Artificial synapses can boost neuromorphic computing to overcome the inherent limitations of von Neumann architecture. As a promising memristor candidate, ferroelectric tunnel junctions (FTJ) enable the authors to successfully emulate spike-timing-dependent synapses. However, the nonlinear and asymmetric synaptic weight update under repeated presynaptic stimulation hampers neuromorphic computing by favoring the runaway of synaptic weights during learning. Here, the authors demonstrate an FTJ whose conductivity varies linearly and symmetrically by judiciously combining ferroelectric domain switching and oxygen vacancy migration. The artificial neural network based on this FTJ-synapse achieves classification accuracy of 96.7% during supervised learning, which is the closest to the maximum theoretical value of 98% achieved to date. This artificial synapse also demonstrates stable unsupervised learning in a noisy environment for its well-balanced spike-timing-dependent plasticity response. The novel concept of controlling ionic migration in ferroelectric materials paves the way toward highly reliable and reproducible supervised and unsupervised learning strategies.

No abstract

Flexible synaptic devices that emulate biological synapses are the key to developing wearable intelligent equipment. However, there are still challenges to achieving multiple synaptic functions in flexible synaptic devices and realizing large‐scale production of synaptic devices. Herein, a flexible synaptic transistor on polyimide substrate is designed and fabricated using a solid‐state electrolyte gate and a VO2 Mott insulator thin film channel. Four orders conductance modulation is obtained in the VO2 thin film channel across the metal‐insulator transition. Several essential synaptic behaviors, including potentiation, depression, and short‐term/long‐term plasticity, are successfully demonstrated in this flexible synaptic device. More importantly, this synaptic transistor exhibits high cycling stability and good tolerance to bending deformation. A simulated artificial neural network built from this synaptic device achieves high recognition accuracy of 98% in the flat state and 92% in the bending state. In addition, this flexible device is fabricated on polyimide substrate by magnetron sputtering method, which has low cost and is compatible with modern semiconductor technology. This work can pave the way for large‐scale production of flexible synaptic Mott transistor devices.