No Prior, No Leakage: Revisiting Reconstruction Attacks in Trained Neural Networks
This work addresses privacy concerns for machine learning practitioners by showing that reconstruction attacks are less reliable than previously thought, though it is incremental in refining theoretical understanding.
The paper tackled the problem of training data reconstruction attacks in neural networks by analyzing their limitations, proving that without prior knowledge, reconstruction is fundamentally unreliable and exact duplication occurs only by chance. It found that more extensively trained networks are less susceptible to attacks, reconciling privacy with generalization.
The memorization of training data by neural networks raises pressing concerns for privacy and security. Recent work has shown that, under certain conditions, portions of the training set can be reconstructed directly from model parameters. Some of these methods exploit implicit bias toward margin maximization, suggesting that properties often regarded as beneficial for generalization may actually compromise privacy. Yet despite striking empirical demonstrations, the reliability of these attacks remains poorly understood and lacks a solid theoretical foundation. In this work, we take a complementary perspective: rather than designing stronger attacks, we analyze the inherent weaknesses and limitations of existing reconstruction methods and identify conditions under which they fail. We rigorously prove that, without incorporating prior knowledge about the data, there exist infinitely many alternative solutions that may lie arbitrarily far from the true training set, rendering reconstruction fundamentally unreliable. Empirically, we further demonstrate that exact duplication of training examples occurs only by chance. Our results refine the theoretical understanding of when training set leakage is possible and offer new insights into mitigating reconstruction attacks. Remarkably, we demonstrate that networks trained more extensively, and therefore satisfying implicit bias conditions more strongly -- are, in fact, less susceptible to reconstruction attacks, reconciling privacy with the need for strong generalization in this setting.