Winner-Takes-All Neural Network Based on 3D NAND Flash With 1-Bit Erase Scheme

3D NAND flash memory constitutes strong competitors for neuromorphic computing due to its high density and mature technology. Neural networks based on differential pair have been reported since conventional block erase scheme hinders individual conductance modulation. However, the halves efficient synaptic operation is still inevitable when one of devices reaches minimum conductance. To overcome these shortcomings, 1-bit was proposed gate induced drain leakage (GIDL) effect. But in this scheme, there are risks unintended erase. In work, gradient bias adopted eliminate potential scheme. Both TCAD simulation experimental results manifest better inhibition Based improved a winner-takes-all neural network constructed, where doubled free from dilemma previous work