High-Speed and Low-Power Superconducting Neuromorphic Circuits Based on Quantum Phase-Slip Junctions

Data_Sheet_1.PDF

“…In additional to biasing and impedance matching challenges, which were discussed in a previous section, device tolerance is also a concern for practical applications. For example, we have performed a small number of simulations to determine the tolerance for each QPSJ in the previously-presented multiweight synaptic circuit (Cheng et al, 2021). The results indicated that the tolerance for identical parallel QPSJs is generally low (< 1%) while the tolerance for other QPSJs in the circuit is usually between a few to tens of percent.…”

“…Ideally, the critical voltage difference between Q 1 and Q 2 or Q 2 and Q 3 should be the same as the voltage change on node 1 after receiving an input voltage pulse from V w , which is ∼ 2e/C 1 . In simulation, the tolerance of these parallel QPSJs was found to be less than ∼ 1% (Cheng et al, 2021). The weight can be increased by applying negative pulses at V w or decreased by applying positive pulses at V w .…”

“…A fanout circuit allows one neuron to be able to connect to multiple other neurons. We previously designed and presented a fan-out circuit to split current pulses from an IFN circuit to provide a means to connect to other IFN circuits, which is briefly reviewed here (Cheng et al, 2021). Charge-flux converters (Goteti and Hamilton, 2019) were used to convert quantized current pulses to single flux quantum (SFQ) pulses that can in turn switch multiple QPSJs, as shown in Figure 3A.…”

Data_Sheet_2.zip

“…A synapse is connected between two neurons to transmit weighted spiking signals. We previously designed and presented a QPSJ-based multi-weight synaptic circuit to transmit weighted current pulses between neuron circuits, which is briefly reviewed here (Cheng et al, 2021). The circuit shown in Figure 2A is a multi-weight synaptic circuit that can generate different numbers of sequential current pulses, which correspond to a weight of 0, 1, 2 or 3.…”

“…The simulation results have demonstrated the fan-out function of this circuit. This circuit does not appear to have a limit for the maximum fanout in simulation, but can be limited by the practical circuits due to fabrication challenges (Cheng et al, 2021).…”