Simple Entity-Centric Questions Challenge Dense Retrievers
This work highlights a critical limitation in dense retrievers for open-domain QA, affecting robustness across input distributions, though it is incremental in identifying and analyzing the issue.
The paper tackles the problem that dense retrieval models underperform on simple entity-centric questions, showing they only generalize to common entities unless the question pattern is seen in training, with sparse methods outperforming them drastically on the EntityQuestions dataset.
Open-domain question answering has exploded in popularity recently due to the success of dense retrieval models, which have surpassed sparse models using only a few supervised training examples. However, in this paper, we demonstrate current dense models are not yet the holy grail of retrieval. We first construct EntityQuestions, a set of simple, entity-rich questions based on facts from Wikidata (e.g., "Where was Arve Furset born?"), and observe that dense retrievers drastically underperform sparse methods. We investigate this issue and uncover that dense retrievers can only generalize to common entities unless the question pattern is explicitly observed during training. We discuss two simple solutions towards addressing this critical problem. First, we demonstrate that data augmentation is unable to fix the generalization problem. Second, we argue a more robust passage encoder helps facilitate better question adaptation using specialized question encoders. We hope our work can shed light on the challenges in creating a robust, universal dense retriever that works well across different input distributions.