Traditional IR rivals neural models on the MS MARCO Document Ranking Leaderboard

12/15/2020
by   Leonid Boytsov, et al.
0

This short document describes a traditional IR system that achieved MRR@100 equal to 0.298 on the MS MARCO Document Ranking leaderboard (on 2020-12-06). Although inferior to most BERT-based models, it outperformed several neural runs (as well as all non-neural ones), including two submissions that used a large pretrained Transformer model for re-ranking. We provide software and data to reproduce our results.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
09/20/2020

Longformer for MS MARCO Document Re-ranking Task

Two step document ranking, where the initial retrieval is done by a clas...
research
02/25/2021

Significant Improvements over the State of the Art? A Case Study of the MS MARCO Document Ranking Leaderboard

Leaderboards are a ubiquitous part of modern research in applied machine...
research
08/31/2021

Shallow pooling for sparse labels

Recent years have seen enormous gains in core IR tasks, including docume...
research
01/21/2021

Assessing the Benefits of Model Ensembles in Neural Re-Ranking for Passage Retrieval

Our work aimed at experimentally assessing the benefits of model ensembl...
research
02/24/2022

Finding Inverse Document Frequency Information in BERT

For many decades, BM25 and its variants have been the dominant document ...
research
03/14/2020

Document Ranking with a Pretrained Sequence-to-Sequence Model

This work proposes a novel adaptation of a pretrained sequence-to-sequen...
research
05/12/2020

Interpreting Neural Ranking Models using Grad-CAM

Recently, applying deep neural networks in IR has become an important an...

Please sign up or login with your details

Forgot password? Click here to reset