Search NASA⌕ Search

SEARCH · Search NASA

Results for “adversarial robustness”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Gumby: Quantifying multi-modal model resiliency

With the rise of cheap data and sensors, more use cases are emerging for multi-input models. Research has shown that including multiple data modalities can improve performance, suggesting that deep learning models can successfully learn to leverage complementary information from different modalities. However, this improved predictive power comes with unanticipated costs: additional inputs change model resiliency and expand the threat space for adversarial attacks. We first provide theoretical underpinnings for how adversarial success scales with input dimension. We then characterize the performance of a suite of multispectral deep learning models with different fusion approaches, quantify their relative reliance on different input bands, and evaluate their robustness to naturalistic and adversarial image corruptions.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

Graph interpolating activation improves both natural and robust accuracies in data-efficient deep learning

Improving the accuracy and robustness of deep neural nets (DNNs) and adapting them to small training data are primary tasks in deep learning (DL) research. In this paper, we replace the output activation function of DNNs, typically the data-agnostic softmax function, with a graph Laplacian-based high-dimensional interpolating function which, in the continuum limit, converges to the solution of a Laplace–Beltrami equation on a high-dimensional manifold. Furthermore, we propose end-to-end training and testing algorithms for this new architecture. The proposed DNN with graph interpolating activation integrates the advantages of both deep learning and manifold learning. Compared to the conventional DNNs with the softmax function as output activation, the new framework demonstrates the following major advantages: First, it is better applicable to data-efficient learning in which we train high capacity DNNs without using a large number of training data. Second, it remarkably improves both natural accuracy on the clean images and robust accuracy on the adversarial images crafted by both white-box and black-box adversarial attacks. Third, it is a natural choice for semi-supervised learning. This paper is a significant extension of our earlier work published in NeurIPS, 2018. For reproducibility, the code is available at https://github.com/BaoWangMath/DNN-DataDependentActivation .

Mathematics↗
DOE OSTI2020

On the vulnerability of data-driven structural health monitoring models to adversarial attack

Many approaches at the forefront of structural health monitoring rely on cutting-edge techniques from the field of machine learning. Recently, much interest has been directed towards the study of so-called adversarial examples; deliberate input perturbations that deceive machine learning models while remaining semantically identical. This article demonstrates that data-driven approaches to structural health monitoring are vulnerable to attacks of this kind. In the perfect information or ‘white-box’ scenario, a transformation is found that maps every example in the Los Alamos National Laboratory three-storey structure dataset to an adversarial example. Also presented is an adversarial threat model specific to structural health monitoring. The threat model is proposed with a view to motivate discussion into ways in which structural health monitoring approaches might be made more robust to the threat of adversarial attack.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2020

Probing for Artifacts: Detecting Imagenet Model Evasions

While deep learning models have made incredible progress across a variety of machine learning tasks, they remain vulnerable to adversarial examples crafted to fool otherwise trustworthy models. In this work we approach this problem through the lens of a detection framework. We propose a classification network that uses the hidden layer activations of a trained model as inputs to detect adversarial artifacts in an input. We train this classification network simultaneously against multiple adversarial algorithms to create a more robust detector and show higher detection rates than several alternatives. The novelty of our approach is in the scale and scope of probing Imagenet models for adversarial artifacts. In addition, we propose an improvement to feature squeezing, another common adversarial example detection method.

Rounds, Jeremiah↗
DOE OSTI2021

Distributed State Estimation Over Time-Varying Graphs: Exploiting the Age-of-Information

Here, we study the problem of designing a distributed observer for an LTI system over a time-varying communication graph. The limited existing work on this topic imposes various restrictions either on the observation model or on the sequence of communication graphs. In contrast, we propose a single-time-scale distributed observer that works under mild assumptions. Specifically, our communication model only requires strong-connectivity to be preserved over non-overlapping, contiguous intervals that are even allowed to grow unbounded over time. We show that under suitable conditions that bound the growth of such intervals, joint observability is sufficient to track the state of any discrete-time LTI system exponentially fast, at any desired rate. We also develop a variant of our algorithm that is provably robust to worst-case adversarial attacks, provided the sequence of graphs is sufficiently connected over time. The key to our approach is the notion of a "freshness-index" that keeps track of the age-of-information being diffused across the network. Such indices enable nodes to reject stale estimates of the state, and, in turn, contribute to stability of the error dynamics.

