SLT: Robust Quantum Neural Networks for Noisy-Label Medical Image Classification via Supermartingale-based Label Transition
Quick Answer
The proposed Supermartingale-based Label Transition (SLT) framework enhances quantum neural networks (QNNs) for noisy-label medical image classification, achieving improved stability and performance over traditional methods.
Quick Take
Experiments on small-scale datasets show consistent improvements in classification accuracy, addressing the challenges posed by noisy labels in medical imaging.
Key Points
- SLT offers an anchor-free loss correction framework for QNNs.
- It models entropy reduction as a supermartingale for stable transition updates.
- Experiments show SLT outperforms classic noise-label learning baselines.
- The framework reduces noise-driven oscillations during QNN training.
- Convergence analysis indicates a steady state is reached in transition-refinement.
DeepSignal Analysis
What happened
The Supermartingale-based Label Transition (SLT) framework was introduced to enhance quantum neural networks (QNNs) for medical image classification with noisy labels. This framework aims to improve stability and performance in small-scale datasets, addressing the challenges of noisy labels. Experiments indicate consistent improvements in classification accuracy compared to traditional methods.
Key evidence
- SLT is designed to address the challenges of noisy-label learning in small-scale medical image classification, which has been a significant obstacle for deep neural networks.
- The framework models entropy reduction in predictive distributions as a supermartingale, allowing for dynamic updates to the transition matrix during QNN training.
- Experiments on multiple public small-scale medical image datasets show that SLT consistently improves QNN-based classification, outperforming classic noise-label learning baselines.
Why it matters
The introduction of SLT could represent a significant advancement in the application of quantum neural networks for medical imaging, particularly in scenarios where data is limited and labels may be noisy. By improving classification accuracy, this framework may enhance diagnostic capabilities in medical settings, potentially leading to better patient outcomes. The findings could also stimulate further research into the use of quantum computing in machine learning.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Noisy-label learning in small-scale medical image classification is challenging and hinders the superiority of deep neural networks. Recent studies suggest that quantum neural networks (QNNs) have shown potential in limited-data regimes, yet their use for noisy-label learning remains under-explored. A key obstacle is QNNs' intrinsic "natural smoothness", which may regularize training but also obscure high-confidence samples needed for noise-transition estimation. We propose Supermartingale-based Label Transition (SLT), an anchor-free loss correction framework for robust QNN-based medical image classification under noisy labels. SLT models entropy reduction in predictive distributions as a supermartingale and uses its monotonic behavior to identify stable transition-matrix refinement steps. This enables dynamic transition updates while reducing noise-driven oscillations during QNN training. We further provide a convergence analysis showing that the proposed transition-refinement process reaches a steady state. Experiments on multiple public small-scale medical image datasets demonstrate that SLT consistently improves QNN-based classification and stably outperforms classic noise-label learning baselines under synthetic and real-world label noise.
| Comments: | Preprint, under review |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG) |
| Cite as: | arXiv:2607.16293 [cs.CV] |
| (or arXiv:2607.16293v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.16293 arXiv-issued DOI via DataCite |
Submission history
From: Jun Zhuang [view email]
[v1]
Mon, 13 Jul 2026 06:37:29 UTC (2,893 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →ProMoE-FL: Prototype-conditioned Mixture of Experts for Multimodal Federated Learning with Missing Modalities
ProMoE-FL introduces a Prototype-conditioned Mixture-of-Experts framework for multimodal federated learning, effectively addressing missing modalities. It outperforms existing methods on four chest X-ray datasets, demonstrating superior feature synthesis capabilities in both homogeneous and heterogeneous settings.