# Automatic Generation of Topic Labels

arXiv: https://arxiv.org/abs/2006.00127 · HTML: https://mlsanity.com/paper/2006.00127/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 May 2020.

## Built on by (most cited first, 8 of 8)

- [Not All Comments are Equal: Insights into Comment Moderation from a Topic-Aware Model](https://mlsanity.com/paper/2109.10033/) (2021)
- [Iterative Topic Taxonomy Induction with LLMs: A Case Study of Electoral Advertising](https://mlsanity.com/paper/2510.15125/) (2025)
- [Automatic Scene-based Topic Channel Construction System for E-Commerce](https://mlsanity.com/paper/2210.02643/) (2022)
- [Moving beyond word lists: towards abstractive topic labels for human-like topics of scientific documents](https://mlsanity.com/paper/2211.05599/) (2022)
- [Labeled Interactive Topic Models](https://mlsanity.com/paper/2311.09438/) (2023)
- [An Information Retrieval and Extraction Tool for Covid-19 Related Papers](https://mlsanity.com/paper/2401.16430/) (2024)
- [Efficient Topic Extraction via Graph-Based Labeling: A Lightweight Alternative to Deep Models](https://mlsanity.com/paper/2511.04248/) (2025)
- [Industry-Aligned Granular Topic Modeling](https://mlsanity.com/paper/2601.11762/) (2026)

## Abstract (arXiv)

Topic modelling is a popular unsupervised method for identifying the underlying themes in document collections that has many applications in information retrieval. A topic is usually represented by a list of terms ranked by their probability but, since these can be difficult to interpret, various approaches have been developed to assign descriptive labels to topics. Previous work on the automatic assignment of labels to topics has relied on a two-stage approach: (1) candidate labels are retrieved from a large pool (e.g. Wikipedia article titles); and then (2) re-ranked based on their semantic similarity to the topic terms. However, these extractive approaches can only assign candidate labels from a restricted set that may not include any suitable ones. This paper proposes using a sequence-to-sequence neural-based approach to generate labels that does not suffer from this limitation. The model is trained over a new large synthetic dataset created using distant supervision. The method is evaluated by comparing the labels it generates to ones rated by humans.