Centroid Spectral Coupling: Implicit Alignment and Representation Diversity in Self-Supervised Learning

M. Hadi Sepanj, Benyamin Ghojogh, Saed Moradi, Paul FiegutharXiv 2026 · arXiv

TL;DR

Centroid Spectral Coupling (CSC) is a new self-supervised learning objective that replaces complex, multi-term loss functions with a single, elegant spectral term. By applying a ridge-based spectral functional to the unrenormalized centroids of augmented image views, CSC implicitly balances view alignment and representation diversity while providing theoretical guarantees against representation collapse.

What Problem Does This Paper Address?

In Self-Supervised Learning (SSL), the goal is to learn useful features from unlabeled data. To do this, models typically try to satisfy two competing requirements:

  1. Alignment: Different "views" (augmentations) of the same image should result in similar representations.
  2. Diversity: Representations of different images should be spread out across the available feature space to avoid "collapse," where the model maps every input to the same constant vector.

Existing "non-contrastive" methods like VICReg or Barlow Twins usually handle this by combining multiple loss terms (e.g., one for invariance, one for variance, one for covariance) and manually tuning weights to balance them. This is often computationally expensive and theoretically messy. The authors of this paper ask: Can we achieve alignment, diversity, and collapse resistance using just one simple loss term?

Methods

The authors introduce Centroid Spectral Coupling (CSC). The process works as follows:

  • Centroid Construction: For each image in a batch, the model generates two augmented views. These views are independently normalized. Then, their arithmetic mean—the centroid—is calculated.
  • The "Secret Sauce": Crucially, these centroids are not renormalized. If the two views are perfectly aligned, the centroid magnitude is 1. If they are poorly aligned (e.g., pointing in opposite directions), the magnitude shrinks toward 0. This magnitude naturally encodes alignment information.
  • Ridge Spectral Objective: The loss function is based on the "ridge effective dimension" of the second-moment matrix of these centroids. Mathematically, it uses a strictly concave functional of the eigenvalues of the centroid matrix.
  • Implicit Coupling: Because the functional is concave, it rewards spreading "energy" across many spectral directions (promoting diversity). Because it operates on unrenormalized centroids, it simultaneously rewards maximizing centroid magnitude (promoting alignment).

Datasets

The researchers evaluated CSC using a ResNet-50 backbone on three standard image classification benchmarks:

  • CIFAR-10
  • CIFAR-100
  • Imagenette (a 10-class subset of ImageNet)

Key Insights and Ideas

  • Alignment via Magnitude: The most elegant insight is that by not renormalizing the average of two vectors, you turn a diversity-seeking spectral loss into a joint alignment-diversity loss. As shown in Figure 1, renormalizing the centroids actually destroys the model's ability to align views.
Figure 1. Effect of centroid renormalization on augmented-view alignment and spectral diversity. Training dynamics on CIFAR-10 using ResNet-50. Left: Positive-pair cosine alignment across training epochs. Right: Representation effective rank across training epochs. Unlike the original CSC, centroid renormalization substantially reduces positive-pair alignment despite producing a higher effective rank. CSC maintains strong alignment and nontrivial spectral diversity, illustrating the importance of preserving centroid magnitudes.
  • The $\kappa$ Trade-off: CSC introduces a single hyperparameter, $\kappa$. The authors prove that $\kappa$ acts as a dial: increasing it prioritizes alignment, while decreasing it prioritizes spectral rank (diversity). This is visualized in their sensitivity study (Figure 2).
Figure 2. Effect of the ridge parameter κ on alignment and rep- resentation diversity. Training trajectories on CIFAR-100 using ResNet-50. Alignment is measured on normalized projector out- puts, whereas effective rank is computed from centered encoder- feature covariance. Stars indicate the final training epoch. Increas- ing κ favors alignment, while smaller values preserve greater spec- tral diversity.
  • Mathematical Robustness: The authors prove that "complete collapse" (where all outputs are the same) is a strict saddle point. This means that while it's a stationary point for the loss, it's unstable, and standard optimization techniques like SGD are theoretically guaranteed to move away from it.

Key Results

CSC consistently outperformed other prominent non-contrastive methods (like VICReg, Barlow Twins, and MMCR) across all tested benchmarks.

  • Linear Probing: CSC achieved the highest Top-1 accuracy on all three datasets, notably outperforming the next best baseline by 2.15% on CIFAR-100.
  • kNN Evaluation: CSC showed particularly strong local structure, beating baselines by over 5.4% in kNN Top-1 accuracy on CIFAR-100.
Table 1. Frozen ResNet-50 evaluation on CIFAR-10, CIFAR-100, and Imagenette. Held-out classification accuracy (%) using linear probing and kNN classification (seed 0). All evaluated methods are non-contrastive SSL approaches. Input resolutions are 32 × 32 for CIFAR and 128 × 128 for Imagenette. Bold indicates the best result and underlining the second-best result within each dataset and metric.

Limitations and Future Work

While the results are impressive on small-to-medium datasets, the paper notes a few areas for future exploration:

  • Scalability: The current experiments use ResNet-50 on relatively small image resolutions. Testing CSC on massive datasets like ImageNet-1k with Vision Transformers (ViTs) would be the next logical step to prove its industrial viability.
  • Semantic Gap: While the paper provides rigorous mathematical proofs for representation-level diversity, it acknowledges that the relationship between these spectral properties and high-level "semantic" classes is still an area for deeper study.
  • Computational Cost: Spectral methods involve eigenvalue decompositions (SVD), which can be more expensive than simple vector operations, though the authors provide a minimal PyTorch implementation that suggests this is manageable for standard batch sizes.

Summary updated

Abstract

Self-Supervised Learning (SSL) aims to align augmented views while preserving representation diversity and preventing collapse. Existing non-contrastive methods often address these goals using multiple weighted loss terms. We introduce Centroid Spectral Coupling (CSC), a single-term SSL objective that implicitly couples view alignment, representation diversity, and collapse resistance. CSC constructs centroids from independently normalized augmented-view pairs and applies a ridge-based spectral objective to their uncentered second-moment matrix. Preserving centroid magnitudes encodes view agreement while encouraging balanced representation energy across spectral directions. We establish an exact decomposition into nonnegative alignment and spectral-imbalance components, characterize the geometry of global optima, and derive finite-loss guarantees for alignment and diversity. We also show that complete constant-output collapse is a strict saddle in normalized representation space. Experiments on CIFAR-10, CIFAR-100, and Imagenette show that CSC achieves the highest linear Top-1 accuracy among the evaluated non-contrastive methods on all three benchmarks, and the highest kNN Top-1 accuracy on both CIFAR datasets. A ridge-parameter sensitivity study reveals an empirical alignment--diversity trade-off: increasing $κ$ strengthens view alignment but reduces representation effective rank. These results show that a unified spectral objective yields competitive visual representations without explicit contrastive negatives or separately weighted alignment and diversity penalties.
BibTeX
@misc{sepanj2026centroid,
  author = {M. Hadi Sepanj and Benyamin Ghojogh and Saed Moradi and Paul Fieguth},
  title = {{Centroid Spectral Coupling: Implicit Alignment and Representation Diversity in Self-Supervised Learning}},
  year = {2026},
  eprint = {2611.22158},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2611.22158},
}
Figure