Published November 27, 2017 | Version v1
Conference paper Open

Learning How to Simplify From Explicit Labeling of Complex-Simplified Text Pairs

  • 1. University of Sheffield
  • 2. University of Copenhagen
  • 3. University
  • 4. Univer


Current research in text simplification has been hampered by two central problems: (i) the small amount of high-quality parallel simplification data available, and (ii) the lack of explicit annotations of simplification operations, such as deletions or substitutions, on existing data. While the recently introduced Newsela corpus has alleviated the first problem, simplifications still need to be learned directly from parallel text using black-box, end-to-end approaches rather than from explicit annotations. These complex-simple parallel sentence pairs often differ to such a high degree that generalization becomes difficult.  End-to-end models also make it hard to interpret what is actually learned from data.  We propose a method that decomposes the task of TS into its sub-problems. We devise a way to automatically identify operations in a parallel corpus and introduce a sequence-labeling approach based on these annotations. Finally, we provide insights on the types of transformations that different approaches can model.



Files (179.3 kB)

Name Size Download all
179.3 kB Preview Download

Additional details


SIMPATICO – SIMplifying the interaction with Public Administration Through Information technology for Citizens and cOmpanies 692819
European Commission