Accurate Forgetting for All-in-One Image Restoration Model

TL;DR

This paper introduces a low-cost method for unlearning specific image degradation data from a unified image restoration model, maintaining overall performance while removing sensitive information.

Contribution

It proposes an instance-wise unlearning technique using adversarial examples and gradient ascent to selectively forget certain degradation types in all-in-one image restoration models.

Findings

Effectively unlearns specific degradation data while preserving overall model performance.

Reduces computational cost compared to retraining from scratch.

Maintains robustness of the model after unlearning.

Abstract

Privacy protection has always been an ongoing topic, especially for AI. Currently, a low-cost scheme called Machine Unlearning forgets the private data remembered in the model. Specifically, given a private dataset and a trained neural network, we need to use e.g. pruning, fine-tuning, and gradient ascent to remove the influence of the private dataset on the neural network. Inspired by this, we try to use this concept to bridge the gap between the fields of image restoration and security, creating a new research idea. We propose the scene for the All-In-One model (a neural network that restores a wide range of degraded information), where a given dataset such as haze, or rain, is private and needs to be eliminated from the influence of it on the trained model. Notably, we find great challenges in this task to remove the influence of sensitive data while ensuring that the overall model performance remains robust, which is akin to directing a symphony orchestra without specific instruments while keeping the playing soothing. Here we explore a simple but effective approach: Instance-wise Unlearning through the use of adversarial examples and gradient ascent techniques. Our approach is a low-cost solution compared to the strategy of retraining the model from scratch, where the gradient ascent trick forgets the specified data and the performance of the adversarial sample maintenance model is robust. Through extensive experimentation on two popular unified image restoration models, we show that our approach effectively preserves knowledge of remaining data while unlearning a given degradation type.