The symposium will examine memristive devices through three complementary lenses: physical mechanisms, mathematical and compact modeling, and circuit-level exploitation for AI-oriented computing. The goal is to provide a focused view of memristors not simply as emerging memory elements, but as a cross-disciplinary platform connecting nanoscale device physics to new computational paradigms.
Particular attention will be devoted to the relationship between material properties, switching dynamics, analog behavior, and device non-idealities, and to how these aspects can be captured through physics-based and compact models suitable for circuit and system design. In this perspective, variability, stochasticity, and reliability are not only critical issues to be addressed, but also features that may influence the way memristive devices are effectively employed in computation.
By bringing together experts in device characterization, modeling, and circuit design, the symposium will highlight both the opportunities and the challenges of using memristive phenomena in analog, in-memory, and neuromorphic systems. More broadly, the session aims to promote a cross-layer view of memristors, showing how progress in the understanding of device physics can translate into more accurate models and, in turn, into more effective hardware solutions for emerging AI circuits.