Improved Autoregressive Modeling with Distribution Smoothing
This work addresses a specific bottleneck in autoregressive modeling for image generation, offering an incremental improvement in sample quality.
The paper tackled the problem of poor sample quality in autoregressive models for image generation by incorporating randomized smoothing to model a smoothed data distribution and then reverse it, resulting in drastically improved sample quality on synthetic and real-world image datasets while maintaining competitive likelihoods.
While autoregressive models excel at image compression, their sample quality is often lacking. Although not realistic, generated images often have high likelihood according to the model, resembling the case of adversarial examples. Inspired by a successful adversarial defense method, we incorporate randomized smoothing into autoregressive generative modeling. We first model a smoothed version of the data distribution, and then reverse the smoothing process to recover the original data distribution. This procedure drastically improves the sample quality of existing autoregressive models on several synthetic and real-world image datasets while obtaining competitive likelihoods on synthetic datasets.