My proposal aims to accelerate future semantic annotation projection research by creating two main deliverables. First, before the GSoC 2020 starts, I intend to complete a Python API for retrieving the annotation dataset. Throughout the GSoC 2020, I will modify existing parsers and create neural networks to project semantic frames and frame elements to unannotated sentences. I mainly work with projections on the manually aligned sentences and aligned sentences generated through transitivity. In addition to suggesting techniques for tackling m:n alignment, I also propose metrics for evaluating the projections in binary and graded manners as well as a phraseological method for analyzing projections for metaphorical phrases. If time permits, I would create sentence alignment models for aligning EN-JP, EN-FR, EN-UR, and EN-HI sentences.



Zheng Xin Yong


  • Oliver Czulo
  • Collin Baker
  • Maryam Foradi
  • Tiago Timponi Torrent
  • DebanjanaKar
  • Ely Matos