Hybrid Retrieval and Multi-stage Text Ranking Solution at TREC 2022 Deep Learning Track
This work addresses text retrieval challenges for practical business applications, but it is incremental as it combines existing techniques.
The paper tackled large-scale text retrieval by proposing a hybrid retrieval and multi-stage ranking method, achieving 1st and 4th place on the TREC 2022 Deep Learning Track for passage and document ranking, respectively.
Large-scale text retrieval technology has been widely used in various practical business scenarios. This paper presents our systems for the TREC 2022 Deep Learning Track. We explain the hybrid text retrieval and multi-stage text ranking method adopted in our solution. The retrieval stage combined the two structures of traditional sparse retrieval and neural dense retrieval. In the ranking stage, in addition to the full interaction-based ranking model built on large pre-trained language model, we also proposes a lightweight sub-ranking module to further enhance the final text ranking performance. Evaluation results demonstrate the effectiveness of our proposed approach. Our models achieve the 1st and 4th rank on the test set of passage ranking and document ranking respectively.