RoTaR: Efficient Row-Based Table Representation Learning via Teacher-Student Training (Short Paper)

Abstract

We propose RoTaR, a row-based table representation learning method, to address the efficiency and scalability issues faced by existing table representation learning methods. The key idea of RoTaR is to generate query-agnostic row representations that could be re-used via query-specific aggregation. In addition to the row-based architecture, we introduce several techniques: cell-aware position embedding, AutoEncoder objective in transformer models, teacher-student training paradigm, and selective backward to improve the performance of RoTaR model.

Cite

Text

Chen et al. "RoTaR: Efficient Row-Based Table Representation Learning via Teacher-Student Training (Short Paper)." NeurIPS 2022 Workshops: TRL, 2022.

Markdown

[Chen et al. "RoTaR: Efficient Row-Based Table Representation Learning via Teacher-Student Training (Short Paper)." NeurIPS 2022 Workshops: TRL, 2022.](https://mlanthology.org/neuripsw/2022/chen2022neuripsw-rotar/)

BibTeX

@inproceedings{chen2022neuripsw-rotar,
  title     = {{RoTaR: Efficient Row-Based Table Representation Learning via Teacher-Student Training (Short Paper)}},
  author    = {Chen, Zui and Cao, Lei and Madden, Samuel},
  booktitle = {NeurIPS 2022 Workshops: TRL},
  year      = {2022},
  url       = {https://mlanthology.org/neuripsw/2022/chen2022neuripsw-rotar/}
}