Neuromorphic and in-materio computation research are heralding advanced, brain-like computational capabilities that have the potential to supplement current General-Purpose Interface (GPI)-based AI and machine learning technologies. Although this field remains largely exploratory, formidable obstacles must be surmounted for it to become a revolutionary technology. The development of the atomic switch and related neuromorphic junctions in synaptic networks has led to emergent behavior, including phenomena such as persistent fluctuations in conductivity power law, dragon king distributions, hysteresis, chaotic attractor dynamics, and avalanche criticality [1]. Several research groups have adopted these devices for real computational applications, such as speech and image recognition, and predictive analysis, primarily as reservoir or echo state machines [2,3]. The range of fabrication strategies and nanoarchitectonic material systems has broadened to include molecular and polymeric systems [4], complementing the original electro-ionic, magnetic, and optical approaches for creating complex adaptive systems. In my presentation, I will discuss the current state of these technologies and their potential applications in addressing critical challenges [5] based on our latest findings and consider the possible integration with biological brains.

References: