Neural Collapse in Test-Time Adaptation

• URL: http://arxiv.org/abs/2512.10421v1
• Date: Thu, 11 Dec 2025 08:34:58 GMT
• Title: Neural Collapse in Test-Time Adaptation
• Authors: Xiao Chen, Zhongjing Du, Jiazhen Huang, Xu Jiang, Li Lu, Jingyan Jiang, Zhi Wang,
• Abstract summary: Test-Time Adaptation (TTA) enhances robustness to out-of-distribution data by updating the model online during inference.<n>Recently, Neural Collapse (NC) has been proposed as an emergent geometric property of deep neural networks (DNNs)<n>We identify that the performance degradation stems from sample-wise misalignment in adaptation which exacerbates under larger distribution shifts.<n>We propose NCTTA, a novel feature-classifier alignment method with hybrid targets to mitigate the impact of unreliable pseudo-labels.
• Score: 12.636904518839303
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Test-Time Adaptation (TTA) enhances model robustness to out-of-distribution (OOD) data by updating the model online during inference, yet existing methods lack theoretical insights into the fundamental causes of performance degradation under domain shifts. Recently, Neural Collapse (NC) has been proposed as an emergent geometric property of deep neural networks (DNNs), providing valuable insights for TTA. In this work, we extend NC to the sample-wise level and discover a novel phenomenon termed Sample-wise Alignment Collapse (NC3+), demonstrating that a sample's feature embedding, obtained by a trained model, aligns closely with the corresponding classifier weight. Building on NC3+, we identify that the performance degradation stems from sample-wise misalignment in adaptation which exacerbates under larger distribution shifts. This indicates the necessity of realigning the feature embeddings with their corresponding classifier weights. However, the misalignment makes pseudo-labels unreliable under domain shifts. To address this challenge, we propose NCTTA, a novel feature-classifier alignment method with hybrid targets to mitigate the impact of unreliable pseudo-labels, which blends geometric proximity with predictive confidence. Extensive experiments demonstrate the effectiveness of NCTTA in enhancing robustness to domain shifts. For example, NCTTA outperforms Tent by 14.52% on ImageNet-C.

Related papers

• Space Alignment Matters: The Missing Piece for Inducing Neural Collapse in Long-Tailed Learning [8.526510873614034]
  We show that under class-balanced conditions, the class feature means and classifier weights spontaneously align into a simplexangular tight frame (ETF)<n>In long-tailed regimes, however, severe sample imbalance tends to prevent the emergence of the NC phenomenon, resulting in poor generalization performance.<n>We propose three explicit alignment strategies that plug-and-play into existing long-tail methods without architectural change.
  arXiv  Detail & Related papers  (2025-11-25T07:38:40Z)
• Open-World Test-Time Adaptation with Hierarchical Feature Aggregation and Attention Affine [17.151364853811128]
  Test-time adaptation (TTA) refers to adjusting the model during the testing phase to cope with changes in sample distribution.<n>We propose a Hierarchical Ladder Network that extracts OOD features from class tokens aggregated across all Transformer layers.<n>We also introduce an Attention Affine Network (AAN) that adaptively refines the self-attention mechanism conditioned on the token information to better adapt to domain drift.
  arXiv  Detail & Related papers  (2025-11-16T14:05:23Z)
• Neural Collapse under Gradient Flow on Shallow ReLU Networks for Orthogonally Separable Data [52.737775129027575]
  We show that gradient flow on a two-layer ReLU network for classifying orthogonally separable data provably exhibits Neural Collapse (NC)<n>We reveal the role of the implicit bias of the training dynamics in facilitating the emergence of NC.
  arXiv  Detail & Related papers  (2025-10-24T01:36:19Z)
• Control the GNN: Utilizing Neural Controller with Lyapunov Stability for Test-Time Feature Reconstruction [15.066912209426542]
  The performance of graph neural networks (GNNs) is susceptible to discrepancies between training and testing sample distributions.<n>We propose a novel node feature reconstruction method grounded in Lyapunov stability theory.<n>We validate the effectiveness of our approach through extensive experiments across multiple datasets, demonstrating significant performance improvements.
  arXiv  Detail & Related papers  (2024-10-13T03:34:19Z)
• PseudoNeg-MAE: Self-Supervised Point Cloud Learning using Conditional Pseudo-Negative Embeddings [55.55445978692678]
  PseudoNeg-MAE enhances global feature representation of point cloud masked autoencoders by making them both discriminative and sensitive to transformations.<n>We propose a novel loss that explicitly penalizes invariant collapse, enabling the network to capture richer transformation cues while preserving discriminative representations.
  arXiv  Detail & Related papers  (2024-09-24T07:57:21Z)
• SINDER: Repairing the Singular Defects of DINOv2 [61.98878352956125]
  Vision Transformer models trained on large-scale datasets often exhibit artifacts in the patch token they extract. We propose a novel fine-tuning smooth regularization that rectifies structural deficiencies using only a small dataset.
  arXiv  Detail & Related papers  (2024-07-23T20:34:23Z)
• FRGNN: Mitigating the Impact of Distribution Shift on Graph Neural Networks via Test-Time Feature Reconstruction [13.21683198528012]
  A distribution shift can adversely affect the test performance of Graph Neural Networks (GNNs) We propose FR-GNN, a general framework for GNNs to conduct feature reconstruction. Notably, the reconstructed node features can be directly utilized for testing the well-trained model.
  arXiv  Detail & Related papers  (2023-08-18T02:34:37Z)
• Collapsed Inference for Bayesian Deep Learning [36.1725075097107]
  We introduce a novel collapsed inference scheme that performs Bayesian model averaging using collapsed samples. A collapsed sample represents uncountably many models drawn from the approximate posterior. Our proposed use of collapsed samples achieves a balance between scalability and accuracy.
  arXiv  Detail & Related papers  (2023-06-16T08:34:42Z)
• Benign Overfitting in Deep Neural Networks under Lazy Training [72.28294823115502]
  We show that when the data distribution is well-separated, DNNs can achieve Bayes-optimal test error for classification. Our results indicate that interpolating with smoother functions leads to better generalization.
  arXiv  Detail & Related papers  (2023-05-30T19:37:44Z)
• Can pruning improve certified robustness of neural networks? [106.03070538582222]
  We show that neural network pruning can improve empirical robustness of deep neural networks (NNs) Our experiments show that by appropriately pruning an NN, its certified accuracy can be boosted up to 8.2% under standard training. We additionally observe the existence of certified lottery tickets that can match both standard and certified robust accuracies of the original dense models.
  arXiv  Detail & Related papers  (2022-06-15T05:48:51Z)
• Attribute-Guided Adversarial Training for Robustness to Natural Perturbations [64.35805267250682]
  We propose an adversarial training approach which learns to generate new samples so as to maximize exposure of the classifier to the attributes-space. Our approach enables deep neural networks to be robust against a wide range of naturally occurring perturbations.
  arXiv  Detail & Related papers  (2020-12-03T10:17:30Z)

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.