Simple Entity-Centric Questions Challenge Dense Retrievers

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.

PDF Abstract EMNLP 2021 PDF EMNLP 2021 Abstract

Datasets


Introduced in the Paper:

EntityQuestions

Used in the Paper:

Natural Questions PAQ
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Passage Retrieval EntityQuestions BM25 Recall@20 0.720 # 3
Passage Retrieval EntityQuestions DPR-NQ Recall@20 0.497 # 7
Passage Retrieval EntityQuestions DPR-multi Recall@20 0.567 # 6

Methods


No methods listed for this paper. Add relevant methods here