42 ENGINEERING↗
NASA NTRSDate not supplied

NNREPAIR: Constraint-based Repair of Neural Network Classifiers

We present NNREPAIR, a constraint-based technique for repairing neural network classifiers by i) fixing the logic of the network at an intermediate layer, and ii) fixing the decision conditions of the network at the last layer. NNREPAIR first uses fault localization to find potentially faulty network parameters (specifically weights) and then performs repair using constraint solving to apply small modifications to remedy the defects. We present novel strategies to enable precise yet efficient repair such as inferring correctness specifications to act as oracles for intermediate layer repair, and generation of experts wrt each class. We demonstrate the technique in the context of three different scenarios: (1) Improving the over all accuracy of a model, (2) Fixing security vulnerabilities caused by poisoning of training data and (3) Improving the robustness of the network against adversarial attacks. Our evaluation on MNIST and CIFAR-10 models shows that NNREPAIR can improve the accuracy by 45.56% on poisoned data and 11.11% on adversarial data. NNREPAIR also provides improvement, albeit modest, in the overall accuracy of models, without requiring new data or re-training.

Repair↗
DOE OSTI2024

Single-Shot Decoding of Good Quantum LDPC Codes

Abstract Quantum Tanner codes constitute a family of quantum low-density parity-check codes with good parameters, i.e., constant encoding rate and relative distance. In this article, we prove that quantum Tanner codes also facilitate single-shot quantum error correction (QEC) of adversarial noise, where one measurement round (consisting of constant-weight parity checks) suffices to perform reliable QEC even in the presence of measurement errors. We establish this result for both the sequential and parallel decoding algorithms introduced by Leverrier and Zémor. Furthermore, we show that in order to suppress errors over multiple repeated rounds of QEC, it suffices to run the parallel decoding algorithm for constant time in each round. Combined with good code parameters, the resulting constant-time overhead of QEC and robustness to (possibly time-correlated) adversarial noise make quantum Tanner codes alluring from the perspective of quantum fault-tolerant protocols.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Exploiting the Local Parabolic Landscapes of Adversarial Losses to Accelerate Black-Box Adversarial Attack

Existing black-box adversarial attacks on image classifiers update the perturbation at each iteration from only a small number of queries of the loss function. Since the queries contain very limited information about the loss, black-box methods usually require much more queries than white-box methods. We propose to improve the query efficiency of black-box methods by exploiting the smoothness of the local loss landscape. However, many adversarial losses are not locally smooth with respect to pixel perturbations. To resolve this issue, our first contribution is to theoretically and experimentally justify that the adversarial losses of many standard and robust image classifiers behave like parabolas with respect to perturbations in the Fourier domain. Our second contribution is to exploit the parabolic landscape to build a quadratic approximation of the loss around the current state, and use this approximation to interpolate the loss value as well as update the perturbation without additional queries. Since the local region is already informed by the quadratic fitting, we use large perturbation steps to explore far areas. We demonstrate the efficiency of our method on MNIST, CIFAR-10 and ImageNet datasets for various standard and robust models, as well as on Google Cloud Vision. The experimental results show that exploiting the loss landscape can help significantly reduce the number of queries and increase the success rate. Our codes are available at https://github.com/HoangATran/BABIES.

