Improving Accuracy-robustness Trade-off via Pixel Reweighted Adversarial Training

• URL: http://arxiv.org/abs/2406.00685v1
• Date: Sun, 2 Jun 2024 09:43:34 GMT
• Title: Improving Accuracy-robustness Trade-off via Pixel Reweighted Adversarial Training
• Authors: Jiacheng Zhang, Feng Liu, Dawei Zhou, Jingfeng Zhang, Tongliang Liu,
• Abstract summary: We propose Pixel-reweighted AdveRsarial Training (PART), a new framework that partially reduces $epsilon$ for less influential pixels. Part achieves a notable improvement in accuracy without compromising robustness on CIFAR-10, SVHN and TinyImagenet-200.
• Score: 61.61368146268329
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Adversarial training (AT) trains models using adversarial examples (AEs), which are natural images modified with specific perturbations to mislead the model. These perturbations are constrained by a predefined perturbation budget $\epsilon$ and are equally applied to each pixel within an image. However, in this paper, we discover that not all pixels contribute equally to the accuracy on AEs (i.e., robustness) and accuracy on natural images (i.e., accuracy). Motivated by this finding, we propose Pixel-reweighted AdveRsarial Training (PART), a new framework that partially reduces $\epsilon$ for less influential pixels, guiding the model to focus more on key regions that affect its outputs. Specifically, we first use class activation mapping (CAM) methods to identify important pixel regions, then we keep the perturbation budget for these regions while lowering it for the remaining regions when generating AEs. In the end, we use these pixel-reweighted AEs to train a model. PART achieves a notable improvement in accuracy without compromising robustness on CIFAR-10, SVHN and TinyImagenet-200, justifying the necessity to allocate distinct weights to different pixel regions in robust classification.

Related papers

• Motion-Aware Adaptive Pixel Pruning for Efficient Local Motion Deblurring [87.56382172827526]
  We propose a trainable mask predictor that identifies blurred regions in the image.<n>We also develop an intra-frame motion analyzer that translates relative pixel displacements into motion trajectories.<n>Our method is trained end-to-end using a combination of reconstruction loss, reblur loss, and mask loss guided by annotated blur masks.
  arXiv  Detail & Related papers  (2025-07-10T12:38:27Z)
• Thera: Aliasing-Free Arbitrary-Scale Super-Resolution with Neural Heat Fields [52.11475771410058]
  Recent approaches to arbitrary-scale single image super-resolution (ASR) use neural fields to represent continuous signals that can be sampled at arbitrary resolutions. Existing methods attempt to mitigate this by approximating an integral version of the field at each scaling factor, compromising both fidelity and generalization. We introduce neural heat fields, a novel neural field formulation that inherently models a physically exact PSF. Our formulation enables analytically correct anti-aliasing at any desired output resolution, and -- unlike supersampling -- at no additional cost.
  arXiv  Detail & Related papers  (2023-11-29T14:01:28Z)
• Pixel-Inconsistency Modeling for Image Manipulation Localization [59.968362815126326]
  Digital image forensics plays a crucial role in image authentication and manipulation localization. This paper presents a generalized and robust manipulation localization model through the analysis of pixel inconsistency artifacts. Experiments show that our method successfully extracts inherent pixel-inconsistency forgery fingerprints.
  arXiv  Detail & Related papers  (2023-09-30T02:54:51Z)
• Look at the Neighbor: Distortion-aware Unsupervised Domain Adaptation for Panoramic Semantic Segmentation [5.352137021024213]
  The aim is to tackle the domain gaps caused by the style disparities and distortion problem from the non-uniformly distributed pixels of equirectangular projection (ERP) We propose a novel UDA framework that can effectively address the distortion problems for panoramic semantic segmentation.
  arXiv  Detail & Related papers  (2023-08-10T10:47:12Z)
• Residual Pattern Learning for Pixel-wise Out-of-Distribution Detection in Semantic Segmentation [38.784135463275305]
  This paper introduces a new residual pattern learning (RPL) module that assists the segmentation model to detect OoD pixels without affecting the inlier segmentation performance. We also present a novel context-robust contrastive learning (CoroCL) that enforces RPL to robustly detect OoD pixels among various contexts. Our approach improves by around 10% FPR and 7% AuPRC the previous state-of-the-art in Fishyscapes, Segment-Me-If-You-Can, and RoadAnomaly datasets.
  arXiv  Detail & Related papers  (2022-11-26T08:32:28Z)
• Pixel-by-Pixel Cross-Domain Alignment for Few-Shot Semantic Segmentation [16.950853152484203]
  We consider the task of semantic segmentation in autonomous driving applications. In this context, aligning the domains is made more challenging by the pixel-wise class imbalance. We propose a novel framework called Pixel-By-Pixel Cross-Domain Alignment (PixDA)
  arXiv  Detail & Related papers  (2021-10-22T08:27:17Z)
• Spatially-Adaptive Image Restoration using Distortion-Guided Networks [51.89245800461537]
  We present a learning-based solution for restoring images suffering from spatially-varying degradations. We propose SPAIR, a network design that harnesses distortion-localization information and dynamically adjusts to difficult regions in the image.
  arXiv  Detail & Related papers  (2021-08-19T11:02:25Z)
• Patch-wise++ Perturbation for Adversarial Targeted Attacks [132.58673733817838]
  We propose a patch-wise iterative method (PIM) aimed at crafting adversarial examples with high transferability. Specifically, we introduce an amplification factor to the step size in each iteration, and one pixel's overall gradient overflowing the $epsilon$-constraint is properly assigned to its surrounding regions. Compared with the current state-of-the-art attack methods, we significantly improve the success rate by 35.9% for defense models and 32.7% for normally trained models.
  arXiv  Detail & Related papers  (2020-12-31T08:40:42Z)
• SegFix: Model-Agnostic Boundary Refinement for Segmentation [75.58050758615316]
  We present a model-agnostic post-processing scheme to improve the boundary quality for the segmentation result that is generated by any existing segmentation model. Motivated by the empirical observation that the label predictions of interior pixels are more reliable, we propose to replace the originally unreliable predictions of boundary pixels by the predictions of interior pixels.
  arXiv  Detail & Related papers  (2020-07-08T17:08:08Z)
• Image Fine-grained Inpainting [89.17316318927621]
  We present a one-stage model that utilizes dense combinations of dilated convolutions to obtain larger and more effective receptive fields. To better train this efficient generator, except for frequently-used VGG feature matching loss, we design a novel self-guided regression loss. We also employ a discriminator with local and global branches to ensure local-global contents consistency.
  arXiv  Detail & Related papers  (2020-02-07T03:45:25Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.