Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Adversarial examples”

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.

At least 37 records · Page 2

An Example of Counter-Adversarial Community Detection Analysis

Community detection is often used to understand the nature of a network. However, there may exist an adversarial member of the network who wishes to evade that understanding. We analyze one such specific situation, quantifying the efficacy of certain attacks against a particular analytic use of community detection and providing a preliminary assessment of a possible defense.

97 MATHEMATICS AND COMPUTING↗

Model-based Hierarchical Reinforcement Learning for Improved Physical Security Design: A Prototype

Prior work in FY24 developed an adversarial AI agent aid in path analysis of physical protection systems. This agent, trained using a model-based reinforcement learning algorithm, was able to successfully learn the most vulnerable path in facilities. It was able to extend the current state of practice for physical protection design by exhibiting dynamic behavior based on current environmental conditions. Whereas PathTrace largely performs a static, graph-based analysis, the AI agent was able to make decisions based on relative position in the facility, current conditions (was the adversarial agnet discovered?), and proximity to secondary targets. The agent demonstrated some novel capabilities, but had limitations that need to be resolved before it can be used for production purposes. For example, the adversarial agent generalizes poorly and takes a relatively long time to train. Nonetheless, there is still considerable promise for developing the adversarial agent further in order to explore even richer, more dynamic behaviors (e.g., adversary motivations, environmental debris, and more). This work considers a complementary idea; development of a planning agent. The planning agent is envisioned as an auto-complete-like tool that can help accelerate security system design by human experts. The agent would respect existing barriers and sensors placed by a human expert while offering cost-effective suggestions (i.e., implicitly balancing effectiveness with cost) to improve the design. The goal is for this agent to be part of an expert’s toolbox, not to totally upend the current state-of-practice, or to displace human experts. The ultimate goal would be concurrent training of both the adversarial and planning agent together, to learn entirely through self-play. This would represent an entirely new way of performing system deign. We selected a hierarchical, model-based reinforcement learning algorithm to serve as the planning agent. This is an extension of concepts used in the prior FY24 adversarial agent work. There, we had a single agent acting an environment. Here, we have two different sub-agents (policies), working together, to form a complete agent. There is a manager policy, which can select abstract goals on slower time scales, and a worker, which performs primitive actions to reach goals selected by the manager. It is worth noting that this class of algorithm is challenging to work with. From our understanding, our work is one of the first successful uses of model-based reinforcement learning (MBRL) in nuclear energy1 , and likely the first hierarchical model-based reinforcement learning application in nuclear energy. Further, this work is one of the first known attempts to apply AI to perform a design tasks in nuclear energy. Consequently, there were significant implementation challenges and the bulk of the work was focused on successful implementation and algorithm design. The results presented here are very low technology readiness level as a consequence of the lack of related literature, but still represent a significant step forward in the pursuit of applied AI for design.

42 ENGINEERING↗

A comparison of approximate reasoning results using information uncertainty

An Approximate Reasoning (AR) model is a useful alternative to a probabilistic model when there is a need to draw conclusions from information that is qualitative. For certain systems, much of the information available is elicited from subject matter experts (SME). One such example is the risk of attack on a particular facility by a pernicious adversary. In this example there are several avenues of attack, i.e. scenarios, and AR can be used to model the risk of attack associated with each scenario. The qualitative information available and provided by the SME is comprised of linguistic values which are well suited for an AR model but meager for other modeling approaches. AR models can produce many competing results. Associated with each competing AR result is a vector of linguistic values and a respective degree of membership in each value. A suitable means to compare and segregate AR results would be an invaluable tool to analysts and decisions makers. A viable method would be to quantify the information uncertainty present in each AR result then use the measured quantity comparatively. One issue of concern for measuring the infornlation uncertainty involved with fuzzy uncertainty is that previously proposed approaches focus on the information uncertainty involved within the entire fuzzy set. This paper proposes extending measures of information uncertainty to AR results, which involve only one degree of membership for each fuzzy set included in the AR result. An approach to quantify the information uncertainty in the AR result is presented.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Bennett-Brassard 1984 quantum key distribution using conjugate homodyne detection

