English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Undirected neural sequence models have achieved performance competitive withthe state-of-the-art directed sequence models that generate monotonically fromleft to right in machine translation tasks. In this work, we train a policythat learns the generation order for a pre-trained, undirected translationmodel via reinforcement learning. We show that the translations decoded by ourlearned orders achieve higher BLEU scores than the outputs decoded from left toright or decoded by the learned order from Mansimov et al. (2019) on the WMT'14German-English translation task. On examples with a maximum source and targetlength of 30 from De-En, WMT'16 English-Romanian, and WMT'21 English-Chinesetranslation tasks, our learned order outperforms all heuristic generationorders on four out of six tasks. We next carefully analyze the learned orderpatterns via qualitative and quantitative analysis. We show that our policygenerally follows an outer-to-inner order, predicting the left-most andright-most positions first, and then moving toward the middle while skippingless important words at the beginning. Furthermore, the policy usually predictspositions for a single syntactic constituent structure in consecutive steps. Webelieve our findings could provide more insights on the mechanism of undirectedgeneration models and encourage further research in this direction. Our code ispublicly available at https://github.com/jiangycTarheel/undirected-generation

Learning and Analyzing Generation Order for Undirected Sequence Models