While the algorithms at the heart of traditional networks are set during training, when these systems are fed reams of data to calibrate the best values for their weights, liquid neural nets are more ...