Optical homodyne detection has been widely used in continuous-variable (CV) quantum information processing for measuring field quadrature. In this paper we explore the possibility of operating a conjugate homodyne detection system in “photon counting” mode to implement discrete-variable (DV) quantum key distribution (QKD). A conjugate homodyne detection system, which consists of a beam splitter followed by two optical homodyne detectors, can simultaneously measure a pair of conjugate quadratures X and P of the incoming quantum state. In classical electrodynamics, X 2 + P 2 is proportional to the energy (the photon number) of the input light. In quantum optics, X and P do not commute and thus the above photon-number measurement is intrinsically noisy. This implies that a blind application of standard security proofs of QKD could result in pessimistic performance. We overcome this obstacle by taking advantage of two special features of the proposed detection scheme. First, the fundamental detection noise associated with vacuum fluctuations cannot be manipulated by an external adversary. Second, the ability to reconstruct the photon number distribution at the receiver's end can place additional constraints on possible attacks from the adversary. As an example, we study the security of the BB84 QKD using conjugate homodyne detection and evaluate its performance through numerical simulations. This study may open the door to a family of QKD protocols, complementary to the well-established DV-QKD based on single-photon detection and CV-QKD based on coherent detection.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Achieving the physical limits of the bounded-storage model

Secure two-party cryptography is possible if the adversary's quantum storage device suffers imperfections. For example, security can be achieved if the adversary can store strictly less then half of the qubits transmitted during the protocol. This special case is known as the bounded-storage model, and it has long been an open question whether security can still be achieved if the adversary's storage were any larger. Here, we answer this question positively and demonstrate a two-party protocol which is secure as long as the adversary cannot store even a small fraction of the transmitted pulses. We also show that security can be extended to a larger class of noisy quantum memories.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Adversarial Sampling-Based Motion Planning

In this report there are many scenarios in which a mobile agent may not want its path to be predictable. Examples include preserving privacy or confusing an adversary. However, this desire for deception can conflict with the need for a low path cost. Optimal plans such as those produced by RRT* may have low path cost, but their optimality makes them predictable. Similarly, a deceptive path that features numerous zig-zags may take too long to reach the goal. We address this trade-off by drawing inspiration from adversarial machine learning. We propose a new planning algorithm, which we title Adversarial RRT*. Adversarial RRT* attempts to deceive machine learning classifiers by incorporating a predicted measure of deception into the planner cost function. Adversarial RRT* considers both path cost and a measure of predicted deceptiveness in order to produce a trajectory with low path cost that still has deceptive properties. We demonstrate the performance of Adversarial RRT*, with two measures of deception, using a simulated Dubins vehicle. We show how Adversarial RRT* can decrease cumulative RNN accuracy across paths to 10%, compared to 46% cumulative accuracy on near-optimal RRT* paths, while keeping path length within 16% of optimal. We also present an example demonstration where the Adversarial RRT* planner attempts to safely deliver a high value package while an adversary observes the path and tries to intercept the package.

42 ENGINEERING↗

Emerging Threats in Transportation Security Related to Intelligent Transportation Systems (ITS)

Transport of high-consequence shipments requires a resilient and robust systems of systems to guarantee cargo arrival. Furthermore, rising adoption of technologies such as connected and automated vehicles (CAVs), intelligent infrastructure, and vehicle-to-everything (V2X) communication presents unique challenges for securing transportation systems. Within these Intelligent Transportation Systems (ITS), several additional vulnerabilities exist that create pathways for adversarial attacks and cargo interception. For example, connectivity provides cyber pathways directly into vehicle systems and infrastructure for malicious actors. Furthermore, advanced vehicle automation exposes additional vehicle control necessary for shipment interception otherwise unavailable to adversaries. Within this paper, we will discuss the specific threats introduced by ITS-enabled technologies currently deployed and in development. These include those mentioned related to connectivity and automation, but will be expanded into grid, infrastructure, and vehicle specific threats. In addition, we will discuss how to potentially mitigate these emerging challenges as well as how to safeguard transportation systems from next generation attacks.

Cook, Adian [ORNL] (ORCID:0000000160825395)↗

Radio Frequency Spectrum Audit to Inventory Private Cellular Base Station Infrastructure

The ever-changing cellular communication landscape makes it difficult to identify, map, and localize cellular base stations. Localizing cellular base stations provides various advantages, including information security, cybersecurity, spectrum management, and interference detection. For example, the MITRE ATT&CK® (Adversarial Tactics, Techniques, and Common Knowledge architecture) [1] and Common Attack Pattern Enumeration and Classification [2] emphasize the importance of being able to minimize the cyber security threat presented by unregulated private cellular base stations (PCBS). The majority of published research looks at the malicious use of PCBSs and focuses on using data retrieved from user equipment (UE), data obtained from an application on the UE, or data shared between the UE and a mobile network to locate it. This innovative strategy, however, focuses on the passively discovered uniqueness of radio frequency (RF) transmissions from commercial cellular infrastructure received in a designated monitoring position (DMP).

42 ENGINEERING↗

GEDet: Detecting Erroneous Nodes with A Few Examples

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.