Tran, Hoang↗
DOE OSTI2021

Robustness of deep learning algorithms in astronomy -- galaxy morphology studies

Deep learning models are being increasingly adopted in wide array of scientific domains, especially to handle high-dimensionality and volume of the scientific data. However, these models tend to be brittle due to their complexity and overparametrization, especially to the inadvertent adversarial perturbations that can appear due to common image processing such as compression or blurring that are often seen with real scientific data. It is crucial to understand this brittleness and develop models robust to these adversarial perturbations. To this end, we study the effect of observational noise from the exposure time, as well as the worst case scenario of a one-pixel attack as a proxy for compression or telescope errors on performance of ResNet18 trained to distinguish between galaxies of different morphologies in LSST mock data. We also explore how domain adaptation techniques can help improve model robustness in case of this type of naturally occurring attacks and help scientists build more trustworthy and stable models.

79 ASTRONOMY AND ASTROPHYSICS↗
NASA NTRS2012

Purifying, Separating, and Concentrating Cells From a Sample Low in Biomass

Frequently there is an inability to process and analyze samples of low biomass due to limiting amounts of relevant biomaterial in the sample. Furthermore, molecular biological protocols geared towards increasing the density of recovered cells and biomolecules of interest, by their very nature, also concentrate unwanted inhibitory humic acids and other particulates that have an adversarial effect on downstream analysis. A novel and robust fluorescence-activated cell-sorting (FACS)-based technology has been developed for purifying (removing cells from sampling matrices), separating (based on size, density, morphology), and concentrating cells (spores, prokaryotic, eukaryotic) from a sample low in biomass. The technology capitalizes on fluorescent cell-sorting technologies to purify and concentrate bacterial cells from a low-biomass, high-volume sample. Over the past decade, cell-sorting detection systems have undergone enhancements and increased sensitivity, making bacterial cell sorting a feasible concept. Although there are many unknown limitations with regard to the applicability of this technology to environmental samples (smaller cells, few cells, mixed populations), dogmatic principles support the theoretical effectiveness of this technique upon thorough testing and proper optimization. Furthermore, the pilot study from which this report is based proved effective and demonstrated this technology capable of sorting and concentrating bacterial endospore and bacterial cells of varying size and morphology. Two commercial off-the-shelf bacterial counting kits were used to optimize a bacterial stain/dye FACS protocol. A LIVE/DEAD BacLight Viability and Counting Kit was used to distinguish between the live and dead cells. A Bacterial Counting Kit comprising SYTO BC (mixture of SYTO dyes) was employed as a broad-spectrum bacterial counting agent. Optimization using epifluorescence microscopy was performed with these two dye/stains. This refined protocol was further validated using varying ratios and mixtures of cells to ensure homogenous staining compared to that of individual cells, and were utilized for flow analyzer and FACS labeling. This technology focuses on the purification and concentration of cells from low-biomass spacecraft assembly facility samples. Currently, purification and concentration of low-biomass samples plague planetary protection downstream analyses. Having a capability to use flow cytometry to concentrate cells out of low-biomass, high-volume spacecraft/ facility sample extracts will be of extreme benefit to the fields of planetary protection and astrobiology. Successful research and development of this novel methodology will significantly increase the knowledge base for designing more effective cleaning protocols, and ultimately lead to a more empirical and true account of the microbial diversity present on spacecraft surfaces. Refined cleaning and an enhanced ability to resolve microbial diversity may decrease the overall cost of spacecraft assembly and/or provide a means to begin to assess challenging planetary protection missions.

Benardini, James N.↗
NASA NTRS2018

Engineering Trade-off Considerations Regarding Design-for-Security, Design-for-Verification, and Design-for-Test

The United States government has identified that application specific integrated circuit (ASIC) and field programmable gate array (FPGA) hardware are at risk from a variety of adversary attacks. This finding affects system security and trust. Consequently, processes are being developed for system mitigation and countermeasure application. The scope of this tutorial pertains to potential vulnerabilities and countermeasures within the ASIC/FPGA design cycle. The presentation demonstrates how design practices can affect the risk for the adversary to: change circuitry, steal intellectual property, and listen to data operations. An important portion of the design cycle is assuring the design is working as specified or as expected. This is accomplished by exhaustive testing of the target design. Alternatively, it has been shown that well established schemes for test coverage enhancement (design-for-verification (DFV) and design-for-test (DFT)) can create conduits for adversary accessibility. As a result, it is essential to perform a trade between robust test coverage versus reliable design implementation. The goal of this tutorial is to explain the evolution of design practices; review adversary accessibility points due to DFV and DFT circuitry insertion (back door circuitry); and to describe common engineering trade-off considerations for test versus adversary threats.

