Investigating Active Learning in Interactive Neural Machine Translation
Kamal Kumar Gupta, Dhanvanth Boppana, Rejwanul Haque, Asif Ekbal, Pushpak Bhattacharyya · NORMA · 2021
Interactive-predictive translation is a collaborative iterative process, where human translators produce translations with the help of machine translation (MT) systems interactively. Various sampling techniques in active learning (AL) exist to update the neural MT (NMT) model in the interactive-predictive scenario. In this paper, we explore term based (named entity count (NEC)) and quality based (quality estimation (QE), sentence similarity (Sim)) sampling techniques – which are used to find the ideal candidates from the incoming data – for human supervision and MT model’s weight updation. We carried out experiments with three language pairs, viz. German-English, Spanish-English and Hindi-English. Our proposed sampling technique yields 1.82, 0.77 and 0.81 BLEU points improvements for German-English, Spanish-English and Hindi-English, respectively, over random sampling based baseline. It also improves the present state-of-the-art by 0.35 and 0.12 BLEU points for German-English and Spanish-English, respectively. Human editing effort in terms of number-of-words-changed also improves by 5 and 4 points for German-English and Spanish-English, respectively, compared to the state-of-the-art.