When doing beam search in sequence to sequence models, one explores next words in order of their likelihood. However, during decoding, there may be other constraints we have or objectives we wish to maximize. For example, sequence length, BLEU score, or mutual information between the target and source sentences. In order to accommodate these additional desiderata, the authors add an additional term Q onto the likelihood capturing the appropriate criterion and then choose words based on this combined objective.
We’re pleased to announce a new feature in our interface that improves the translation review process. Previously, reviewers could make general text comments about errors and introduce categories by using hashtags to indicate the type of error (such as #punctuation or #mistranslation). However, it was still a manual process to collect the different hashtags and identify precisely where the error occurred.