Conjecture: machine translation research is not advancing as smoothly as it could. There are hardly any incentives to prioritize genuine scientific progress.
In this talk I will highlight what I believe to be the drivers of actual progress. In short, making results reproducible is paramount and the methodology must allow for meaningful conclusions to be drawn from those results.
This will enable us to discuss what you, as a young and promising scientist, can do to bring about change.
Mathias is a final-year PhD student at the University of Zurich. His strongest suit by far is that he likes to admit he is wrong.