Meng Chen, Philip Arthur, Qianyu Feng, Cong Duy Vu Hoang, Yu-Heng Hong, Mahdi Kazemi Moghaddam, Omid Nezami, Duc Thien Nguyen, Gioacchino Tangari, Duy Vu, Thanh Vu, Mark Johnson, Krishnaram Kenthapadi, Don Dharmasiri, Long Duong, Yuan-Fang Li. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Biometric recognition systems are pattern-recognition systems designed to identify users based on a vector of features, which may pertain to physiological or behavioral characteristics. These features are typically stored in databases. Despite the encryption algorithms employed to secure these databases, there is always a chance, however small, that an attacker could decrypt the data and gain access to the full set of features. Subsequently, the attacker could compare the acquired biometric information with templates corresponding to user identities, potentially leading to unauthorized identification. This study focuses on leveraging convolutional neural networks (CNNs) to compress images in a manner that increases the difficulty of restoring them, thereby reducing the effectiveness of such attacks. We employ CNNs to generate deep representations (features) and utilize deep convolutional generative adversarial networks (DCGANs) to reconstruct the images from these features. Finally, the accuracy of the reconstructed images is analyzed using Bayesian beta regression and Random Forest to identify the CNN hyperparameters that have the greatest impact on reconstruction accuracy. Our findings indicate that for simpler datasets, such as black-and-white images of letters or digits (e.g., MNIST), the number of convolutional layers and the presence of pooling layers are the two most critical structural hyperparameters in securing image features. However, for more complex datasets, such as colored facial images (e.g., FaceScrub), activation functions and optimization algorithms become more significant in determining the security of biometric representations.

OBJECTIVE: We conduct a first large-scale analysis of mobile health (mHealth) apps available on Google Play with the goal of providing a comprehensive view of mHealth apps' security features and gauging the associated risks for mHealth users and their data. MATERIALS AND METHODS: We designed an app collection platform that discovered and downloaded more than 20 000 mHealth apps from the Medical and Health & Fitness categories on Google Play. We performed a suite of app code and traffic measurements to highlight a range of app security flaws: certificate security, sensitive or unnecessary permission requests, malware presence, communication security, and security-related concerns raised in user reviews. RESULTS: Compared to baseline non-mHealth apps, mHealth apps generally adopt more reliable signing mechanisms and request fewer dangerous permissions. However, significant fractions of mHealth apps expose users to serious security risks. Specifically, 1.8% of mHealth apps package suspicious codes (eg, trojans), 45.0% rely on unencrypted communication, and as much as 23.0% of personal data (eg, location information and passwords) is sent on unsecured traffic. An analysis of the app reviews reveals that mHealth app users are largely unaware of the surfaced security issues. CONCLUSION: Despite being better aligned with security best practices than non-mHealth apps, mHealth apps are still far from ensuring robust security guarantees. App users, clinicians, technology developers, and policy makers alike should be cognizant of the uncovered security issues and weigh them carefully against the benefits of mHealth apps.

Network operators have recently been developing multi-Gbps traffic monitoring tools on commodity hardware, as part of the packet-processing pipelines realizing software dataplanes. These solutions allow the execution of sophisticated per-packet monitoring using the processing power available on servers. Although advances in packet capture have enabled the interception of packets at high rates, bottlenecks can still arise in the monitoring process as a result of concurrent access to shared processor resources, variations of the traffic skew, and unbalanced packet-rate spikes. In this paper we present an adaptive monitoring framework, MONA, which is resilient to bottlenecks while maintaining the accuracy of monitoring reports above a user-specified threshold. MONA dynamically reduces the measurement task sets under adverse conditions, and reconfigures them to recover potential accuracy degradations. To quantify the monitoring accuracy at run time, MONA adopts a novel task-independent technique that generates accuracy estimates according to recently observed traffic characteristics. With a prototype implementation based on a generic packet-processing pipeline, and using well-known measurements tasks, we show that MONA achieves lossless traffic monitoring for a wide range of conditions, significantly enhances the level of monitoring accuracy, and performs adaptations at the time scale of milliseconds with limited overhead.

The software-defined networking (SDN) paradigm can allow network management solutions to automatically and frequently reconfigure network resources. When developing SDN-based management architectures, it is of paramount importance to design a monitoring system that can provide timely and consistent updates to heterogeneous management applications. To support such applications operating with low latency requirements, the monitoring system should scale with increasing network size and provide precise network views with minimum overhead on the available resources. In this paper, we present a novel, self-adaptive, decentralized framework for resource monitoring in SDN. Our framework enables accurate statistics to be collected with limited burden on the network resources. This is realized through a self-tuning, adaptive monitoring mechanism that automatically adjusts its settings based on the traffic dynamics. We evaluate our proposal based on a realistic use case scenario, where a content distribution service and an on-demand gaming platform are deployed within an ISP network. The results show that reduced monitoring latencies are obtained with the proposed framework, thus enabling shorter reconfiguration control loops. In addition, the proposed adaptive monitoring method achieves significant gain in terms of monitoring overhead, while preserving the performance of the services considered.

Network operators have recently been developing multi-Gbps traffic monitoring tools that execute on commodity hardware and are part of the packet-processing pipelines realizing software dataplanes. These solutions allow sophisticated tasks to be performed on a per-packet basis, without relying on sampling or passive trace analysis, by leveraging the processing power available on servers. Although advances in packet capture have enabled intercepting packets from network cards at high rates, bottlenecks can still arise in the monitoring process as a result of concurrent access to shared processor resources, variations of the traffic skew, and unbalanced packet-rate spikes. In this paper we present an adaptive traffic monitoring approach that copes with emerging bottlenecks by timely detecting changes in the operational conditions and reconfiguring monitoring-related operations for subsets of traffic flows. Our solution performs responsive adaptations at the time scale of milliseconds and does not require a significant amount of resources. To demonstrate the capabilities of our approach we implemented it as part of a generic packet-processing pipeline and show that lossless traffic monitoring can be achieved for a wide range of conditions.

The Software-Defined Networking (SDN) paradigm can allow network management solutions to automatically and frequently reconfigure network resources. When developing SDN-based management architectures, it is of paramount importance to design a monitoring system that can provide frequent and consistent updates to heterogeneous management applications. For the monitoring functionality to scale according to the requirements of large-scale networks a distributed monitoring approach is required. In this paper we present a decentralized approach for resource monitoring in SDN, which is designed to support a wide range of measurement tasks and requirements in terms of monitoring rates and information granularity levels. Our solution leverages effective processing of the monitoring requests to reduce the consumption of limited resources, such as the control plane bandwidth of OpenFlow switches. To demonstrate the benefits of the proposed approach, our evaluation is based on a realistic and demanding use case, where a distributed management application coordinates a content distribution service in an ISP network.