Brief

A new research paper proposes a unified framework for integrating various data-driven priors into Bayesian inverse problems. The study demonstrates how diverse priors, including regularization-by-denoising, normalizing flow-based priors, and score-based generative models, can be unified through their score functions. This approach allows for effective integration into a proposed sampling algorithm, with experimental validation in image inpainting and super-resolution tasks. AI

Researchers have developed MaCo-GAN, a new framework for single image super-resolution that addresses artifact generation in conventional Generative Adversarial Networks (GANs). This novel approach replaces the standard adversarial loss with a supervised contrastive objective, utilizing a dynamic synthesizer to create challenging, perceptually plausible fake images. The MaCo-GAN framework trains the generator to distinguish between on-manifold and off-manifold fakes, leading to improved perception-distortion trade-offs across various benchmarks. AI