Guan, Sheng↗

Statistical Hypothesis Testing using CNN Features for Synthesis of Adversarial Counterexamples to Human and Object Detection Vision Systems

Validating the correctness of human detection vision systems is crucial for safety applications such as pedestrian collision avoidance in autonomous vehicles. The enormous space of possible inputs to such an intelligent system makes it difficult to design test cases for such systems. In this report, we present our tool MAYA that uses an error model derived from a convolutional neural network (CNN) to explore the space of images similar to a given input image, and then tests the correctness of a given human or object detection system on such perturbed images. We demonstrate the capability of our tool on the pre-trained Histogram-of-Oriented-Gradients (HOG) human detection algorithm implemented in the popular OpenCV toolset and the Caffe object detection system pre-trained on the ImageNet benchmark. Our tool may serve as a testing resource for the designers of intelligent human and object detection systems.

97 MATHEMATICS AND COMPUTING↗

The Role of Nuclear-Conventional Intermingling on State Decision-making and the Risk of Inadvertent Escalation

The questions answered by this report are: What are the implications of nuclear and conventional intermingling on crisis stability and the potential risk of miscalculation?, and Specifically, how might entanglement influence US and competitor decision making during crisis and conflict? In practice, there are three main forms of nuclear-conventional intermingling. First, intermingling between nuclear and non-nuclear weapon systems can occur via the fielding of dual capable delivery systems like missiles or aircraft. Second, intermingling can happen due to the co-location of nuclear and non-nuclear forces and their support structures—for example, the co-location of strategic bombers and general-purpose aircraft, or the co-location of strategic submarines and general-purpose vessels. Third, intermingling can occur via convolving nuclear and conventional military command and control systems, to include ballistic missile early warning and potentially space surveillance systems as well. All three forms of nuclear-conventional intermingling have significantly increased since the end of the Cold War, driven by both technological and doctrinal changes. However, there are important differences in the rationale behind, and also the risks associated with these three different forms of intermingling. The mere existence of dual capable systems is not new— deploying such systems can increase the effectiveness of forces, and it can also provide more flexibility. The major powers have both employed and threatened with dual capable systems for decades, and they have done so without nuclear escalation. Similarly, the major powers co-located nuclear and conventional systems in the Cold War, and they did so for variety of reasons that had nothing to do with complicating the adversary’s risk calculus. The Soviet Union, for example, decided to co-locate its nuclear and conventional forces for economic and administrative reasons. Although today it might be recognized as a useful deterrent tool, it was not their primary intention. Major powers want to convince rivals that the co-location of forces creates a high bar for targeting and raises the risk of nuclear escalation, but they also want to have the flexibility of this not being true in an actual crisis or conflict.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Quantum-assisted associative adversarial network: applying quantum annealing in deep learning

Abstract Generative models have the capacity to model and generate new examples from a dataset and have an increasingly diverse set of applications driven by commercial and academic interest. In this work, we present an algorithm for learning a latent variable generative model via generative adversarial learning where the canonical uniform noise input is replaced by samples from a graphical model. This graphical model is learned by a Boltzmann machine which learns low-dimensional feature representation of data extracted by the discriminator. A quantum processor can be used to sample from the model to train the Boltzmann machine. This novel hybrid quantum-classical algorithm joins a growing family of algorithms that use a quantum processor sampling subroutine in deep learning, and provides a scalable framework to test the advantages of quantum-assisted learning. For the latent space model, fully connected, symmetric bipartite and Chimera graph topologies are compared on a reduced stochastically binarized MNIST dataset, for both classical and quantum sampling methods. The quantum-assisted associative adversarial network successfully learns a generative model of the MNIST dataset for all topologies. Evaluated using the Fréchet inception distance and inception score, the quantum and classical versions of the algorithm are found to have equivalent performance for learning an implicit generative model of the MNIST dataset. Classical sampling is used to demonstrate the algorithm on the LSUN bedrooms dataset, indicating scalability to larger and color datasets. Though the quantum processor used here is a quantum annealer, the algorithm is general enough such that any quantum processor, such as gate model quantum computers, may be substituted as a sampler.

Wilson, Max (ORCID:0000000207983391)↗

Optimization-Based Resiliency Verification in Microgrids via Maximal Adversarial Set Characterization

Critical energy infrastructures are increasingly relying on advanced sensing and control technologies for efficient and optimal utilization of flexible energy resources. Algorithmic procedures are needed to ensure that such systems are designed to be resilient to a wide range of cyber-physical adversarial events. This paper provides a robust optimization framework to quantify the largest adversarial perturbation that a system can accommodate without violating pre-specified resiliency metrics. We formulate the maximal adversarial set characterization as a bi-level optimization problem which is solved via Lagrangian relaxations.We illustrate the proposed algorithm on an islanded microgrid example: a modified IEEE 123-node feeder with distributed energy resources. Simulations are carried out to characterize the tolerable adversarial perturbations for varying levels of available flexibility (energy reserves).

