This is the second year of the TREC Deep Learning Track, with the goal of studying ad hoc ranking in the large training data regime . We evaluate using single-shotTREC-style evaluation, to give us a picture of which ranking methods work best . This year we have further evidence thatrankers with BERT-style pretraining outperform other rankers in large dataregime. This year, we have more evidence that rankers with

Author(s) : Nick Craswell, Bhaskar Mitra, Emine Yilmaz, Daniel Campos

Links : PDF - Abstract

Code :

Keywords : year - style - evidence - trec - deep -

Leave a Reply

Your email address will not be published. Required fields are marked *