Abstract:Denoising Diffusion Probabilistic Models (DDPMs) have achieved significant success in generation tasks. Nevertheless, the exposure bias issue, i.e., the natural discrepancy between the training (the output of each step is calculated individually by a given input) and inference (the output of each step is calculated based on the input iteratively obtained based on the model), harms the performance of DDPMs. To our knowledge, few works have tried to tackle this issue by modifying the training process for DDPMs, but they still perform unsatisfactorily due to 1) partially modeling the discrepancy and 2) ignoring the prediction error accumulation. To address the above issues, in this paper, we propose a multi-step denoising scheduled sampling (MDSS) strategy to alleviate the exposure bias for DDPMs. Analyzing the formulations of the training and inference of DDPMs, MDSS 1) comprehensively considers the discrepancy influence of prediction errors on the output of the model (the Gaussian noise) and the output of the step (the calculated input signal of the next step), and 2) efficiently models the prediction error accumulation by using multiple iterations of a mathematical formulation initialized from one-step prediction error obtained from the model. The experimental results, compared with previous works, demonstrate that our approach is more effective in mitigating exposure bias in DDPM, DDIM, and DPM-solver. In particular, MDSS achieves an FID score of 3.86 in 100 sample steps of DDIM on the CIFAR-10 dataset, whereas the second best obtains 4.78. The code will be available on GitHub.

Enhancing Diffusion Models for Inverse Problems with Covariance-Aware Posterior Sampling

Improving Diffusion Models for Inverse Problems Using Optimal Posterior Covariance

Multi-Step Denoising Scheduled Sampling: Towards Alleviating Exposure Bias for Diffusion Models

Enhancing Diffusion Posterior Sampling for Inverse Problems by Integrating Crafted Measurements

Diffusion Model Based Posterior Sampling for Noisy Linear Inverse Problems

Tweedie Moment Projected Diffusions For Inverse Problems

Divide-and-Conquer Posterior Sampling for Denoising Diffusion Priors

Think Twice Before You Act: Improving Inverse Problem Solving With MCMC

Diffusion Posterior Sampling for General Noisy Inverse Problems

Diffusion Prior-Based Amortized Variational Inference for Noisy Inverse Problems

Principled Probabilistic Imaging using Diffusion Models as Plug-and-Play Priors

Improving Decoupled Posterior Sampling for Inverse Problems using Data Consistency Constraint

A Variational Perspective on Solving Inverse Problems with Diffusion Models

Gaussian is All You Need: A Unified Framework for Solving Inverse Problems via Diffusion Posterior Sampling

Diffusion with Forward Models: Solving Stochastic Inverse Problems Without Direct Supervision

Variational Diffusion Posterior Sampling with Midpoint Guidance

Improving Diffusion Inverse Problem Solving with Decoupled Noise Annealing

Ambient Diffusion Posterior Sampling: Solving Inverse Problems with Diffusion Models trained on Corrupted Data

Improving Probabilistic Diffusion Models With Optimal Covariance Matching

Accelerating Diffusion Models for Inverse Problems through Shortcut Sampling