Running Projects

We collected and shared proposed projects before the Marathon. As usual, project topics settled on the first day of the MT Marathon. Most of the projects actually made it to the final presentation and some will continue even (long) after the Marathon.

Projects proposed (see the project proposals):

  1. Lattice training prototype
  2. PortS mtplz to Joshua - (Slides Δ)
  3. Starting an Open Source Toolkit for Working with Comparable Corpora
  4. Twitter Crowd Translation (was MT Evaluation and other bits for Twitter Crowd Translation) - (Slides Δ)
  5. Twitter Crowd Translation (was MultiMT) - (Slides Δ)
  6. MT with NLTK: Yet Another MT Decoder (Slides Δ)
  7. OxLM: Extend neural language models by conditioning on source sentences (Slides Δ)
  8. User adaptive SMT system (Moses for CAT)
  9. KenLM Fun with language model - (Slides Δ)
  10. Features for Syntax-based Moses (Slides Δ)
  11. Docent FTW! (Project material)
  12. GPU Phrase-Based Decoder
  13. Inline Tag Handling Continued
  14. Decoding with sampling and nonlocal features - (Slides Δ)
  15. Web-service for format converters for CAT
  16. Domain adaptation via biased sampling (Project material) (Slides Δ)