# Addressing Model Vulnerability to Distributional Shifts over Image Transformation Sets

## TL;DR

This paper introduces a novel approach to enhance computer vision models' robustness against distributional shifts caused by image transformations by identifying vulnerable regions and iteratively training with targeted data augmentation.

## Contribution

It formulates a combinatorial optimization framework to pinpoint vulnerable image regions and integrates this into a training process for improved robustness.

## Key findings

- Models trained with the proposed method are more robust to image manipulations.
- The approach improves performance on classification and segmentation tasks under distributional shifts.
- Empirical results demonstrate increased resilience against content-preserving transformations.

## Abstract

We are concerned with the vulnerability of computer vision models to distributional shifts. We formulate a combinatorial optimization problem that allows evaluating the regions in the image space where a given model is more vulnerable, in terms of image transformations applied to the input, and face it with standard search algorithms. We further embed this idea in a training procedure, where we define new data augmentation rules according to the image transformations that the current model is most vulnerable to, over iterations. An empirical evaluation on classification and semantic segmentation problems suggests that the devised algorithm allows to train models that are more robust against content-preserving image manipulations and, in general, against distributional shifts.