Exploring and Leveraging Class Vectors for Classifier Editing

• URL: http://arxiv.org/abs/2510.11268v2
• Date: Fri, 17 Oct 2025 00:53:12 GMT
• Title: Exploring and Leveraging Class Vectors for Classifier Editing
• Authors: Jaeik Kim, Jaeyoung Do,
• Abstract summary: We introduce Class Vectors, which capture class-specific representation adjustments during fine-tuning.<n>Whereas task vectors encode task-level changes in weight space, Class Vectors disentangle each class's adaptation in the latent space.<n>We show that Class Vectors capture each class's semantic shift and that classifier editing can be achieved either by steering latent features along these vectors or by mapping them into weight space to update the decision boundaries.
• Score: 6.328734263302503
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Image classifiers play a critical role in detecting diseases in medical imaging and identifying anomalies in manufacturing processes. However, their predefined behaviors after extensive training make post hoc model editing difficult, especially when it comes to forgetting specific classes or adapting to distribution shifts. Existing classifier editing methods either focus narrowly on correcting errors or incur extensive retraining costs, creating a bottleneck for flexible editing. Moreover, such editing has seen limited investigation in image classification. To overcome these challenges, we introduce Class Vectors, which capture class-specific representation adjustments during fine-tuning. Whereas task vectors encode task-level changes in weight space, Class Vectors disentangle each class's adaptation in the latent space. We show that Class Vectors capture each class's semantic shift and that classifier editing can be achieved either by steering latent features along these vectors or by mapping them into weight space to update the decision boundaries. We also demonstrate that the inherent linearity and orthogonality of Class Vectors support efficient, flexible, and high-level concept editing via simple class arithmetic. Finally, we validate their utility in applications such as unlearning, environmental adaptation, adversarial defense, and adversarial trigger optimization.

Related papers

• Semantic Prioritization in Visual Counterfactual Explanations with Weighted Segmentation and Auto-Adaptive Region Selection [50.68751788132789]
  This study introduces an innovative methodology named as Weighted Semantic Map with Auto-adaptive Candidate Editing Network (WSAE-Net)<n>The generation of the weighted semantic map is designed to maximize the reduction of non-semantic feature units that need to be computed.<n>The auto-adaptive candidate editing sequences are designed to determine the optimal computational order among the feature units to be processed.
  arXiv  Detail & Related papers  (2025-11-17T05:34:10Z)
• Towards Fine-Grained Adaptation of CLIP via a Self-Trained Alignment Score [11.74414842618874]
  We show that modeling fine-grained cross-modal interactions during adaptation produces more accurate, class-discriminative pseudo-labels.<n>We introduce Fine-grained Alignment and Interaction Refinement (FAIR), an innovative approach that dynamically aligns localized image features with descriptive language embeddings.<n>Our approach, FAIR, delivers a substantial performance boost in fine-grained unsupervised adaptation, achieving a notable overall gain of 2.78%.
  arXiv  Detail & Related papers  (2025-07-13T12:38:38Z)
• REACT: Representation Extraction And Controllable Tuning to Overcome Overfitting in LLM Knowledge Editing [42.89229070245538]
  We introduce REACT, a framework for precise and controllable knowledge editing.<n>In the initial phase, we utilize tailored stimuli to extract latent factual representations.<n>In the second phase, we apply controllable perturbations to hidden states using the obtained vector with a magnitude scalar.
  arXiv  Detail & Related papers  (2025-05-25T01:57:06Z)
• Generalize Your Face Forgery Detectors: An Insertable Adaptation Module Is All You Need [3.5424095074777533]
  We introduce an insertable adaptation module that can adapt a trained off-the-shelf detector using only online unlabeled test data.<n> Specifically, we first present a learnable class prototype-based classifier that generates predictions from the revised features and prototypes.<n>We also propose a nearest feature calibrator to further improve prediction accuracy and reduce the impact of noisy pseudo-labels during self-training.
  arXiv  Detail & Related papers  (2024-12-30T08:48:04Z)
• Versatile Teacher: A Class-aware Teacher-student Framework for Cross-domain Adaptation [2.9748058103007957]
  We introduce a novel teacher-student model named Versatile Teacher (VT) VT considers class-specific detection difficulty and employs a two-step pseudo-label selection mechanism to generate more reliable pseudo labels. Our method demonstrates promising results on three benchmark datasets, and extends the alignment methods for widely-used one-stage detectors.
  arXiv  Detail & Related papers  (2024-05-20T03:31:43Z)
• Anomaly Detection using Ensemble Classification and Evidence Theory [62.997667081978825]
  We present a novel approach for novel detection using ensemble classification and evidence theory. A pool selection strategy is presented to build a solid ensemble classifier. We use uncertainty for the anomaly detection approach.
  arXiv  Detail & Related papers  (2022-12-23T00:50:41Z)
• Location-Aware Self-Supervised Transformers [74.76585889813207]
  We propose to pretrain networks for semantic segmentation by predicting the relative location of image parts. We control the difficulty of the task by masking a subset of the reference patch features visible to those of the query. Our experiments show that this location-aware pretraining leads to representations that transfer competitively to several challenging semantic segmentation benchmarks.
  arXiv  Detail & Related papers  (2022-12-05T16:24:29Z)
• Towards Counterfactual Image Manipulation via CLIP [106.94502632502194]
  Existing methods can achieve realistic editing of different visual attributes such as age and gender of facial images. We investigate this problem in a text-driven manner with Contrastive-Language-Image-Pretraining (CLIP) We design a novel contrastive loss that exploits predefined CLIP-space directions to guide the editing toward desired directions from different perspectives.
  arXiv  Detail & Related papers  (2022-07-06T17:02:25Z)
• Prototypical Classifier for Robust Class-Imbalanced Learning [64.96088324684683]
  We propose textitPrototypical, which does not require fitting additional parameters given the embedding network. Prototypical produces balanced and comparable predictions for all classes even though the training set is class-imbalanced. We test our method on CIFAR-10LT, CIFAR-100LT and Webvision datasets, observing that Prototypical obtains substaintial improvements compared with state of the arts.
  arXiv  Detail & Related papers  (2021-10-22T01:55:01Z)
• Learning a Domain Classifier Bank for Unsupervised Adaptive Object Detection [48.19258721979389]
  In this paper, we propose a fine-grained domain alignment approach for object detectors based on deep networks. We develop a bare object detector with the proposed fine-grained domain alignment mechanism as the adaptive detector. Experiments on three popular transferring benchmarks demonstrate the effectiveness of our method.
  arXiv  Detail & Related papers  (2020-07-06T09:12:46Z)

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.