search

lovecambi / qebrain

machine translation and quality estimation
BSD 2-Clause "Simplified" License
34 stars 18 forks source link

Does this version of code include "Boosting Model Performance" in the paper? #12

Closed meettyj closed 5 years ago

meettyj commented 5 years ago

In the paper, you mentioned Human-crafted Features, Artificial QE Data Construction and Greedy Ensemble Selection have been used to improve performance. However, I couldn't find any related part in your code. Am I wrong or you just ignore implementing it in this version of code? Thanks!

lovecambi commented 5 years ago

In the data folder, I have included some tfrecord data, which is an example of hand crafted feature.

Average ensemble also works well. You can simply try it.

meettyj commented 5 years ago

Thanks.