MJMarc Juarez

Marc Juarez

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Marc Juarez has a total of 36 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Brewing Under Pressure: More Realistic Tor Flow Correlation Considering Low FPRs

    by Haeseung Jeon, Yeonseo Cho, Nate Mathews, Rob Jansen, Marc Juarez, Se Eun Oh

  • Open Access

    What Breaks Local Watermarks? A Robustness Benchmark for Local Invisible Image Watermarking

    by Kai Yao, Bence Szilágyi, Sebestyén Kamp, Máté Poór, Máté Szilveszter, Matyas K. Zsoldos, Marc Juarez

  • Open Access

    Can model attribution bridge AI's accountability gap in safety-critical domains?

    by Marc Juarez

    The current trend of deploying machine learning models as remote services obscures which specific models users are interacting with, exacerbating an accountability gap that becomes ever more pressing as these technologies make their way into safety-critical domains. In this paper, we examine how model attribution, a recently proposed accountability property, can address this gap. We review recent technical approaches to provide model attribution and identify limitations, such as computational costs, latency overheads and insufficient guarantees, that render them inadequate for many safety-critical applications. We conclude that although model attribution is an important step towards accountability, current technical solutions fall short of the requirements of safety-critical domains. This article is part of the theme issue ‘Safe, secure and robust AI for safety-critical systems’.


  • Open Access

    Black-Box Evasion Attacks on Data-Driven Open RAN Apps: Tailored Design and Experimental Evaluation

    by Pranshav Gajjar, Molham Khoja, Abiodun Ganiyu, Marc Juarez, Mahesh K. Marina, Andrew Lehane, Vijay K. Shah

    The impending adoption of Open Radio Access Network (O-RAN) is fueling innovation in the RAN towards data-driven operation. Unlike traditional RAN where the RAN data and its usage is restricted within proprietary and monolithic RAN equipment, the O-RAN architecture opens up access to RAN data (i.e., network telemetry), via RAN intelligent controllers (RICs), to third-party machine learning (ML) powered applications - rApps and xApps - to optimize RAN operations. Consequently, a major focus has been placed on leveraging RAN data to unlock greater efficiency gains. However, there is an increasing recognition that RAN data access to apps could become a source of vulnerability and be exploited by malicious actors. Motivated by this, we carry out a comprehensive investigation of data vulnerabilities on both xApps and rApps, respectively hosted in Near- and Non-real-time (RT) RIC components of O-RAN. Our investigation begins by qualitatively analyzing the O-RAN security mechanisms and limitations relevant to xApps and rApps, such as their onboarding authentication process and RIC database access mechanisms. Considering a threat model informed by this analysis, we design a viable and effective black-box evasion attack strategy targeting O-RAN RIC Apps while accounting for the stringent timing constraints (particularly for xApps) and attack effectiveness. The attack strategy employs four key techniques: the model cloning algorithm, input-specific perturbations, universal adversarial perturbations (UAPs), and targeted UAPs. This strategy targets ML models used by both xApps and rApps within the O-RAN system, aiming to degrade network performance. We experimentally validate the effectiveness of the designed evasion attack strategy and quantify the scale of performance degradation using a real-world O-RAN testbed and emulation environments. This evaluation is conducted using the Interference Classification xApp and the Power Saving rApp as representative applications for near-RT and non-RT RICs, respectively. Further, we show that the attack strategy is effective against prominent defense techniques for adversarial ML, such as defensive distillation and adversarial training.


  • Open Access

    Rethinking the Role of Network Stacks for Website Fingerprinting Defenses

    by Elisaveta Lavrentieva, Marc Juarez, Michio Honda

  • Shift Your Shape: Correlating and Defending Mixnet Flows Based on Their Shapes

    by Lennart Oldenburg, Marc Juarez, Enrique Argones Rúa, Claudia Diaz

    When the packet rate of flows in a mixnet depends on the amount of transferred data, it is possible to identify which flow entering is which flow exiting the mixnet based on their shapes. We present a passive shape-based flow correlation attack against state-of-the-art mixnet Nym and a systematic evaluation of countermeasures. Assuming an adversary controlling both the entry and exit gateway-requesters selected by users to access the public Internet through Nym, our attack's artificial neural network assigns correlation scores to flow pairs based on traffic distribution similarities to accurately distinguish paired from unpaired flow tuples. From data we collected on the live Nym mixnet, we generate $ \mathbf {45}$ datasets and $ \mathbf {119}$ testing scenarios for different defense configurations. After one minute of attacking flow pairs on default Nym, we achieve a PR-AUC of $ \mathbf {0.9998}$ at a base rate of $ \mathbf {1.9 \times 10^{-4}}$ paired flow tuples. However, (combinations of) the five evaluated defense strategies indicate that the right choice and scale of countermeasure(s) can offer meaningful protection. Our evaluation also informs on the resources overhead spent on defenses. We discuss steps a mixnet such as Nym can take to make our attack both less likely and less accurate.

