SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2201.09745 · IJCAI · 2022

Table Pre-training: A Survey on Model Architectures, Pre-training Objectives, and Downstream Tasks

Fan Zhou, Dongmei Zhang, Shi Han, Ao Liu, Zhoujun Cheng, Haoyu Dong, Mengyu Zhou, Xinyi He, Anda Zhou

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Following the success of pre-training paradigm in the natural language domain, a flurry of table pretraining frameworks have been proposed and have achieved new state-of-the-arts on various downstream tasks such as table question answering, table type recognition, column relation classification, table search, and formula prediction. Various model architectures have been explored to best leverage the characteristics of structured tables, especially specially-designed attention mechanisms. Moreover, to fully exploit the supervision signals in unlabeled tables, diverse pre-training objectives have been designed and evaluated, for example, denoising cell values, predicting numerical relationships, and learning a neural SQL executor. This survey aims to provide a review of model designs, pretraining objectives, and downstream tasks for table pre-training, and we further share our thoughts on existing challenges and future opportunities.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2201.09745")
get_code_for_paper("2201.09745")
have("2201.09745")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2201.09745.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2201.09745.svg)](https://syntology.ai/paper/2201.09745)

Connect an agent — have() is free.