Neural Learning Methods TRIZ Cases
- Neural Network Training via Self-Supervised Loss for Algorithmic Tasks
A neural network uses self-supervised loss to train on algorithmic tasks by generating invariant intermediate representations from augmented datasets.