Superclass-Guided Representation Disentanglement for Spurious Correlation Mitigation

• URL: http://arxiv.org/abs/2508.08570v1
• Date: Tue, 12 Aug 2025 02:16:04 GMT
• Title: Superclass-Guided Representation Disentanglement for Spurious Correlation Mitigation
• Authors: Chenruo Liu, Hongjun Liu, Zeyu Lai, Yiqiu Shen, Chen Zhao, Qi Lei,
• Abstract summary: We propose a method that leverages the semantic structure inherent in class labels to reduce reliance on spurious features.<n>Our model employs gradient-based attention guided by a pre-trained vision-language model to disentangle superclass-relevant and irrelevant features.<n>Our approach achieves robustness to more complex spurious correlations without the need to annotate any source samples.
• Score: 16.791911185682256
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: To enhance group robustness to spurious correlations, prior work often relies on auxiliary annotations for groups or spurious features and assumes identical sets of groups across source and target domains. These two requirements are both unnatural and impractical in real-world settings. To overcome these limitations, we propose a method that leverages the semantic structure inherent in class labels--specifically, superclass information--to naturally reduce reliance on spurious features. Our model employs gradient-based attention guided by a pre-trained vision-language model to disentangle superclass-relevant and irrelevant features. Then, by promoting the use of all superclass-relevant features for prediction, our approach achieves robustness to more complex spurious correlations without the need to annotate any source samples. Experiments across diverse datasets demonstrate that our method significantly outperforms baselines in domain generalization tasks, with clear improvements in both quantitative metrics and qualitative visualizations.

Related papers

• Multi-Level Aware Preference Learning: Enhancing RLHF for Complex Multi-Instruction Tasks [81.44256822500257]
  RLHF has emerged as a predominant approach for aligning artificial intelligence systems with human preferences.<n> RLHF exhibits insufficient compliance capabilities when confronted with complex multi-instruction tasks.<n>We propose a novel Multi-level Aware Preference Learning (MAPL) framework, capable of enhancing multi-instruction capabilities.
  arXiv  Detail & Related papers  (2025-05-19T08:33:11Z)
• Refining Filter Global Feature Weighting for Fully-Unsupervised Clustering [0.0]
  In unsupervised learning, effective clustering plays a vital role in revealing patterns and insights from unlabeled data.<n>This paper explores feature weighting for clustering and presents new weighting strategies, including methods based on SHAP (SHapley Additive exPlanations)<n>Our empirical evaluations demonstrate that feature weighting based on SHAP can enhance unsupervised clustering quality, achieving up to a 22.69% improvement over other weighting methods.
  arXiv  Detail & Related papers  (2025-03-12T13:14:09Z)
• Enhancing Interpretability Through Loss-Defined Classification Objective in Structured Latent Spaces [5.2542280870644715]
  We introduce Latent Boost, a novel approach that integrates advanced distance metric learning into supervised classification tasks.<n>Latent Boost improves classification interpretability, as demonstrated by higher Silhouette scores, while accelerating training convergence.
  arXiv  Detail & Related papers  (2024-12-11T16:25:17Z)
• Cluster Specific Representation Learning [1.6727186769396276]
  Despite its widespread application, there is no established definition of a good'' representation.<n>We propose a downstream-agnostic formulation: when inherent clusters exist in the data, the representations should be specific to each cluster.<n>Under this idea, we develop a meta-algorithm that jointly learns cluster-specific representations and cluster assignments.
  arXiv  Detail & Related papers  (2024-12-04T16:59:37Z)
• Learning from Neighbors: Category Extrapolation for Long-Tail Learning [62.30734737735273]
  We offer a novel perspective on long-tail learning, inspired by an observation: datasets with finer granularity tend to be less affected by data imbalance.<n>We introduce open-set auxiliary classes that are visually similar to existing ones, aiming to enhance representation learning for both head and tail classes.<n>To prevent the overwhelming presence of auxiliary classes from disrupting training, we introduce a neighbor-silencing loss.
  arXiv  Detail & Related papers  (2024-10-21T13:06:21Z)
• GCC: Generative Calibration Clustering [55.44944397168619]
  We propose a novel Generative Clustering (GCC) method to incorporate feature learning and augmentation into clustering procedure. First, we develop a discrimirative feature alignment mechanism to discover intrinsic relationship across real and generated samples. Second, we design a self-supervised metric learning to generate more reliable cluster assignment.
  arXiv  Detail & Related papers  (2024-04-14T01:51:11Z)
• Leveraging vision-language models for fair facial attribute classification [19.93324644519412]
  General-purpose vision-language model (VLM) is a rich knowledge source for common sensitive attributes. We analyze the correspondence between VLM predicted and human defined sensitive attribute distribution. Experiments on multiple benchmark facial attribute classification datasets show fairness gains of the model over existing unsupervised baselines.
  arXiv  Detail & Related papers  (2024-03-15T18:37:15Z)
• Enhancing Neural Subset Selection: Integrating Background Information into Set Representations [53.15923939406772]
  We show that when the target value is conditioned on both the input set and subset, it is essential to incorporate an textitinvariant sufficient statistic of the superset into the subset of interest. This ensures that the output value remains invariant to permutations of the subset and its corresponding superset, enabling identification of the specific superset from which the subset originated.
  arXiv  Detail & Related papers  (2024-02-05T16:09:35Z)
• Learning Debiased and Disentangled Representations for Semantic Segmentation [52.35766945827972]
  We propose a model-agnostic and training scheme for semantic segmentation. By randomly eliminating certain class information in each training iteration, we effectively reduce feature dependencies among classes. Models trained with our approach demonstrate strong results on multiple semantic segmentation benchmarks.
  arXiv  Detail & Related papers  (2021-10-31T16:15:09Z)
• Top-$k$ Regularization for Supervised Feature Selection [11.927046591097623]
  We introduce a novel, simple yet effective regularization approach, named top-$k$ regularization, to supervised feature selection. We show that the top-$k$ regularization is effective and stable for supervised feature selection.
  arXiv  Detail & Related papers  (2021-06-04T01:12:47Z)
• Improving cluster recovery with feature rescaling factors [2.4366811507669124]
  We argue that the rescaling procedure should not treat all features identically. We introduce a feature rescaling method that takes into account the within-cluster degree of relevance of each feature. Our comprehensive simulation study, carried out on real and synthetic data, with and without noise features, clearly demonstrates that clustering methods that use the proposed data normalization strategy clearly outperform those that use traditional data normalization.
  arXiv  Detail & Related papers  (2020-12-01T13:29:35Z)
• Learning What Makes a Difference from Counterfactual Examples and Gradient Supervision [57.14468881854616]
  We propose an auxiliary training objective that improves the generalization capabilities of neural networks. We use pairs of minimally-different examples with different labels, a.k.a counterfactual or contrasting examples, which provide a signal indicative of the underlying causal structure of the task. Models trained with this technique demonstrate improved performance on out-of-distribution test sets.
  arXiv  Detail & Related papers  (2020-04-20T02:47:49Z)

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.