Nazir, Mohammad Nawaf↗

Final Report for LDRD Project 02-FS-009 Gigapixel Surveillance Camera

The threats of terrorism and proliferation of weapons of mass destruction add urgency to the development of new techniques for surveillance and intelligence collection. For example, the United States faces a serious and growing threat from adversaries who locate key facilities underground, hide them within other facilities, or otherwise conceal their location and function. Reconnaissance photographs are one of the most important tools for uncovering the capabilities of adversaries. However, current imaging technology provides only infrequent static images of a large area, or occasional video of a small area. We are attempting to add a new dimension to reconnaissance by introducing a capability for large area video surveillance. This capability would enable tracking of all vehicle movements within a very large area. The goal of our project is the development of a gigapixel video surveillance camera for high altitude aircraft or balloon platforms. From very high altitude platforms (20-40 km altitude) it would be possible to track every moving vehicle within an area of roughly 100 km x 100 km, about the size of the San Francisco Bay region, with a gigapixel camera. Reliable tracking of vehicles requires a ground sampling distance (GSD) of 0.5 to 1 m and a framing rate of approximately two frames per second (fps). For a 100 km x 100 km area the corresponding pixel count is 10 gigapixels for a 1-m GSD and 40 gigapixels for a 0.5-m GSD. This is an order of magnitude beyond the 1 gigapixel camera envisioned in our LDRD proposal. We have determined that an instrument of this capacity is feasible.

42 ENGINEERING↗

Decision-Making in Complex Dynamical Systems of Systems With One Opposing Subsystem

Many complex dynamical systems consist of a large number of interacting subsystems that operate harmoniously and make decisions that are designed for the benefit of the entire enterprise. If, in an attempt to disrupt the operation of the entire system, one subsystem gets attacked and is made to operate in a manner that is adversarial with the others, then the entire system suffers, resulting in an adversarial decision-making environment among its subsystems. Such an environment may affect not only the decision-making process of the attacked subsystem but also possibly the other remaining subsystems as well. The disruption caused by the attacked subsystem may cause the remaining subsystems to either coalesce as a unified team making team-based decisions, or disintegrate and act as independent decision-making entities. The decision-making process in these types of complex systems of systems is best analyzed within the general framework of cooperative and non-cooperative game theory. In this paper, we will develop an analysis that provides a theoretical basis for modeling the decision-making process in such complex systems. We show how cooperation among the subsystems can produce Noninferior Nash Strategies (NNS) that are fair and acceptable to all subsystems within the team while at the same time provide the subsystems in the team with the security of the Nash equilibrium against the opposing attacked subsystem. We contrast these strategies with the all Nash Strategies (NS) that would result if the operation of the entire system disintegrated and became adversarial among all subsystems as a result of the attack. An example of a system consisting of three subsystems with one opposing subsystem as a result of an attack is included to illustrate the results.

Talebi, Shahriar↗

Optimization of detection sensitivity for a Fiber Optic Intrusion Detection System (FOIDS) using design of experiments.

The Fiber Optic Intrusion Detection System (FOIDS)1 is a physical security sensor deployed on fence lines to detect climb or cut intrusions by adversaries. Calibration of detection sensitivity can be time consuming because, for example, the FiberSenSys FD-332 has 32 settings that can be adjusted independently to provide a balance between a high probability of detection and a low nuisance alarm rate. Therefore, an efficient method of calibrating the FOIDS in the field, other than by trial and error, was needed. This study was conducted to: x Identify the most significant settings for controlling detection x Develop a way of predicting detection sensitivity for given settings x Develop a set of optimal settings for validation The Design of Experiments (DoE) 2-4 methodology was used to generate small, planned test matrixes, which could be statistically analyzed to yield more information from the test data. Design of Experiments is a statistical methodology for quickly optimizing performance of systems with measurable input and output variables. DoE was used to design custom screening experiments based on 11 FOIDS settings believed to have the most affect on WKH types of fence perimeter intrusions were evaluated: simulated cut intrusions and actual climb intrusions. Two slightly different two-level randomized fractional factorial designed experiment matrixes consisting of 16 unique experiments were performed in the field for each type of intrusion. Three repetitions were conducted for every cut test; two repetitions were conducted for every climb test. Total number of cut tests analyzed was 51; the total number of climb tests was 38. This paper discusses the results and benefits of using Design of Experiments (DoE) to calibrate and optimize the settings for a FOIDS sensor

22 GENERAL STUDIES OF NUCLEAR REACTORS↗