Neural MachineTranslation1990s-2010s:StatisticalMachineTranslationSMTwasahuge researchfieldThebestsystemswereextremelycomplex·Hundreds of important details wehaven'tmentioned hereSystemshadmanyseparately-designedsubcomponents·Lots of feature engineeringNeedto designfeaturesto capture particularlanguagephenomena.Requirecompilingandmaintaining extraresources.Liketablesofequivalentphrases麦通大学.LotsofhumanefforttomaintainRepeatedeffortforeachlanguagepair
l 1990s-2010s: Statistical Machine Translation • SMT was a huge research field • The best systems were extremely complex • Hundreds of important details we haven’t mentioned here • Systems had many separately-designed subcomponents • Lots of feature engineering • Need to design features to capture particular language phenomena • Require compiling and maintaining extra resources • Like tables of equivalent phrases • Lots of human effort to maintain • Repeated effort for each language pair! Neural Machine Translation
Outline1.Pre-Neural Machine Translation2.Neural Machine Translation发通大
1. Pre-Neural Machine Translation 2. Neural Machine Translation Outline
Neural Machine Translation2014(dramaticreenactment)
2014 (dramatic reenactment) Neural Machine Translation
Neural Machine Translation2014NeuralMachineTranslationMTresearch(dramaticreenactment)
2014 MT research Ma Neural chine Translation (dramatic reenactment) 19 Neural Machine Translation
Neural MachineTranslationWhatisNeuralMachineTranslation?. Neural Machine Translation (NMT) is a way to doMachine Translation with a single end-to-end neuralnetworkThe neural network architecture is called a sequence.to-sequence model (seq2seq) and itinvolves two RNNs通大
l What is Neural Machine Translation? • Neural Machine Translation (NMT) is a way to do Machine Translation with a single end-to-end neural network • The neural network architecture is called a sequence- to-sequence model (seq2seq) and it involves two RNNs Neural Machine Translation