Long Document Re-ranking with Modular Re-ranker
This addresses a specific problem in information retrieval for researchers and practitioners by improving re-ranking accuracy on long documents, though it is incremental as it builds on existing Transformer frameworks.
The paper tackles the challenge of long document re-ranking by proposing a modular Transformer re-ranker that models full query-to-document interaction to overcome information bottlenecks in encode-and-pool methods, achieving effective re-ranking on datasets like Robust04, ClueWeb09, and MS-MARCO document ranking.
Long document re-ranking has been a challenging problem for neural re-rankers based on deep language models like BERT. Early work breaks the documents into short passage-like chunks. These chunks are independently mapped to scalar scores or latent vectors, which are then pooled into a final relevance score. These encode-and-pool methods however inevitably introduce an information bottleneck: the low dimension representations. In this paper, we propose instead to model full query-to-document interaction, leveraging the attention operation and modular Transformer re-ranker framework. First, document chunks are encoded independently with an encoder module. An interaction module then encodes the query and performs joint attention from the query to all document chunk representations. We demonstrate that the model can use this new degree of freedom to aggregate important information from the entire document. Our experiments show that this design produces effective re-ranking on two classical IR collections Robust04 and ClueWeb09, and a large-scale supervised collection MS-MARCO document ranking.