English

One Configuration to Rule Them All? Towards Hyperparameter Transfer in Topic Models using Multi-Objective Bayesian Optimization

Computation and Language 2022-02-16 v1

Abstract

Topic models are statistical methods that extract underlying topics from document collections. When performing topic modeling, a user usually desires topics that are coherent, diverse between each other, and that constitute good document representations for downstream tasks (e.g. document classification). In this paper, we conduct a multi-objective hyperparameter optimization of three well-known topic models. The obtained results reveal the conflicting nature of different objectives and that the training corpus characteristics are crucial for the hyperparameter selection, suggesting that it is possible to transfer the optimal hyperparameter configurations between datasets.

Keywords

Cite

@article{arxiv.2202.07631,
  title  = {One Configuration to Rule Them All? Towards Hyperparameter Transfer in Topic Models using Multi-Objective Bayesian Optimization},
  author = {Silvia Terragni and Ismail Harrando and Pasquale Lisena and Raphael Troncy and Elisabetta Fersini},
  journal= {arXiv preprint arXiv:2202.07631},
  year   = {2022}
}

Comments

17 pages, 7 figures

R2 v1 2026-06-24T09:39:19.719Z