  • Open Access

    SoK: What Makes Private Learning Unfair?

    by Kai Yao, Marc Juarez

    Differential privacy has emerged as the most studied framework for privacy-preserving machine learning. However, recent studies show that enforcing differential privacy guarantees can not only significantly degrade the utility of the model, but also amplify existing disparities in its predictive performance across demographic groups. Although there is extensive research on the identification of factors that contribute to this phenomenon, we still lack a complete understanding of the mechanisms through which differential privacy exacerbates disparities. The literature on this problem is muddled by varying definitions of fairness, differential privacy mechanisms, and inconsistent experimental settings, often leading to seemingly contradictory results. This survey provides the first comprehensive overview of the factors that contribute to the disparate effect of training models with differential privacy guarantees. We discuss their impact and analyze their causal role in such a disparate effect. Our analysis is guided by a taxonomy that categorizes these factors by their position within the machine learning pipeline, allowing us to draw conclusions about their interaction and the feasibility of potential mitigation strategies. We find that factors related to the training dataset and the underlying distribution play a decisive role in the occurrence of disparate impact, highlighting the need for research on these factors to address the issue.

  • Towards Understanding the Interplay of Generative Artificial Intelligence and the Internet

    by Gonzalo Martínez, Lauren Watson, Pedro Reviriego, Hernández, José Alberto, Marc Juarez, Rik Sarkar


  • Open Access

    MixMatch: Flow Matching for Mixnet Traffic

    by Lennart Oldenburg, Marc Juarez, Enrique Argones Rúa, Claudia Diaz

    Mixnets provide communication anonymity against network adversaries by routing packets independently via multiple hops, delaying them artificially at each hop, and introducing cover traffic. We show that these features (particularly the use of cover traffic) significantly diminish the effectiveness of state-of-the-art flow correlation techniques developed to link the two ends of a Tor connection. In this work, we propose novel methods to determine whether a set of endpoints exchanges packets via a mixnet and demonstrate their effectiveness by applying them to the Nym mixnet. We consider Nym in both an idealized lab setup and the official live network, and propose and compare three classifiers to conduct flow matching on it. Our statistical classifier tests whether egress packet timestamps are consistent with ingress timestamps and the (known) routing delay characteristic of the mixnet. In contrast, our two deep learning (DL) classifiers learn to distinguish matched from unmatched flow pairs from collected datasets directly, rather than relying on priors that describe the delay distribution. All three classifiers use our flow merging technique, which enables testing a match for sets of communicating endpoints of any cardinality. Considering a use case where two observed endpoints communicate exclusively to exchange a file through Nym, we find that flow matching is fast and accurate in the idealized lab setup. If flow pairs are aligned using all network observations in a download, we achieve a TPR of circa 0.6 (DL) and 0.47 (statistical) at an FPR of 10-2 after only processing 100 observations. We evaluate classifier performance under key variations of this setup: the absence of loop cover traffic, an increased or decreased average per-mix delay, larger communicating sets (three endpoints) with faster responders, and the presence of realistic network effects (live network). The classifiers' matching performance diminishes on the live network where packet losses and variable propagation delays exist, reducing DL TPR to circa 0.26 and statistical TPR to circa 0.28 at an FPR of 10-2. Informed by the insights of our analyses, we outline countermeasures that can be deployed in mixnets such as Nym to mitigate flow matching threats.


  • Open Access

    Online Platforms and the Fair Exposure Problem Under Homophily

    by Jakob Schoeffer, Alexander Ritchie, Keziah Naggita, Faidra Monachou, Jessie Finocchiaro, Marc Juarez

    In the wake of increasing political extremism, online platforms have been criticized for contributing to polarization. One line of criticism has focused on echo chambers and the recommended content served to users by these platforms. In this work, we introduce the fair exposure problem: given limited intervention power of the platform, the goal is to enforce balance in the spread of content (e.g., news articles) among two groups of users through constraints similar to those imposed by the Fairness Doctrine in the United States in the past. Groups are characterized by different affiliations (e.g., political views) and have different preferences for content. We develop a stylized framework that models intra- and intergroup content propagation under homophily, and we formulate the platform's decision as an optimization problem that aims at maximizing user engagement, potentially under fairness constraints. Our main notion of fairness requires that each group see a mixture of their preferred and non-preferred content, encouraging information diversity. Promoting such information diversity is often viewed as desirable and a potential means for breaking out of harmful echo chambers. We study the solutions to both the fairness-agnostic and fairness-aware problems. We prove that a fairness-agnostic approach inevitably leads to group-homogeneous targeting by the platform. This is only partially mitigated by imposing fairness constraints: we show that there exist optimal fairness-aware solutions which target one group with different types of content and the other group with only one type that is not necessarily the group's most preferred. Finally, using simulations with real-world data, we study the system dynamics and quantify the price of fairness.