
Transformer-Representation Neural Topic Model (TNTM) unifies transformer embeddings with probabilistic modelling for enhanced topic coherence and diversity.


coremsg

probabilistic-topic-modelling-with-transformer-representations-a-comprehensive-analysis


Probabilistic Topic Modelling with Transformer Representations: A Comprehensive Analysis


title_rewrite


The author proposes the Transformer-Representation Neural Topic Model (TNTM) to combine transformer-based embedding spaces with probabilistic modeling for topic representation. The approach unifies powerful topics based on transformer embeddings with fully probabilistic modeling.
