They attempted to solution the trouble with various symbolic procedures, together with what had been then termed "neural networks"; these have been generally perceptrons along with other products that were later found to become reinventions on the generalised linear models of figures.[twenty] Probabilistic reasoning was also employed, particularly