Design for reliability (DFR)↗
DOE OSTI2020

MimicGAN: Robust Projection onto Image Manifolds with Corruption Mimicking

In the past few years, Generative Adversarial Networks (GANs) have dramatically advanced our ability to represent and parameterize high-dimensional, non-linear image manifolds. As a result, they have been widely adopted across a variety of applications, ranging from challenging inverse problems like image completion, to problems such as anomaly detection and adversarial defense. A recurring theme in many of these applications is the notion of projecting an image observation onto the manifold that is inferred by the generator. In this context, Projected Gradient Descent (PGD) has been the most popular approach, which essentially optimizes for a latent vector that minimizes the discrepancy between a generated image and the given observation. However, PGD is a brittle optimization technique that fails to identify the right projection (or latent vector) when the observation is corrupted, or perturbed even by a small amount. Such corruptions are common in the real world, for example images in the wild come with unknown crops, rotations, missing pixels, or other kinds of non-linear distributional shifts which break current encoding methods, rendering downstream applications unusable. To address this, we propose corruption mimicking—a new robust projection technique, that utilizes a surrogate network to approximate the unknown corruption directly at test time, without the need for additional supervision or data augmentation. The proposed method is significantly more robust than PGD and other competing methods under a wide variety of corruptions, thereby enabling a more effective use of GANs in real-world applications. Finally, more importantly, we show that our approach produces state-of-the-art performance in several GAN-based applications—anomaly detection, domain adaptation, and adversarial defense, that benefit from an accurate projection.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

ASK: Adversarial Soft k-Nearest Neighbor Attack and Defense

K-Nearest Neighbor (kNN)-based deep learning methods have been applied to many applications due to their simplicity and geometric interpretability. However, the robustness of kNN-based deep classification models has not been thoroughly explored and kNN attack strategies are underdeveloped. In this paper, we first propose an Adversarial Soft kNN (ASK) loss for developing more effective kNN-based deep neural network attack strategies and designing better defense methods against them. Our ASK loss provides a differentiable surrogate of the expected kNN classification error. It is also interpretable as it preserves the mutual information between the perturbed input and the in-class-reference data. We use the ASK loss to design a novel attack method called the ASK-Attack (ASK-Atk), which shows superior attack efficiency and accuracy degradation relative to previous kNN attacks on hidden layers. We then derive an ASK-Defense (ASK-Def) method that optimizes the worst-case ASK training loss. Experiments on CIFAR-10 (ImageNet) show that (i) ASK-Atk achieves ≥13% (≥ 13% ) improvement in attack success rate over previous kNN attacks, and (ii) ASK-Def outperforms the conventional adversarial training method by ≥ 6.9% (≥ 3.5% ) in terms of robustness improvement. Relevant codes are available at https://github.com/wangren09/ASK .

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Evaluating Deception Detection Model Robustness To Linguistic Variation

With the increasing use of automated, machine learning-driven tools and the downstream impact that algorithmic judgements can have, it is critical to develop models that are robust to evolving or manipulated inputs. Evaluating the reliability of multimodal models across linguistic variations to understand model susceptibility to intentional linguistic adversarial attacks as well as natural linguistic variations is essential in this pursuit. We present extensive analysis of model robustness and susceptibility to linguistic variations in the setting of deceptive news detection, a difficult classification task that is an increasingly important problem to solve with the impact of misinformation spread online. We evaluate the effectiveness of incorporating adversarial defense strategies and measure model susceptibility to state-of-the-art adversarial attacks using two types of linguistic attacks — character and word perturbations. We consider two multiclass prediction tasks — a 3-way classification of tweets as trustworthy, propaganda, or disinformation; and a 4-way classification as clickbait, hoax, satire, or conspiracy — and compare the performance of three embeddings that have been state-of-the-art for several NLP tasks — GloVe, ELMo, and BERT — to highlight consistent trends in susceptibility, high confidence misclassifications, and high impact failures. We find that character or mixed ensemble models are the most effective defense mechanisms and that character perturbations are a more effective attack than word perturbations for deception classification.

