SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Topological Data Analysis for Adversarial Detection (LANL O4937) - Detects adversarial examples in vision-language models using persistent homology and two-sample testing. Combines TDA features from CLIP embeddings with statistical methods (ME, SCF, SAMMD, C2ST) for robust detection across ImageNet, CIFAR-10/100.
The well-known vulnerability of Deep Neural Networks to adversarial samples has led to a rapid cycle of increasingly sophisticated attack algorithms and proposed defenses. While most contemporary defenses have been shown to be vulnerable to carefully configured attacks, methods based on gradient regularization and out-of-distribution detection have attracted much interest recently by demonstrating higher resilience to a broad range of attack algorithms. However, no study has yet investigated the effect of combining these techniques. In this paper, we consider the effect of Jacobian matrix regularization on the detectability of adversarial samples on the CIFAR-10 image benchmark dataset. We find that regularization has a significant effect on detectability, and in some cases can make an undetectable attack on a baseline model detectable. In addition, we give evidence that regularization may mitigate the known weaknesses of detectors to high-confidence adversarial samples. The defenses we consider here are highly generalizable, and we believe they will be useful for further investigations to transfer machine learning robustness to other data domains.
Recent year has brought considerable advancements in Electric Vehicles (EVs) and associated infrastructures/ communications. Intrusion Detection Systems (IDS) are widely deployed for anomaly detection in such critical infrastructures. This paper presents an Interpretable Anomaly Detection System (RX-ADS) for intrusion detection in CAN protocol communication in EVs. Contributions include: 1) window based feature extraction method; 2) deep Autoencoder based anomaly detection method; and 3) adversarial machine learning based explanation generation methodology. The presented approach was tested on two benchmark CAN datasets: OTIDS and Car Hacking. The anomaly detection performance of RX-ADS was compared against the state-of-the-art approaches on these datasets: HIDS and GIDS. The RX-ADS approach presented performance comparable to the HIDS approach (OTIDS dataset) and has outperformed HIDS and GIDS approaches (Car Hacking dataset). Further, the proposed approach was able to generate explanations for detected abnormal behaviors arising from various intrusions. Furthermore, these explanations were later validated by information used by domain experts to detect anomalies. Other advantages of RX-ADS include: 1) the method can be trained on unlabeled data; 2) explanations help experts in understanding anomalies and root course analysis, and also help with AI model debugging and diagnostics, ultimately improving user trust in AI systems.
Machine learning for cyber defense offers the promise of detecting adversarial activity against the ground data systems managing critical space assets. A fundamental challenge facing machine learning research in cybersecurity is the lack of high-fidelity, shareable datasets for robust evaluation and testing of machine learning-based solutions. High-fidelity, real-world datasets are necessary for reliable benchmarking of nominal system behavior and malicious activity. Unfortunately, such realistic datasets of both nominal and adversarial activity are rarely shared publicly by data owners due to security and privacy concerns. Besides, the available adversarial data is sparse, which makes training models on malicious activity much harder. This situation has impeded and continues to impede the research and successful adoption of machine learning methods for cyber defense. Researchers have dealt with this problem by generating data within a low-fidelity lab environment, using classified and thus unshareable datasets, or downloading low-fidelity public datasets made available by others. We propose an innovative solution to the problem by employing machine learning methods to generate high-fidelity data. Specifically, we propose the use of Generative Adversarial Networks (GANs) to generate high-fidelity data for cybersecurity purposes. GANs have found successful image processing and natural language applications, but have not yet been investigated for cyber data generation. Our proposed approach first involves training the `discriminator' network of the GAN with a sample of real-world data consisting of malicious and nominal samples. We then use the `generator' network to generate new high-fidelity data samples consisting of an appropriate mix of malicious and nominal activity. We demonstrate applications of our architecture by generating high-fidelity cybersecurity data containing both malicious and nominal samples. We thoroughly evaluate the fidelity of our generated data using heuristics and evaluate its usefulness for machine learning applications using three different datasets. Overall, our approach results in high-fidelity, shareable datasets.
