DiffETM: Diffusion Process Enhanced Embedded Topic Model
This work addresses a limitation in topic modeling for researchers and practitioners, but it is incremental as it builds directly on the existing ETM framework.
The authors tackled the oversimplified logistic normal distribution assumption in embedded topic models (ETM) by introducing a diffusion process into the sampling of document-topic distributions, resulting in improved topic modeling performance validated on two mainstream datasets.
The embedded topic model (ETM) is a widely used approach that assumes the sampled document-topic distribution conforms to the logistic normal distribution for easier optimization. However, this assumption oversimplifies the real document-topic distribution, limiting the model's performance. In response, we propose a novel method that introduces the diffusion process into the sampling process of document-topic distribution to overcome this limitation and maintain an easy optimization process. We validate our method through extensive experiments on two mainstream datasets, proving its effectiveness in improving topic modeling performance.