Tethering Broken Themes: Aligning Neural Topic Models with Labels and Authors
This addresses the issue of poor alignment in topic models for researchers and practitioners working with document collections, though it appears incremental as it builds on existing neural topic models.
The paper tackles the problem of neural topic models not aligning well with human intentions by introducing FANToM, a method that incorporates labels and authorship metadata to produce more interpretable topics and author distributions. Experimental results show improvements in topic quality and alignment compared to existing models.
Topic models are a popular approach for extracting semantic information from large document collections. However, recent studies suggest that the topics generated by these models often do not align well with human intentions. Although metadata such as labels and authorship information are available, it has not yet been effectively incorporated into neural topic models. To address this gap, we introduce FANToM, a novel method to align neural topic models with both labels and authorship information. FANToM allows for the inclusion of this metadata when available, producing interpretable topics and author distributions for each topic. Our approach demonstrates greater expressiveness than conventional topic models by learning the alignment between labels, topics, and authors. Experimental results show that FANToM improves existing models in terms of both topic quality and alignment. Additionally, it identifies author interests and similarities.