Semantic Parsing on WikiSQL (test)
82.6Execution AccuracyTypeSQL+TC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TypeSQL+TCUse table content=true2018.10 | 82.6 | — | — | |
| COARSE2FINEoracle sketch=yes2018.05 | 79.6 | 73 | — | |
| TRANXParent feeding=true, Answer pruning=true, Use table content=true2018.10 | 78.6 | 68.4 | — | |
| TRANXParent feeding=false, Answer pruning=true, Use table content=true2018.10 | 78.6 | 68.6 | — | |
| COARSE2FINE2018.05 | 78.5 | 71.7 | — | |
| COARSE2FINEsketch encoder=no2018.05 | 77.7 | 70.8 | — | |
| ONESTAGE2018.05 | 75.9 | 68.8 | — | |
| COARSE2FINEtable-aware input encoder=no2018.05 | 75.6 | 68.6 | — | |
| STAMP+RLUse table content=true2018.10 | 74.6 | 61 | — | |
| STAMPUse table content=true2018.10 | 74.4 | 60.7 | — | |
| TypeSQLUse table content=false2018.10 | 73.5 | — | — | |
| TRANXParent feeding=false, Answer pruning=false, Use table content=false2018.10 | 71.7 | 62.9 | — | |
| TRANXParent feeding=true, Answer pruning=false, Use table content=false2018.10 | 71.6 | 62.6 | — | |
| SQLNET2018.05 | 68 | 61.3 | — | |
| SQLNetUse table content=false2018.10 | 68 | — | — | |
| PT-MAMLUse table content=false2018.10 | 68 | 62.8 | — | |
| PointSQLUse table content=false2018.10 | 66.8 | 61.5 | — | |
| Seq2SQLreinforcement learning=true2017.08 | 59.4 | 48.3 | — | |
| SEQ2SQL2018.05 | 59.4 | 48.3 | — | |
| Seq2SQLUse table content=false2018.10 | 59.4 | 48.3 | — | |
| Seq2SQLreinforcement learning=false2017.08 | 57.1 | 47.4 | — | |
| Augmented Pointer Network2017.08 | 53.3 | 43.3 | — | |
| Aug Ptr Network2018.05 | 53.3 | 43.3 | — | |
| Baseline2017.08 | 35.9 | 23.4 | — | |
| SEQ2SEQ2018.05 | 35.9 | 23.4 | — | |
| Seq2SeqUse table content=false2018.10 | 35.9 | 23.4 | — | |
| SEQ2TREEUse table content=false2018.10 | 35.9 | 23.4 | — | |
| Abstract ProgramsSupervision Type=Denotations, Attention Type=standard attention2019.09 | — | — | 75.3 | |
| Agarwal et al.Supervision Type=Denotations2019.09 | — | — | 74.8 | |
| Dong and LapataSupervision Type=Programs2019.09 | — | — | 78.5 | |
| GPT-2Backbone=Smallest pre-trained GPT-2 model, Training epochs=9, Decoding=Greedy2019.09 | — | 70.7 | — | |
| Huang et al.Supervision Type=Programs2019.09 | — | — | 68 | |
| Liang et al.Supervision Type=Denotations2019.09 | — | — | 72.1 | |
| MQANEvaluation mode=Single task scores2019.09 | — | 72.6 | — | |
| Shi et al.Supervision Type=Programs2019.09 | — | — | 83.7 | |
| Structured-Attention ParserSupervision Type=Denotations, Attention Type=structured attention, Configuration=Abstract Programs2019.09 | — | — | 79.3 | |
| Sun et al.Supervision Type=Programs2019.09 | — | — | 74.6 | |
| Typed Seq2SeqSupervision Type=Denotations2019.09 | — | — | 74.7 | |
| Wang et al.Supervision Type=Programs2019.09 | — | — | 66.8 | |
| Xu et al.Supervision Type=Programs2019.09 | — | — | 68 | |
| Yu et al.Supervision Type=Programs2019.09 | — | — | 73.5 | |
| Zhong et al.Supervision Type=Programs2019.09 | — | — | 59.4 |