Last Words: Boring Problems are Sometimes the Most Interesting

Abstract In a recent position paper, Turing Award Winners Yoshua Bengio, Geoffrey Hinton, and Yann LeCun make the case that symbolic methods are not needed in AI that, while there still many issues to be resolved, will solved using purely neural methods. this piece I issue challenge: Demonstrate approach problem of text normalization is possible. Various groups have tried, but so far nobody has eliminated unrecoverable errors, errors where, due insufficient training data or faulty generalization, system substitutes some other reading for correct one. Solutions been proposed involve marriage traditional finite-state with models, thus shown can alone. Though hardly an “exciting” problem, argue until one solve “boring” problems like methods, cannot claim success.