: This highly regarded paper from arXiv challenges current automated metrics and advocates for human-centered evaluation in neural models.
: Best for smaller datasets where interpretability is key and computational resources are limited.
: A comprehensive review available on Springer that categorizes methods, applications, and future directions for the field.
: Published in Information Processing & Management , this paper presents a way to improve how distinct and interpretable topics are within a model.
: Typically use Variational Autoencoders (VAEs) and often outperform classical models on large-scale datasets, though evaluation can be more complex.
: This is the seminal paper by David Blei and colleagues that introduced the LDA model, forming the foundation for most modern topic modeling techniques.
Searching for specific papers on topic modeling involves looking at foundational research and state-of-the-art developments in and Latent Dirichlet Allocation (LDA) . Top Papers for Topic Modeling