1 paper · 1 filter
Junhao Su, Chenghao He, Feiyu Zhu +3
Traditional deep learning relies on end-to-end backpropagation for training, but it suffers from drawbacks such as high memory consumption and not aligning with biological neural n…