Reproduce “Topic Modelling Meets Deep Neural Networks: A Survey” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

He Zhao, Dinh Phung, Viet Huynh, Yuan Jin, Lan Du, Wray Buntine

Researchers from Monash University provide a comprehensive survey of Neural Topic Models (NTMs), categorizing over a hundred variants based on their underlying deep learning frameworks. The work details how NTMs address limitations of traditional topic models by leveraging neural networks for improved inference, scalability, and topic coherence across diverse applications.

Key insights

  • Neural Topic Models (NTMs) combine the interpretability of traditional topic models with the expressive power and flexibility of deep neural networks.
  • The survey proposes a taxonomy categorizing NTMs based on their backbone frameworks: VAE-NTMs, DocNADE-NTMs, GAN-NTMs, and GNN-NTMs.
  • NTMs find applications across various domains, including text mining, information retrieval, recommendation systems, and medical text analysis.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.