This is a corrigendum for the article “Quantitative Measurement of Cyber Resilience: Modeling and Experimentation” published in ACM Trans. Cyber-Phys. Syst. 9, 1, Article 1 (January 2025), 25 pages.

Cyber resilience is the ability of a system to resist and recover from a cyber attack, thereby restoring the system’s functionality. Effective design and development of a cyber resilient system requires experimental methods and tools for quantitative measuring of cyber resilience. This article describes an experimental method and test bed for obtaining resilience-relevant data as a system (in our case—a truck) traverses its route, in repeatable, systematic experiments. We model a truck equipped with an autonomous cyber-defense system and which also includes inherent physical resilience features. When attacked by malware, this ensemble of cyber-physical features (i.e., “bonware”) strives to resist and recover from the performance degradation caused by the malware’s attack. We propose parsimonious mathematical models to aid in quantifying systems’ resilience to cyber attacks. Using the models, we identify quantitative characteristics obtainable from experimental data and show that these characteristics can serve as useful quantitative measures of cyber resilience.

Ultraviolet (UV) communications has been proposed as a promising modality for short-range military communications, as it is often presumed to have low-probability-of-detection characteristics, has desirable non-line-of-sight properties, and resides within an underutilized frequency band. Recent research efforts have sought to formalize the first presumption of the detection of UV communications. This effort seeks to begin the study of the localization of UV communication sources after they are detected. We focus here exclusively on the range estimation problem. Using a phenomenon relating UV received power and range over short-to-medium distances (≪ 1 km), we develop a range estimator using only the received signal strength. The approach does not require information about other system or environmental parameters. We also theoretically study the performance of the estimator using the Cramér-Rao bound, via simulations, and using previously collected data.

Machine learning-based network intrusion detection systems have demonstrated state-of-the-art accuracy in flagging malicious traffic. However, machine learning has been shown to be vulnerable to adversarial examples, particularly in domains such as image recognition. In many threat models, the adversary exploits the unconstrained nature of images–the adversary is free to select some arbitrary amount of pixels to perturb. However, it is not clear how these attacks translate to domains such as network intrusion detection as they contain domain constraints, which limit which and how features can be modified by the adversary. In this paper, we explore whether the constrained nature of networks offers additional robustness against adversarial examples versus the unconstrained nature of images. We do this by creating two algorithms: (1) the Adapative-JSMA, an augmented version of the popular JSMA which obeys domain constraints, and (2) the Histogram Sketch Generation which generates adversarial sketches: targeted universal perturbation vectors that encode feature saliency within the envelope of domain constraints. To assess how these algorithms perform, we evaluate them in a constrained network intrusion detection setting and an unconstrained image recognition setting. The results show that our approaches generate misclassification rates in network intrusion detection applications that were comparable to those of image recognition applications (greater than 95%). Our investigation shows that the constrained attack surface exposed by network intrusion detection systems is still sufficiently large to craft successful adversarial examples – and thus, network constraints do not appear to add robustness against adversarial examples. Indeed, even if a defender constrains an adversary to as little as five random features, generating adversarial examples is still possible.

Machine learning is vulnerable to adversarial examples--inputs designed to cause models to perform poorly. However, it is unclear if adversarial examples represent realistic inputs in the modeled domains. Diverse domains such as networks and phishing have domain constraints--complex relationships between features that an adversary must satisfy for an attack to be realized (in addition to any adversary-specific goals). In this paper, we explore how domain constraints limit adversarial capabilities and how adversaries can adapt their strategies to create realistic (constraint-compliant) examples. In this, we develop techniques to learn domain constraints from data, and show how the learned constraints can be integrated into the adversarial crafting process. We evaluate the efficacy of our approach in network intrusion and phishing datasets and find: (1) up to 82% of adversarial examples produced by state-of-the-art crafting algorithms violate domain constraints, (2) domain constraints are robust to adversarial examples; enforcing constraints yields an increase in model accuracy by up to 34%. We observe not only that adversaries must alter inputs to satisfy domain constraints, but that these constraints make the generation of valid adversarial examples far more challenging.

Abstract Source code authorship attribution can be used for many types of intelligence on binaries and executables, including forensics, but introduces a threat to the privacy of anonymous programmers. Previous work has shown how to attribute individually authored code files and code segments. In this work, we examine authorship segmentation, in which we determine authorship of arbitrary parts of a program. While previous work has performed segmentation at the textual level, we attempt to attribute subtrees of the abstract syntax tree (AST). We focus on two primary problems: identifying the primary author of an arbitrary AST subtree and identifying on which edges of the AST primary authorship changes. We demonstrate that the former is a difficult problem but the later is much easier. We also demonstrate methods by which we can leverage the easier problem to improve accuracy for the harder problem. We show that while identifying the author of subtrees is difficult overall, this is primarily due to the abundance of small subtrees: in the validation set we can attribute subtrees of at least 25 nodes with accuracy over 80% and at least 33 nodes with accuracy over 90%, while in the test set we can attribute subtrees of at least 33 nodes with accuracy of 70%. While our baseline accuracy for single AST nodes is 20.21% for the validation set and 35.66% for the test set, we present techniques by which we can increase this accuracy to 42.01% and 49.21% respectively. We further present observations about collaborative code found on GitHub that may drive further research.

Challenging Army-relevant communications and networking environments require novel methods and systems with sufficient robustness and resilience to enable operation in a covert fashion, especially in the presence of sophisticated adversaries. Recent studies suggest that regions of electromagnetic spectrum, such as deep ultraviolet (UV) spectrum, that are not being utilized in existing systems can provide unique advantages, particularly for more-covert short- range operations. However, the UV channel is not well understood, and techniques for the effective and efficient use of this channel for networking, such as spatial multiplexing, are still being investigated. In this paper, we report on experiments that validate previously hypothesized channel modeling behaviors, where we isolate and accurately measure the line-of-sight (LOS), single-scattering, and multiple-scattering components of measured returns and compare the measurements with theory. These experimental measurements exhibit good agreement with theoretical predictions, and demonstrate, in particular, predicted variations from ideal behavior in the multiple-scattering case. The validated behaviors play a key role in a previously proposed UV networking optimization framework. As such, we conclude this paper with additional development and examples of that framework in order to demonstrate the utility of the validated channel modeling.

Abstract Program authorship attribution has implications for the privacy of programmers who wish to contribute code anonymously. While previous work has shown that individually authored complete files can be attributed, these efforts have focused on such ideal data sets as contest submissions and student assignments. We explore the problem of authorship attribution “in the wild,” examining source code obtained from open-source version control systems, and investigate how contributions can be attributed to their authors, either on an individual or a per-account basis. In this work, we present a study of attribution of code collected from collaborative environments and identify factors which make attribution of code fragments more or less successful. For individual contributions, we show that previous methods (adapted to be applied to short code fragments) yield an accuracy of approximately 50% or 60%, depending on whether we average by sample or by author, at identifying the correct author out of a set of 104 programmers. By ensembling the classification probabilities of a sufficiently large set of samples belonging to the same author we achieve much higher accuracy for assigning the set of samples to the correct author from a known suspect set. Additionally, we propose the use of calibration curves to identify which samples are by unknown and previously unencountered authors.

We present a novel method for reconstructing three dimensional surfaces from stereo intensity data. We employ a set of competing surface hypotheses based on parameterized models. We use maximum a posteriori (MAP) estimation and demonstrate a connection to the Hough transform. Experimental results are given showing the effectiveness of the algorithm.