Spectral Surgery: Class-Targeted Post-Hoc Rebalancing via Hessian Spike Perturbation

• URL: http://arxiv.org/abs/2605.07790v1
• Date: Fri, 08 May 2026 14:27:41 GMT
• Title: Spectral Surgery: Class-Targeted Post-Hoc Rebalancing via Hessian Spike Perturbation
• Abstract summary: Hessian spectrum of trained deep networks exhibits a characteristic structure: a continuous bulk of near-zero eigenvalues and a small number of large outlier eigenvalues (spikes)<n>We propose Spectral Surgery, a post-hoc optimization method that directly perturbs model weights along spike eigenvectors to rebalance per-class accuracy without retraining.<n>We obtain encouraging results on CIFAR-10 and ISIC 2019 on both balanced accuracy and standard deviation.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The Hessian spectrum of trained deep networks exhibits a characteristic structure: a continuous bulk of near-zero eigenvalues and a small number of large outlier eigenvalues (spikes), confirming the relevance of Random Matrix Theory in deep learning. The spike count matches the number of classes minus one. While prior work has described this structure, no method has exploited it operationally to improve classification performance. We propose Spectral Surgery, a post-hoc optimization method that directly perturbs model weights along spike eigenvectors to rebalance per-class accuracy without retraining. We introduce (i) a spike-class sensitivity matrix that quantifies the directional derivative of each class's accuracy along each spike eigenvector, (ii) a constrained optimization of perturbation coefficients that targets weak classes while preserving strong ones, and (iii) an adaptive amplitude control that raises or lowers the perturbation budget based on iteration-level improvement signals. We obtain encouraging results on CIFAR-10 and ISIC-2019 on both balanced accuracy and standard deviation.

Related papers

• When Does Scale-Invariant Optimization Become Unstable? An Exact Schedule Law with Weight Decay [6.625270757109195]
  We show that a single quantity captures all schedule and decay forcing, while norm growth induces an opposing geometric self-quenching effect.<n>This yields a sharp boundary that cleanly separates contraction- and expansion-dominated effective learning rate regimes.
  arXiv  Detail & Related papers  (2026-09-08T17:46:17Z)
• Spectral-Aware Analytic Class-Incremental Learning for Long-Tailed Distributions [17.58096703161218]
  Analytic Continual Learning offers a computationally efficient alternative to gradient-based approaches.<n>Recent ACL methods falter significantly in Class-Incremental Learning scenarios characterized by Long-Tailed distributions.<n>We propose Geometry-Spectral Rectification (GSR), a theoretically grounded framework that treats long-tailed learning as a spectral regularization problem.
  arXiv  Detail & Related papers  (2026-07-24T22:12:15Z)
• Gradient Clipping Beyond Vector Norms: A Spectral Approach for Matrix-Valued Parameters [48.3175117923623]
  Most gradient clipping rules treat all parameters as heavy vectors and ignore the matrix structure of modern vectors.<n>We show empirically that data outliers often amplify only a small number of leading singular values.<n>Motivated by this phenomenon, we propose clipping, which stabilizes by singular values.
  arXiv  Detail & Related papers  (2026-05-12T09:24:59Z)
• ORTHOBO: Orthogonal Bayesian Hyperparameter Optimization [31.941089890706348]
  noisy acquisition estimates can lead to unstable decisions.<n>We introduce OrthoBO: a Bayesian optimization framework with ensemble surrogates and an outer log transformation.<n>We show theoretically that our estimator preserves the target, leads to variance reduction, and improves pairwise ranking stability.
  arXiv  Detail & Related papers  (2026-05-07T15:49:03Z)
• Variational Rectification Inference for Learning with Noisy Labels [74.85528327499662]
  We propose variational rectification inference (VRI) to formulate the adaptive rectification for loss functions.<n>VRI is constructed as a hierarchical Bayes by treating the rectifying vector as a latent variable.<n>By introducing the variational term in VRI, the conditional posterior is estimated accurately and avoids collapsing to a Dirac delta function.
  arXiv  Detail & Related papers  (2026-03-18T01:25:08Z)
• Purifying Shampoo: Investigating Shampoo's Heuristics by Decomposing its Preconditioner [25.56447816096372]
  Recent success of Shampoo in the computationPerf contest has sparked renewed interest in Kroneckerfactorization-based algorithms for training neural networks.<n>We show that grafting from Adam directly mitigates the staleness and mis-scaling of the preconditioner's eigenvalues.<n>We propose an adaptive criterion for determining the eigenbasis frequency motivated by terminating a warm-started QR algorithm.
  arXiv  Detail & Related papers  (2025-06-04T05:55:41Z)
• Benign Overfitting and the Geometry of the Ridge Regression Solution in Binary Classification [75.01389991485098]
  We show that ridge regression has qualitatively different behavior depending on the scale of the cluster mean vector.<n>In regimes where the scale is very large, the conditions that allow for benign overfitting turn out to be the same as those for the regression task.
  arXiv  Detail & Related papers  (2025-03-11T01:45:42Z)
• Deep Imbalanced Regression via Hierarchical Classification Adjustment [50.19438850112964]
  Regression tasks in computer vision are often formulated into classification by quantizing the target space into classes. The majority of training samples lie in a head range of target values, while a minority of samples span a usually larger tail range. We propose to construct hierarchical classifiers for solving imbalanced regression tasks. Our novel hierarchical classification adjustment (HCA) for imbalanced regression shows superior results on three diverse tasks.
  arXiv  Detail & Related papers  (2023-10-26T04:54:39Z)
• The Lipschitz-Variance-Margin Tradeoff for Enhanced Randomized Smoothing [85.85160896547698]
  Real-life applications of deep neural networks are hindered by their unsteady predictions when faced with noisy inputs and adversarial attacks. We show how to design an efficient classifier with a certified radius by relying on noise injection into the inputs. Our novel certification procedure allows us to use pre-trained models with randomized smoothing, effectively improving the current certification radius in a zero-shot manner.
  arXiv  Detail & Related papers  (2023-09-28T22:41:47Z)
• Consensus-Adaptive RANSAC [104.87576373187426]
  We propose a new RANSAC framework that learns to explore the parameter space by considering the residuals seen so far via a novel attention layer. The attention mechanism operates on a batch of point-to-model residuals, and updates a per-point estimation state to take into account the consensus found through a lightweight one-step transformer.
  arXiv  Detail & Related papers  (2023-07-26T08:25:46Z)
• Noise Stability Optimization for Finding Flat Minima: A Hessian-based Regularization Approach [18.009376840944284]
  We present an algorithm that can effectively regularize the Hessian loss matrices leading to regions with bound loss surfaces. Our approach is effective for improving generalization in pretraining CLIP and chain-of-thought fine-tuning datasets.
  arXiv  Detail & Related papers  (2023-06-14T14:58:36Z)
• Consistency Regularization for Certified Robustness of Smoothed Classifiers [89.72878906950208]
  A recent technique of randomized smoothing has shown that the worst-case $ell$-robustness can be transformed into the average-case robustness. We found that the trade-off between accuracy and certified robustness of smoothed classifiers can be greatly controlled by simply regularizing the prediction consistency over noise.
  arXiv  Detail & Related papers  (2020-06-07T06:57:43Z)

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.