed by applied voltages. They exhibit two fundamental characteristics—static conductance and dynamic switching—both of which can serve as computing primitives to accelerate artificial intelligence (AI) algorithms. Computations based on these physical properties offer several key advantages, including massive parallelism, low latency, and reduced data movement. This Review integrates perspectives from device physics to system-level implementation, beginning with the fundamentals of memristive behavior and progressing to two primary application domains. First, we examine how static conductance characteristics are exploited in in-memory analog computing, particularly for matrix–vector multiplication and matrix equation solving, and discuss how these primitives enable efficient neural network inference and training. Second, we survey dynamic switching behaviors and their algorithmic applications, including stateful logic for digital in-memory acceleration, attractor networks for associative memory, reservoir computing and spatiotemporal signal processing using transient device dynamics, biologically inspired spike-timing-dependent plasticity, and stochastic computation. In addition, we discuss key challenges such as device variability, stochastic switching, interconnect parasitics, peripheral DAC/ADC overhead, and endurance limitations. We also highlight opportunities for future development, emphasizing algorithm–hardware co-design to leverage application-specific error tolerance and mitigate device non-idealities. Finally, we outline promising research directions aimed at realizing robust, scalable, and energy-efficient memristor-based AI systems.