Latent Diffusion Energy-Based Model for Interpretable Text Modeling
This work addresses interpretable text modeling for researchers and practitioners, offering an incremental improvement by combining existing techniques to mitigate sampling issues in latent EBMs.
The paper tackled the problem of poor generation quality and training instability in latent space Energy-Based Models (EBMs) for text modeling by introducing a latent diffusion energy-based model that integrates diffusion models with latent EBMs in a variational framework, achieving superior performance on interpretable text modeling tasks over strong counterparts.
Latent space Energy-Based Models (EBMs), also known as energy-based priors, have drawn growing interests in generative modeling. Fueled by its flexibility in the formulation and strong modeling power of the latent space, recent works built upon it have made interesting attempts aiming at the interpretability of text modeling. However, latent space EBMs also inherit some flaws from EBMs in data space; the degenerate MCMC sampling quality in practice can lead to poor generation quality and instability in training, especially on data with complex latent structures. Inspired by the recent efforts that leverage diffusion recovery likelihood learning as a cure for the sampling issue, we introduce a novel symbiosis between the diffusion models and latent space EBMs in a variational learning framework, coined as the latent diffusion energy-based model. We develop a geometric clustering-based regularization jointly with the information bottleneck to further improve the quality of the learned latent space. Experiments on several challenging tasks demonstrate the superior performance of our model on interpretable text modeling over strong counterparts.