Metric: EM (higher is better)
| # | Model↕ | EM▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | Cluster-Former (#C=512) | 68 | No | Cluster-Former: Clustering-based Sparse Transfor... | 2020-09-13 | - |
| 2 | Locality-Sensitive Hashing | 66 | No | Reformer: The Efficient Transformer | 2020-01-13 | Code |
| 3 | Multi-passage BERT | 65.1 | No | Multi-passage BERT: A Globally Normalized BERT M... | 2019-08-22 | - |
| 4 | Sparse Attention | 64.7 | No | Generating Long Sequences with Sparse Transformers | 2019-04-23 | Code |
| 5 | DECAPROP | 62.2 | No | Densely Connected Attention Propagation for Read... | 2018-11-10 | Code |
| 6 | Denoising QA | 58.8 | No | - | - | Code |
| 7 | DecaProp | 56.8 | No | Densely Connected Attention Propagation for Read... | 2018-11-10 | Code |
| 8 | R^3 | 49 | No | R$^3$: Reinforced Reader-Ranker for Open-Domain ... | 2017-08-31 | Code |
| 9 | DrQA | 41.9 | No | Reading Wikipedia to Answer Open-Domain Questions | 2017-03-31 | Code |