# Active Learning with Tabular Language Models

arXiv: https://arxiv.org/abs/2211.04128 · HTML: https://mlsanity.com/paper/2211.04128/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 11 references that are in the map.
- First posted on arXiv 8 Nov 2022.

## Builds on (most cited first, 8 of 8)

- [Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Bounds](https://mlsanity.com/paper/1906.03671/) (2019)
- [A Survey of Deep Active Learning](https://mlsanity.com/paper/2009.00236/) (2020)
- [Deep Active Learning for Named Entity Recognition](https://mlsanity.com/paper/1707.05928/) (2017)
- [Diverse mini-batch Active Learning](https://mlsanity.com/paper/1901.05954/) (2019)
- [Revisiting Uncertainty-based Query Strategies for Active Learning with Transformers](https://mlsanity.com/paper/2107.05687/) (2021)
- [Table Pre-training: A Survey on Model Architectures, Pre-training Objectives, and Downstream Tasks](https://mlsanity.com/paper/2201.09745/) (2022)
- [TURL: Table Understanding through Representation Learning](https://mlsanity.com/paper/2006.14806/) (2020)
- [Named Entity Recognition in Industrial Tables using Tabular Language Models](https://mlsanity.com/paper/2209.14812/) (2022)

## Abstract (arXiv)

Despite recent advancements in tabular language model research, real-world applications are still challenging. In industry, there is an abundance of tables found in spreadsheets, but acquisition of substantial amounts of labels is expensive, since only experts can annotate the often highly technical and domain-specific tables. Active learning could potentially reduce labeling costs, however, so far there are no works related to active learning in conjunction with tabular language models. In this paper we investigate different acquisition functions in a real-world industrial tabular language model use case for sub-cell named entity recognition. Our results show that cell-level acquisition functions with built-in diversity can significantly reduce the labeling effort, while enforced table diversity is detrimental. We further see open fundamental questions concerning computational efficiency and the perspective of human annotators.