Sequence-to-Sequence Learning as Beam-Search Optimization

EMNLP 2016 Sam WisemanAlexander M. Rush

Sequence-to-Sequence (seq2seq) modeling has rapidly become an important general-purpose NLP tool that has proven effective for many text-generation and sequence-labeling tasks. Seq2seq builds on deep neural language modeling and inherits its remarkable accuracy in estimating local, next-word distributions... (read more)

PDF Abstract

Evaluation results from the paper

Task Dataset Model Metric name Metric value Global rank Compare
Machine Translation IWSLT2015 German-English Word-level CNN w/attn, input feeding BLEU score 24.0 # 16