They attempted to method the problem with several symbolic strategies, along with what were then termed "neural networks"; these ended up largely perceptrons and various versions which were later discovered to become reinventions on the generalised linear styles of figures.[20] Probabilistic reasoning was also employed, particularly in automated me