Improving Neural Topic Modeling with Semantically-Grounded Soft Label Distributions
This work addresses a domain-specific problem for researchers and practitioners in natural language processing by improving topic modeling, though it is incremental as it builds on existing neural topic models.
The paper tackled the problem of neural topic models overlooking contextual information and data sparsity by proposing a method using language models to create semantically-grounded soft labels, resulting in substantial improvements in topic coherence and purity on three datasets.
Traditional neural topic models are typically optimized by reconstructing the document's Bag-of-Words (BoW) representations, overlooking contextual information and struggling with data sparsity. In this work, we propose a novel approach to construct semantically-grounded soft label targets using Language Models (LMs) by projecting the next token probabilities, conditioned on a specialized prompt, onto a pre-defined vocabulary to obtain contextually enriched supervision signals. By training the topic models to reconstruct the soft labels using the LM hidden states, our method produces higher-quality topics that are more closely aligned with the underlying thematic structure of the corpus. Experiments on three datasets show that our method achieves substantial improvements in topic coherence, purity over existing baselines. Additionally, we also introduce a retrieval-based metric, which shows that our approach significantly outperforms existing methods in identifying semantically similar documents, highlighting its effectiveness for retrieval-oriented applications.