Fast agreement-driven device-calibrated local learning paradigms for spiking neural networks.

We introduce Spike Agreement Dependent Plasticity (SADP) and its correlation-based variant Spike Correlation Dependent Plasticity (SCDP), biologically inspired synaptic learning rules for Spiking Neural Networks (SNNs) that rely on the agreement between pre- and post-synaptic spike trains rather than precise spike-pair timing. SADP generalizes classical Spike-Timing-Dependent Plasticity (STDP) by replacing pairwise temporal updates with population-level agreement metric Cohen's κ, while SCDP employs raw Pearson correlation to capture direct co-firing intensity. Both rules admit linear-time complexity and support efficient hardware implementation via event-driven, bitwise logic. Empirical evaluations on MNIST and Fashion-MNIST show that SADP and SCDP learn discriminative features significantly faster than classical STDP, achieving strong downstream classification accuracy within an order of magnitude less training time. While the present study validates these rules in software simulation on single-layer networks, our results establish agreement-driven plasticity as a fast and local learning framework that can, in principle, be adapted to neuromorphic substrates. This framework bridges the gap between biological plausibility and computational speed, offering a viable unsupervised learning mechanism for next-generation neuromorphic systems with possibilities of extension to a fast and local supervised learning paradigm.