Naver Labs Europe (SPLADE) @ TREC Deep Learning 2022

This paper describes our participation to the 2022 TREC Deep Learning challenge. We submitted runs to all four tasks, with a focus on the full retrieval passage task. The strategy is almost the same as 2021, with first stage retrieval being based around SPLADE, with some added ensembling with ColBERTv2 and DocT5. We also use the same strategy of last year for the second stage, with an ensemble of re-rankers trained using hard negatives selected by SPLADE. Initial result analysis show that the strategy is still strong, but is still unclear to us what next steps should we take.

Paper

References (13)

10while our efficient models get very good results on MSMARCO, they struggle on TREC (something we already had seen)
11TREC19, 20, and 21 are biased to techniques that participated in the competition
12we are able to “beat” the best nDCG@10 results for TREC2019 and 21, but not for 2020, while we are able to increase mAP in all yearsgot

Scroll for more · 1 remaining

Similar papers

© 2026 NYSGPT2525 LLC