The Power of Selecting Key Blocks with Local Pre-ranking for Long Document Information Retrieval
This addresses the computational inefficiency of transformers for long-document information retrieval, but it is incremental as it builds on existing block selection and aggregation methods.
The paper tackles the challenge of processing long documents with transformer models by selecting key blocks via local pre-ranking and aggregating them into a shorter document for BERT-based retrieval. Experiments on standard IR datasets show the approach is effective, though no specific performance numbers are provided.
On a wide range of natural language processing and information retrieval tasks, transformer-based models, particularly pre-trained language models like BERT, have demonstrated tremendous effectiveness. Due to the quadratic complexity of the self-attention mechanism, however, such models have difficulties processing long documents. Recent works dealing with this issue include truncating long documents, in which case one loses potential relevant information, segmenting them into several passages, which may lead to miss some information and high computational complexity when the number of passages is large, or modifying the self-attention mechanism to make it sparser as in sparse-attention models, at the risk again of missing some information. We follow here a slightly different approach in which one first selects key blocks of a long document by local query-block pre-ranking, and then few blocks are aggregated to form a short document that can be processed by a model such as BERT. Experiments conducted on standard Information Retrieval datasets demonstrate the effectiveness of the proposed approach.