Voice communication is convenient while also poses risks of privacy leakage, due to potential interception or eavesdropping during voice transmission. Current protections of voice privacy are almost entirely controlled by communication service providers (CSPs), which operate as a black-box to users thus hard to fully trust. To take back the control of user privacy, in this paper, we introduce SecHeadset, an end-to-end solution for secure voice communication based on voice obfuscation, which is plug-and-play and compatible with various CSPs. Our solution involves two parts. First, we design a voice-like noise masking scheme for voice obfuscation. The noise, mimicking voice characteristics, could effectively obscure users' voices while demonstrating resilience against noise reduction methods. Second, we develop a protocol that enables efficient channel state estimation and secure information exchange between two communication entities. Based on this information, we propose a lightweight algorithm for voice retrieval during communication. We develop a prototype of SecHeadset and evaluate its performance with 8 widely-used applications, including Telegram and Skype. It reduces the voice recognition accuracy of various adversaries to below 15% while maintaining communication quality. We also integrate SecHeadset with off-the-shelf portable devices and verify its real-world effectiveness.

The malicious use and widespread dissemination of deepfake pose a significant crisis of trust. Current deepfake detection models can generally recognize forgery images by training on a large dataset. However, the accuracy of detection models degrades significantly on images generated by new deepfake methods due to the difference in data distribution. To tackle this issue, we present a novel incremental learning framework that improves the generalization of deepfake detection models by continual learning from a small number of new samples. To cope with different data distributions, we propose to learn a domain-invariant representation based on supervised contrastive learning, preventing overfit to the insufficient new data. To mitigate catastrophic forgetting, we regularize our model in both feature-level and label-level based on a multi-perspective knowledge distillation approach. Finally, we propose to select both central and hard representative samples to update the replay set, which is beneficial for both domain-invariant representation learning and rehearsal-based knowledge preserving. We conduct extensive experiments on four benchmark datasets, obtaining the new state-of-the-art average forgetting rate of 7.01 and average accuracy of 85.49 on FF++, DFDC-P, DFD, and CDF2. Our code is released at \textcolorblue https://github.com/DeepFakeIL/DFIL.

Influence maximization (IM) is a problem of selecting the most influential vertices with a limited budget under a given propagation model. A significant challenge faced by many existing algorithms pertains to their inability to reconcile the competing goals of solution quality and computational efficiency, rendering them unsuitable in large-scale social networks. In this paper, we propose an adaptive community-based influence maximization algorithm, named AComA, to solve the IM problem with a balance of effectiveness and efficiency. First, we introduce a community detection method to divide a large-scale network into several communities. An adaptive indicator is then defined to identify vertices with high propagation values in divided community networks. Based on community detection and the adaptive influence indicator, the number of candidate vertices is reduced, which significantly reduces the search space of the problem. Second, to select the final seed set from these candidate vertices, a genetic algorithm (GA) is introduced. The crossover and mutation operations are modified explicitly to adapt to the IM problem. By extracting information from the local neighborhood and the global community structure, AComA achieves a more accurate measurement of the influence spread for each vertex. The method proposed in this paper is tested on several real-world datasets. The experimental results show that AComA is promising.