Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AutoTM 2.0: Automatic Topic Modeling Framework for Documents Analysis

Maria Khodorchenko, Nikolay Butakov, Maxim Zuev, Denis Nasonov

TL;DR

This work presents an AutoTM 2.0 framework for optimizing additively regularized topic models and shows that AutoTM 2.0 achieves better performance compared to the previous AutoTM by providing results on 5 datasets with different features and in two different languages.

Abstract

In this work, we present an AutoTM 2.0 framework for optimizing additively regularized topic models. Comparing to the previous version, this version includes such valuable improvements as novel optimization pipeline, LLM-based quality metrics and distributed mode. AutoTM 2.0 is a comfort tool for specialists as well as non-specialists to work with text documents to conduct exploratory data analysis or to perform clustering task on interpretable set of features. Quality evaluation is based on specially developed metrics such as coherence and gpt-4-based approaches. Researchers and practitioners can easily integrate new optimization algorithms and adapt novel metrics to enhance modeling quality and extend their experiments. We show that AutoTM 2.0 achieves better performance compared to the previous AutoTM by providing results on 5 datasets with different features and in two different languages.

AutoTM 2.0: Automatic Topic Modeling Framework for Documents Analysis

TL;DR

This work presents an AutoTM 2.0 framework for optimizing additively regularized topic models and shows that AutoTM 2.0 achieves better performance compared to the previous AutoTM by providing results on 5 datasets with different features and in two different languages.

Abstract

In this work, we present an AutoTM 2.0 framework for optimizing additively regularized topic models. Comparing to the previous version, this version includes such valuable improvements as novel optimization pipeline, LLM-based quality metrics and distributed mode. AutoTM 2.0 is a comfort tool for specialists as well as non-specialists to work with text documents to conduct exploratory data analysis or to perform clustering task on interpretable set of features. Quality evaluation is based on specially developed metrics such as coherence and gpt-4-based approaches. Researchers and practitioners can easily integrate new optimization algorithms and adapt novel metrics to enhance modeling quality and extend their experiments. We show that AutoTM 2.0 achieves better performance compared to the previous AutoTM by providing results on 5 datasets with different features and in two different languages.
Paper Structure (14 sections, 6 figures, 1 table)

This paper contains 14 sections, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related work
  3. Framework design
  4. Dataset preprocessing
  5. Optimization Approaches
  6. Optimization pipelines
  7. Quality estimation
  8. Surrogate modeling
  9. Distributed mode
  10. Framework Performance
  11. Datasets overview
  12. Pipelines comparison
  13. Quality metrics performance
  14. Conclusion

Figures (6)

  • Figure 1: General design of the AutoTM 2.0 framework.
  • Figure 2: An example of a pipeline with 4 training stages.
  • Figure 3: Vectorization scheme for surrogate modeling in graph-based approach.
  • Figure 4: Distributed mode schema.
  • Figure 5: Average fitness values with 90% confidence interval for 5 datasets ((a) 20 Newsgroups (b) Amazon food (c) Banners (d) Hotel reviews (e) Lenta.ru) by the number of used iterations with the usage of a surrogate model.
  • ...and 1 more figures