“…DNNs need analog conductance modulation, while SNNs require conduction controlled by temporal correlations between signals; thus, an artificial synapse with diverse plasticity is necessary for diverse neural networks. , Ferroelectric thin films have attracted much attention for artificial synapses due to nonvolatility, easy controllability, high stability, and fast speed since their resistive switching comes from a pure electronic mechanism . Recently, ferroelectric synaptic devices have made important progress in functional simulation of biological synapses, such as paired-pulse facilitation (PPF), spike timing-dependent plasticity (STDP), spike rate-dependent plasticity (SRDP), and Bienenstock–Cooper–Munro (BCM) learning rules, and some with sub-nanosecond speed, linear weight update, and high precision. , However, versatile synaptic plasticity is still limited in a single ferroelectric device.…”