Injecting The BM25 Score As Text Improves Bert-based Re-rankers

Abstract

In this paper we propose a novel approach for combining first-stage lexical retrieval models and Transformer-based re-rankers: we inject the relevance score of the lexical model as a token in the middle of the input of the cross-encoder re-ranker. It was shown in prior work that interpolation between the relevance score of lexical and BERT-based re-rankers may not consistently result in higher effectiveness. Our idea is motivated by the finding that BERT models can capture numeric information. We compare several representations of the BM25 score and inject them as text in the input of four different cross-encoders. We additionally analyze the effect for different query types, and investigate the effectiveness of our method for capturing exact matching relevance. Evaluation on the MSMARCO Passage collection and the TREC DL collections shows that the proposed method significantly improves over all cross-encoder re-rankers as well as the common interpolation methods. We show that the impr

Stats

• citations24
• S2 citations—
• github stars0
• HF likes0
• heat score10.48
• arxiv keyaskari2023injecting

Related papers

• On The Interpolation Of Contextualized Term-based Ranking With BM25 For Query-by-example Retrieval (2022)7.50
• How Different Are Pre-trained Transformers For Text Ranking? (2022)7.81
• Refit: Relevance Feedback From A Reranker During Inference (2023)0.00
• HYRR: Hybrid Infused Reranking For Passage Retrieval (2022)0.00
• Enhancing Documents With Multidimensional Relevance Statements In Cross-encoder Re-ranking (2023)0.00
• Parameter-efficient Neural Reranking For Cross-lingual And Multilingual Retrieval (2022)0.00
• Shallow Cross-encoders For Low-latency Retrieval (2024)2.26
• Boosting Zero-shot Cross-lingual Retrieval By Training On Artificially Code-switched Data (2023)4.52