Fei Zeng, Da Ma
This paper reviews recent advances in signal processing and resistive switching mechanisms of niobium oxides. The reported signal-processing functionalities include neural spike firing, phase locking, frequency and amplitude modulation, information-flux control, and quantized encoding. The underlying resistive switching mechanism indicates that the phase composition within conductive filaments, or the local activity of niobium oxides, can be modulated and stabilized by acquiring appropriate energy from external electrical stimuli. Precise control over these two characteristics enables the design of artificial neural networks with high biological plausibility, thereby significantly enhancing computational efficiency.