Detecting nodes with erroneous values in real-world graphs re- mains challenging due to the lack of examples and various error scenarios. We demonstrate GEDet, an error detection engine that can detect erroneous nodes in graphs with a few examples. The GEDet framework tackles error detection as a few-shot node classification problem. We invite the attendees to experience the following unique features. (1) Few-shot detection. Users only need to provide a few examples of erroneous nodes to perform error detection with GEDet. GEDet achieves desirable accuracy with (a) a graph augmentation module, which automatically generates synthetic examples to learn the classifier, and (b) an adversarial detection module, which improves classifiers to better distinguish erroneous nodes from both cleaned nodes and synthetic examples. We show that GEDet significantly improves the state-of-the-art error detection methods. (2) Diverse error scenarios. GEDet profiles data errors with a built-in library of transformation functions from correct values to errors. Users can also easily “plug in” new error types or examples. (3) User-centric detection. GEDet supports (a) an active learning mode to engage users to verify detected results, and adapts the error detection process accordingly; and (b) visual interfaces to interpret and track detected errors.
When searching for adversarial activity within multiple networks, one of the greatest challenges is how to accurately align entities across different channels of information. This task becomes increasingly difficult when minimal additional information is known about each individual besides a name. Within this study, we analyze name rarity and how it can be used to align people on three distinct data channels: Venmo financial transactions, Reddit online discussions, and a bibliographic data source of academic writings. We explore how the uniqueness of a name can be used to decide if a person is likely the same as another across networks, in the absence of any additional ground truth. While 100 percent confidence cannot be gained, we can use this information to clarify when a possible alignment is more or less likely to be the same individual, increasing our confidence of accurately detecting adversarial behavioral patterns. From the data collected, we found that 0.1% of people had the same name across data sets, and 22.5% of those names are considered rare by our threshold. In our study, we also examine the accuracy of our method and show how real names can be extracted from account usernames, and compared in a similar manner.
The rapid digitalization of the electric grid - driven by the integration of inverter-based resources (IBRs), battery energy storage systems (BESS), and advanced grid control platforms - has significantly enhanced grid efficiency, visibility, and flexibility. However, this evolution also introduces new cybersecurity risks, particularly through supply chain dependencies and operational blind spots at the grid edge. To address these challenges, Idaho National Laboratory (INL), through the Department of Energy (DOE) Office of Cybersecurity, Energy Security, and Emergency Response (CESER) Rapid Risk initiative, conducted a series of rapid risk assessment engagements with energy organizations across the United States. Drawing on lessons learned from these engagements, INL developed the following threat hunting guide for asset owners and operators (AOOs) to enhance their cybersecurity visibility within BESS and IBR systems. The guide demonstrates how to use passive network monitoring to baseline device behavior, detect adversarial activity, and investigate anomalies without disrupting operations. By implementing these practices, energy sector stakeholders can improve coordination between cybersecurity and operations teams and strengthen the resilience of distributed energy resources (DERs) within the modern power grid. Prior to implementing any network monitoring, packet capture, or threat hunting activity described in this guide, AOOs are strongly advised to review applicable governance frameworks, legal requirements, and organizational policies. This guide is intended for informational and educational purposes only. It does not replace compliance with any federal, state, or local cybersecurity mandates or industry standards. Implementation of described configurations, technologies, or analytic workflows is performed at the discretion and responsibility of the asset owner and operator.
Adversarial binaries are executable files that have been altered without loss of function by an AI agent in order to deceive malware detection systems. Progress in this emergent vein of research has been constrained by the complex and rigid structure of executable files. Although prior work has demonstrated that these binaries deceive a variety of malware classification models which rely on disparate feature sets, a consensus as to the best approach has not been reached, either in terms of the optimization algorithms or the instrumentation methods. Furthermore, although inconsistencies in the data sets, target classifiers, and functionality verification methods make head-to-head comparisons difficult, here we extract lessons learned and make recommendations for future research.
