This research establishes the first population risk bounds for Kolmogorov-Arnold Networks (KANs) trained with mini-batch SGD, including differentially private SGD (DP-SGD) with correlated noise. It covers more practical scenarios than prior KAN theory and provides sharper results for fixed-second-layer specializations.
This research explores combining privacy-preserving machine learning, specifically differential privacy and active learning, for the maintenance of bio-inspired soft robotics. The work addresses the challenge of retraining predictive maintenance models without exposing proprietary data during critical recovery windows.
This paper systematically explores text decomposition and budget distribution in Differentially Private (DP) text obfuscation. It evaluates multiple techniques for chunking texts and allocating the epsilon budget, revealing that these design choices significantly impact the results even with comparable privacy budgets.
This research explores a differential-privacy based approach to improve generalization and prevent overfitting in Deep Neural Networks. Overfitting, where models learn noise and perform poorly on unseen data, is a growing challenge in modern AI systems.