# Pushing the (Decision) Boundaries: Dynamically Calibrating Differentially Private Noise to Explainability in Federated Learning

arXiv: https://arxiv.org/abs/2609.03851 · HTML: https://mlsanity.com/paper/2609.03851/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 49 references that are in the map.
- First posted on arXiv 3 Sep 2026.

## Builds on (most cited first, 24 of 24)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Communication-Efficient Learning of Deep Networks from Decentralized Data](https://mlsanity.com/paper/1602.05629/) (2016)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [Deep Learning with Differential Privacy](https://mlsanity.com/paper/1607.00133/) (2016)
- [ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseases](https://mlsanity.com/paper/1705.02315/) (2017)
- [Inverting Gradients -- How easy is it to break privacy in federated learning?](https://mlsanity.com/paper/2003.14053/) (2020)
- [Differentially Private Learning with Adaptive Clipping](https://mlsanity.com/paper/1905.03871/) (2019)
- [A Patient-Centric Dataset of Images and Metadata for Identifying Melanomas Using Clinical Context](https://mlsanity.com/paper/2008.07360/) (2020)
- [A Consistent and Efficient Evaluation Strategy for Attribution Methods](https://mlsanity.com/paper/2202.00449/) (2022)
- [Generalization in Adaptive Data Analysis and Holdout Reuse](https://mlsanity.com/paper/1506.02629/) (2015)
- [Individual Privacy Accounting via a Renyi Filter](https://mlsanity.com/paper/2008.11193/) (2020)
- [Accelerated Federated Learning with Decoupled Adaptive Optimization](https://mlsanity.com/paper/2207.07223/) (2022)
- [Fully Adaptive Composition in Differential Privacy](https://mlsanity.com/paper/2203.05481/) (2022)
- [Manifold Integrated Gradients: Riemannian Geometry for Feature Attribution](https://mlsanity.com/paper/2405.09800/) (2024)
- [Adap DP-FL: Differentially Private Federated Learning with Adaptive Noise](https://mlsanity.com/paper/2211.15893/) (2022)
- [Not All Federated Learning Algorithms Are Created Equal: A Performance Evaluation Study](https://mlsanity.com/paper/2403.17287/) (2024)
- [Beyond the Calibration Point: Mechanism Comparison in Differential Privacy](https://mlsanity.com/paper/2406.08918/) (2024)
- [Clients Collaborate: Flexible Differentially Private Federated Learning with Guaranteed Improvement of Utility-Privacy Trade-off](https://mlsanity.com/paper/2402.07002/) (2024)
- [Client-Level Differential Privacy via Adaptive Intermediary in Federated Medical Imaging](https://mlsanity.com/paper/2307.12542/) (2023)
- [Interplay between Federated Learning and Explainable Artificial Intelligence: a Scoping Review](https://mlsanity.com/paper/2411.05874/) (2024)
- [Privacy accounting varepsilonconomics: Improving differential privacy composition via a posteriori bounds](https://mlsanity.com/paper/2205.03470/) (2022)
- [Gaussian Smoothing in Saliency Maps: The Stability-Fidelity Trade-Off in Neural Network Interpretability](https://mlsanity.com/paper/2411.05837/) (2024)
- [The Impact of Generalization Techniques on the Interplay Among Privacy, Utility, and Fairness in Image Classification](https://mlsanity.com/paper/2412.11951/) (2024)

## Abstract (arXiv)

Federated Learning (FL) with Differential Privacy (DP) is increasingly adopted to preserve data confidentiality in distributed machine learning. However, DP noise distorts learned representations and degrades explanation fidelity, limiting differentially private FL where trustworthy explanations are required, such as assistive clinical diagnosis. Prior work adapted DP noise with static feature-importance signals, restricting explainability to post hoc analysis and precluding noise calibration to explanation quality during training. We propose XCal-FL, a closed-loop, explainability-driven local training algorithm for image classification in cross-silo FL that dynamically calibrates DP noise from three complementary signals: (1) prediction logit variations, measuring causal influence on model confidence, (2) counterfactual margins, capturing decision-boundary sensitivity, and (3) saliency concentration, quantifying spatial coherence of model attention, while enforcing formal DP guarantees via adaptive privacy accounting. Experiments on three medical imaging datasets across varying FL configurations show that XCal-FL yields more accurate and interpretable global models, improving predictive performance by over 10% and explanation fidelity by up to 5× over static-noise FL, and outperforming state-of-the-art adaptive DP methods in fidelity. XCal-FL also achieves higher privacy-budget efficiency, turning each unit of cumulative privacy loss into larger gains in both accuracy and explanation fidelity. Our analysis further reveals that, unlike predictive performance, which scales roughly linearly with privacy loss, explanation fidelity exhibits non-linear dynamics. These findings suggest explainability is a distinct dimension of the privacy trade-off that cannot be inferred from utility alone, with implications for training and privacy-budget allocation in decision-critical applications.