The U.S. Department of Energy’s (DOE) Office of Cybersecurity, Energy Security, and Emergency Response (CESER), through the Cybersecurity for the Operational Technology Environment (CyOTE) Program, worked with energy sector asset owners and operators (AOOs), partners, and Idaho National Laboratory (INL) to develop capabilities for AOOs to independently detect adversarial tactics, techniques, and procedures (TTPs) within their operational technology (OT) environments. Unlike the approach taken with commercial security solutions, CyOTE seeks to tie anomalies in cyber operations to a cyber-attack. By stringing together multiple techniques in the OT environment, AOOs can identify attack campaigns with ever decreasing impacts. The CyOTE methodology applies fundamental concepts of perception and comprehension to a universe of knowns and unknowns increasingly disaggregated into observables, anomalies, and triggering events. MITRE’s ATT&CK® Framework for Industrial Control Systems (ICS) is used as a common lexicon to identify a set of triggering events related to three Use Cases – Alarm Logs, Human-Machine Interface (HMI), and Remote Logins – which together account for 87 percent of the techniques commonly used by adversaries. The CyOTE methodology is also appropriate for OT-related anomalies perceived outside the three Use Cases, such as through the energy system itself.
Accurate and speedy detection of power system events is critical to enhancing the reliability and resiliency of power systems. Although supervised deep learning algorithms show great promise in identifying power system events, they require a large volume of high-quality event labels for training. This paper develops a bidirectional anomaly generative adversarial network (GAN)-based algorithm to detect power system events using streaming PMU data, which does not rely on a huge amount of event labels. By introducing conditional entropy constraint in the objective function of GAN and graph signal processing-based PMU sorting technique, our proposed algorithm significantly outperforms state-of-the-art event detection algorithms in terms of accuracy. To facilitate the adoption of the proposed algorithm, a prototype online platform is also developed using Apache Hadoop, Kafka, and Spark to enable real-time event detection. Here, the accuracy and computational efficiency of the proposed algorithm are validated using a large-scale real-world PMU dataset from the Eastern Interconnection of the United States.
Not Available
In nuclear power plants (NPPs), anomalies arising from sensors or human errors (HEs) can undermine the performance and reliability of plant operations. Anomaly detection models can be employed to detect sensor errors and HEs. Additionally, physics-informed machine learning models can utilize the known physics of the system, as described by mathematical equations, to ensure that sensor values are consistent with physical laws. Hence, we propose SPIDARman: System-level Physics-Informed Detection of Anomalies in Reactor Collected Data Considering Human Errors, a holistic physics-informed anomaly detection approach based on generative adversarial networks (GANs) to detect anomalies in both automatically collected sensor data and manually collected surveillance data. Here we test our approach on data collected from a flow loop testbed, showcasing its potential to detect anomalies. Results demonstrate that the proposed model performs better than the baseline GAN-based models in detecting sensor and surveillance anomalies, suggesting the potential of physics-informed anomaly detection GAN models in NPPs.
Dragos provided INL access to their Neighborhood Keeper platform, which contains simulated data, sample reports given to utilities, and access to anonymized program data that is sent from utilities to the cloud. Dragos has specifically asked INL to review the available data and respond to the following questions: 1) What detections or combinations of detections are most useful? 2) What additional analysis would benefit the electricity subsector? 3) How could Neighborhood Keeper reporting be modified to benefit the electricity subsector? 4) What lessons learned from the Cybersecurity for the Operational Technology Environment (CyOTE) program could be provided? 5) Based on INL’s experience with utilities, are there lessons learned about data sharing that could be provided? 6) Are there recommendations that could benefit the electricity subsector?
Human error (HE) is an important concern in safety-critical systems such as nuclear power plants (NPPs). HE has played a role in many accidents and outage incidents in NPPs. Despite the increased automation in NPPs, HE remains unavoidable. Hence, the need for HE detection is as important as HE prevention efforts. In NPPs, HE is rather rare. Hence, anomaly detection, a widely used machine learning technique for detecting rare anomalous instances, can be repurposed to detect potential HE. In this study, we develop an unsupervised anomaly detection technique based on generative adversarial networks (GANs) to detect anomalies in manually collected surveillance data in NPPs. More specifically, our GAN is trained to detect mismatches between automatically recorded sensor data and manually collected surveillance data, and hence, identify anomalous instances that can be attributed to HE. We test our GAN on both a real-world dataset and an external dataset obtained from a testbed, and we benchmark our results against state-of-the-art unsupervised anomaly detection algorithms, including one-class support vector machine and isolation forest. Our results show that the proposed GAN provides improved anomaly detection performance. Our study is promising for the future development of artificial intelligence based HE detection systems.
Abstract not provided.
Explore the source record for details and available documents.
Abstract not provided.