Fig. 5: Performance of AFeFET-based spiking neurons. | Nature Communications

Fig. 5: Performance of AFeFET-based spiking neurons.

From: Reconfigurable neuromorphic functions in antiferroelectric transistors through coupled polarization switching and charge trapping dynamics

Fig. 5

a Illustration of the LIF spiking neuron model and the LIF dynamics. b Repeated firing events of the AFeFET neuron showing self-reset behavior with no firing triggered by low input strength. c Reliable firing performance up to 108 cycles, showing minimal cycle-to-cycle variations. d–f Device responses to pulses with varying amplitudes, pulses widths, and intervals. Corresponding statistical data showing input spike numbers required for firing, with results from ten devices in each box. g Schematic of the simulated hardware implementation. h Test accuracy of the output results, with the inset showing the configuration of the SNN, including input, hidden, and output layers. i Corresponding confusion matrix of the test results.

Back to article page