Passage Re-ranking with BERT

  • 2019-01-30 02:25:25
  • Rodrigo Nogueira, Kyunghyun Cho
  • 0

Abstract

Recently, neural models pretrained on a language modeling task, such as ELMo(Peters et al., 2017), OpenAI GPT (Radford et al., 2018), and BERT (Devlin etal., 2018), have achieved impressive results on various natural languageprocessing tasks such as question-answering and natural language inference. Inthis paper, we describe a simple re-implementation of BERT for query-basedpassage re-ranking. Our system is the state of the art on the TREC-CAR datasetand the top entry in the leaderboard of the MS MARCO passage retrieval task,outperforming the previous state of the art by 27% (relative) in [email protected] Thecode to reproduce our results is available athttps://github.com/nyu-dl/dl4marco-bert

 

Introduction (beta)

None

 

Conclusion (beta)

None