# Reliable Active Learning from Unreliable Labels via Neural Collapse Geometry

arXiv: https://arxiv.org/abs/2510.09740 · HTML: https://mlsanity.com/paper/2510.09740/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 14 references that are in the map.
- First posted on arXiv 10 Oct 2025.

## Builds on (most cited first, 10 of 10)

- [Deep Bayesian Active Learning with Image Data](https://mlsanity.com/paper/1703.02910/) (2017)
- [Prevalence of Neural Collapse during the terminal phase of deep learning training](https://mlsanity.com/paper/2008.08186/) (2020)
- [Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Bounds](https://mlsanity.com/paper/1906.03671/) (2019)
- [Variational Adversarial Active Learning](https://mlsanity.com/paper/1904.00370/) (2019)
- [Contextual Diversity for Active Learning](https://mlsanity.com/paper/2008.05723/) (2020)
- [NECO: NEural Collapse Based Out-of-distribution detection](https://mlsanity.com/paper/2310.06823/) (2023)
- [How does Weight Correlation Affect the Generalisation Ability of Deep Neural Networks](https://mlsanity.com/paper/2010.05983/) (2020)
- [Active Generalized Category Discovery](https://mlsanity.com/paper/2403.04272/) (2024)
- [Linking Neural Collapse and L2 Normalization with Improved Out-of-Distribution Detection in Deep Neural Networks](https://mlsanity.com/paper/2209.08378/) (2022)
- [Joint Out-of-Distribution Filtering and Data Discovery Active Learning](https://mlsanity.com/paper/2503.02491/) (2025)

## Abstract (arXiv)

Active Learning (AL) promises to reduce annotation cost by prioritizing informative samples, yet its reliability is undermined when labels are noisy or when the data distribution shifts. In practice, annotators make mistakes, rare categories are ambiguous, and conventional AL heuristics (uncertainty, diversity) often amplify such errors by repeatedly selecting mislabeled or redundant samples. We propose Reliable Active Learning via Neural Collapse Geometry (NCAL-R), a framework that leverages the emergent geometric regularities of deep networks to counteract unreliable supervision. Our method introduces two complementary signals: (i) a Class-Mean Alignment Perturbation score, which quantifies how candidate samples structurally stabilize or distort inter-class geometry, and (ii) a Feature Fluctuation score, which captures temporal instability of representations across training checkpoints. By combining these signals, NCAL-R prioritizes samples that both preserve class separation and highlight ambiguous regions, mitigating the effect of noisy or redundant labels. Experiments on ImageNet-100 and CIFAR100 show that NCAL-R consistently outperforms standard AL baselines, achieving higher accuracy with fewer labels, improved robustness under synthetic label noise, and stronger generalization to out-of-distribution data. These results suggest that incorporating geometric reliability criteria into acquisition decisions can make Active Learning less brittle to annotation errors and distribution shifts, a key step toward trustworthy deployment in real-world labeling pipelines. Our code is available at https://github.com/Vision-IIITD/NCAL.