Introduction

Foundation models (FMs) are leading a paradigm shift in artificial intelligence, especially ChatGPT in natural language processing and Stable Diffusion in computer vision (Gupta, Tarun, et al. 2024). Nevertheless, they have achieved less promising results in the tabular domain, though tabular data are one of the most common modalities in both industries and sciences. A long-term vision for foundation models in the tabular domain is to develop FMs with better understanding, reasoning, and analysing tabular data which require a higher-level intelligence.

Tables play an essential role in real-world applications.

Citation

Tu, Ruibo. (Sep. 2024). Tabular Foundation Models. Ruibo-blog. https://turuibo.github.io/posts/2024-09-11-tabular-foundation-models/.

or

@article{tu2024tabularllms,
  title   = "Tabular Foundation Models",
  author  = "Tu, Ruibo",
  journal = "turuibo.github.io",
  year    = "2024",
  month   = "Sep",
  url     = "https://turuibo.github.io/posts/2024-09-11-tabular-foundation-models/"
}