Paper: Automatic Voice Selection in Japanese based on Various Linguistic Information

ACL ID W13-2118
Title Automatic Voice Selection in Japanese based on Various Linguistic Information
Venue European Workshop on Natural Language Generation
Session
Year 2013
Authors

This paper focuses on a subtask of natu- ral language generation (NLG), voice se- lection, which decides whether a clause is realised in the active or passive voice ac- cording to its contextual information. Au- tomatic voice selection is essential for re- alising more sophisticated MT and sum- marisation systems, because it impacts the readability of generated texts. How- ever, to the best of our knowledge, the NLG community has been less concerned with explicit voice selection. In this pa- per, we propose an automatic voice se- lection model based on various linguistic information, ranging from lexical to dis- course information. Our empirical evalua- tion using a manually annotated corpus in Japanese demonstrates that the proposed model achieved 0.758 in F-score, outper- forming the two...