# Some Robustness Properties of Label Cleaning

arXiv: https://arxiv.org/abs/2509.11379 · HTML: https://mlsanity.com/paper/2509.11379/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 13 of its 53 references that are in the map.
- First posted on arXiv 14 Sep 2025.

## Builds on (most cited first, 13 of 13)

- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [Cross-Entropy Loss Functions: Theoretical Analysis and Applications](https://mlsanity.com/paper/2304.07288/) (2023)
- [Calibration and Consistency of Adversarial Surrogate Losses](https://mlsanity.com/paper/2104.09658/) (2021)
- [Structured Prediction with Partial Labelling through the Infimum Loss](https://mlsanity.com/paper/2003.00920/) (2020)
- [On the Consistency of Ordinal Regression Methods](https://mlsanity.com/paper/1408.2327/) (2014)
- [A Universal Growth Rate for Learning with Smooth Surrogate Losses](https://mlsanity.com/paper/2405.05968/) (2024)
- [On Structured Prediction Theory with Calibrated Convex Surrogate Losses](https://mlsanity.com/paper/1703.02403/) (2017)
- [Multi-Label Learning with Stronger Consistency Guarantees](https://mlsanity.com/paper/2407.13746/) (2024)
- [Enhanced H-Consistency Bounds](https://mlsanity.com/paper/2407.13722/) (2024)
- [Consistent Structured Prediction with Max-Min Margin Markov Networks](https://mlsanity.com/paper/2007.01012/) (2020)
- [How many labelers do you have? A closer look at gold-standard labels](https://mlsanity.com/paper/2206.12041/) (2022)
- [Learning from Imperfect Annotations](https://mlsanity.com/paper/2004.03473/) (2020)
- [Don't Label Twice: Quantity Beats Quality when Comparing Binary Classifiers on a Budget](https://mlsanity.com/paper/2402.02249/) (2024)

## Abstract (arXiv)

We demonstrate that learning procedures that rely on aggregated labels, e.g., label information distilled from noisy responses, enjoy robustness properties impossible without data cleaning. This robustness appears in several ways. In the context of risk consistency -- when one takes the standard approach in machine learning of minimizing a surrogate (typically convex) loss in place of a desired task loss (such as the zero-one mis-classification error) -- procedures using label aggregation obtain stronger consistency guarantees than those even possible using raw labels. And while classical statistical scenarios of fitting perfectly-specified models suggest that incorporating all possible information -- modeling uncertainty in labels -- is statistically efficient, consistency fails for ``standard'' approaches as soon as a loss to be minimized is even slightly mis-specified. Yet procedures leveraging aggregated information still converge to optimal classifiers, highlighting how incorporating a fuller view of the data analysis pipeline, from collection to model-fitting to prediction time, can yield a more robust methodology by refining noisy signals.