Abstract | ||
---|---|---|
We present GraPPa, an effective pre-training approach for table semantic parsing that learns a compositional inductive bias in the joint representations of textual and tabular data. We construct synthetic question-SQL pairs over high-quality tables via a synchronous context-free grammar (SCFG). We pre-train our model on the synthetic data to inject important structural properties commonly found in semantic parsing into the pre-training language model. To maintain the model\u0027s ability to represent real-world data, we also include masked language modeling (MLM) on several existing table-related datasets to regularize our pre-training process. Our proposed pre-training strategy is much data-efficient. When incorporated with strong base semantic parsers, GraPPa achieves new state-of-the-art results on four popular fully supervised and weakly supervised table semantic parsing tasks. |
Year | Venue | DocType |
---|---|---|
2021 | ICLR | Conference |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
9 |
Name | Order | Citations | PageRank |
---|---|---|---|
Tao Yu | 1 | 25 | 6.78 |
Chien-Sheng Wu | 2 | 35 | 10.91 |
Victoria Lin | 3 | 45 | 3.39 |
Bailin Wang | 4 | 0 | 0.34 |
Yi Chern Tan | 5 | 4 | 2.08 |
Xinyi Yang | 6 | 0 | 0.34 |
Dragomir Radev | 7 | 5167 | 374.13 |
Richard Socher | 8 | 6770 | 230.61 |
Caiming Xiong | 9 | 969 | 69.56 |