Federated Learning (FL) has amassed notable attention for its ability to preserve user privacy while emphasizing the retainment of model training efficiency. Due to this potential, FL has been integrated in many domains, such as healthcare, finance, law, and industrial engineering, where data cannot be easily exchanged due to sensitive information and strict privacy laws. However, current research has indicated that FL protocols are easily compromised by active data reconstruction attacks employed by actively dishonest servers. The malicious modification of global model parameters allows an actively dishonest server to obtain a direct copy of users’ private data via gradient inversion. This class of attacks is highly underexplored and continues to be a major challenge due to the intense threat model. In this article, we propose OASIS as a scalable and modality-agnostic defense based on data augmentation that counteracts active data reconstruction attacks while preserving model performance. To generalize our defense, we uncover the intuition behind gradient inversion that enables these attacks and theoretically establish the conditions by which the defense can be considered robust regardless of attack design. From this, we formulate our defense with data augmentation that illustrates its ability to undermine the attack principle. We evaluate OASIS on five real-world datasets–two image-based (ImageNet and CIFAR100) and three text-based (Wikitext, Stack Overflow, and Shakespeare)–which span diverse uses cases such as vision tasks and language modeling. Comprehensive evaluations on these datasets exhibit the efficacy of OASIS and highlight its feasibility as a solution.

Combating misinformation in the digital age has been a challenging subject with significant social implications, as misinformation continues to shape public responses to elections, pandemics, and other critical events. Despite extensive research efforts, traditional mitigation approaches have fallen short, largely due to their narrow focus on detection accuracy while overlooking the social and emotional influences that drive user behavior. This article proposes to frame information belief as a social norm, enabling us to track and predict its evolution at the individual user level. By operating on an individual level, we can identify and target specific users susceptible to adopting certain norms ahead of time, providing a more powerful approach to mitigating the effects of misinformation. To accomplish this, we develop a temporal graph neural network equipped with projection operators capable of forecasting arbitrary time horizons. The network is trained to predict individual users’ tipping points—the moments when they transition from resistance to acceptance of particular information norms. We conduct rigorous analysis of the temporal smoothness of our predictions, leading to an algorithm that queries the model to identify critical moments before users reach their tipping points, enabling preemptive action. To our knowledge, this represents the first successful attempt at predicting individual tipping points in real-world information networks, marking a significant advance in our ability to address misinformation through preventive rather than reactive measures.