29 citations · 75 across the 17 of their papers we have counts for
Showing 2017Show all
3 papers · 1 filter
cs.CL2017★ 29 cited
Compressing Word Embeddings via Deep Compositional Code Learning
Raphael Shu, Hideki Nakayama
Natural language processing (NLP) models often require a massive number of parameters for word embeddings, resulting in a large storage or memory footprint. Deploying neural NLP mo…
cs.CL2017
Single-Queue Decoding for Neural Machine Translation
Raphael Shu, Hideki Nakayama
Neural machine translation models rely on the beam search algorithm for decoding. In practice, we found that the quality of hypotheses in the search space is negatively affected ow…
cs.CL2017★ 8 cited
Later-stage Minimum Bayes-Risk Decoding for Neural Machine Translation
Raphael Shu, Hideki Nakayama
For extended periods of time, sequence generation models rely on beam search algorithm to generate output sequence. However, the correctness of beam search degrades when the a mode…