How much data is needed for reliable MT evaluation? Using bootstrapping to study human and automatic metrics Paula Estrella author Olivier Hamon author Andrei Popescu-Belis author 2007-sep 10-14 text Proceedings of Machine Translation Summit XI: Papers Bente Maegaard editor Copenhagen, Denmark conference publication estrella-etal-2007-much https://aclanthology.org/2007.mtsummit-papers.23/ 2007-sep 10-14