Work

Beyond Coherence: Improving Temporal Consistency and Interpretability in Dynamic Topic Models

Year: 2026

Type: conference-paper

Abstract: Dynamic topic models aim to reveal how themes emerge, evolve, and dissolve in timestamped corpora, but existing approaches still face three major challenges: (i) encoders capture bag-of-words statistics but fail to align with the rich semantic priors of large pretrained language models, (ii) temporal linkages are often modeled as rigid one-to-one chains, limiting the ability to track non-linear evolution such as topic splits or merges, and (iii) interpretability remains shallow, relying on noisy top-word lists that obscure thematic clarity.We propose L-DNTM (LLM-Augmented for Dynamic Neural Topic Model), a variational framework designed to capture more faithful temporal trajectories.Our model integrates three key components: multiobjective distillation to inject PLM-derived semantic knowledge into the encoder, entropyregularized optimal transport to align entire topic constellations across time for smooth yet flexible evolution, and LLM-guided refinement to sharpen topicword distributions for improved interpretability.Extensive experiments on diverse corpora show that L-DNTM yields more coherent, temporally consistent, and interpretable topic dynamics, and further enhances downstream classification and clustering tasks.

Source: Findings of the Association for Computational Linguistics: EACL

DOI: 10.18653/v1/2026.findings-eacl.187