adversarial evaluation↗
DOE OSTI2025

Security-by-Design: Light Water Small Modular Reactor

The growing demand for nuclear power is increasing pressure to find solutions to cost prohibitive requirements of both construction and security. Offsite response has been proposed as an option to reduce costs associated with training and maintaining an onsite response force. A previous report explored this option and revealed that security could be provided at the required level, but cost savings was not a result of this methodology. An offsite response strategy required costly active and passive delay barriers to provide sufficient time for responders to muster and deploy to a site in time to interrupt a determined and well-equipped adversary. Also, contrary to the hypothesis, the number of responders required for this strategy exceeded that needed for an onsite response force, as the adversaries could avail themselves of advantageous positions within the facility to repel arriving responders. This report builds upon the previous evaluation by using the same hypothetical light water small modular reactor (LWSMR) facility model, but this time an onsite response strategy was assessed. The goal of this analysis was to show that an onsite response strategy could be implemented effectively at a cost point that removes barriers within the industry at this critical time of growth and development. The assessment of the facility design and response strategy was completed through modeling using Scribe3D© and subsequent scenario analysis over the course of a two-day tabletop exercise. Subject matter experts in nuclear security, nuclear facility design, and response strategy and tactics contributed to the effort to ensure accurate representation of hypothetical scenarios. Several adaptations were made to the layout of the LWSMR based on lessons learned during the first day of scenario analysis. The subsequent design evaluated on the second day proved to provide a robust response posture against a large and well-trained adversary force. This report details the process of the analysis and compares the cost of the final facility design with that of the LWSMR model used for evaluation of offsite response. Ultimately, the results of this effort indicate that, when implemented correctly, an onsite response strategy is the best option from a security and cost perspective.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

Architecture Design for Remote Operation of Microreactors: Poster

The nuclear sector is pursuing a number of advanced reactor concepts, one of which is the microreactor, an advanced reactor characterized by a power output of less than 20 MWth. These reactors are intended for use in applications where traditional power solutions are economically or logistically impractical. One key feature required for the successful deployment of microreactors is a remote operation capability, which can dramatically cut staffing expenses by eliminating the need for licensed operators to be present at the site of each reactor and instead concentrate in a centrally located operations center. In moving to a remote operations framework for nuclear reactors, the number of potential attack surfaces for a cyber adversary looking to cause harm or disruption increases. Therefore, a robust cybersecurity architecture is required to mitigate these potential vulnerabilities. This paper explores the functional infrastructure, security, and communication requirements to adapt remote operations to nuclear applications. It then presents a reference architecture for remote operations of microreactors that applies best practices in cybersecurity and remote communications in the context of a digital twin remote operation system.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2025

Deploying Adversarial Attacks in Super-Resolution Models

Reliable super-resolution methods are crucial for applications like remote sensing, grid resilience and disaster impact analysis, and standoff biometrics. These methods infuse additional high-frequency information into reconstructions, allowing for better contextualization and image intelligence. However, super-resolution models can also introduce hallucinations or other unseen vulnerabilities that could be exploited by an adversary. This is further compounded by the prominence of deep learning in these models, as models are often blindly applied on out-of-distribution images. In this work, we implement adversarial attacks in common open-source super-resolution models and examine their impact on reconstructions and downstream classification tasks. We find that an adversarially trained super-resolution model can produce high-quality reconstructions that degrade downstream classifications. Moreover, these attacks do not require access to low-resolution imagery or class labels at inference time. These results demonstrate the vulnerability of super-resolution methods to malicious actors and motivates the development of a detector for super-resolution adversarial attacks. Further exploration of adversarial attacks in this domain is required to ensure trustworthiness and robustness of super-resolution models for national security applications.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗
DOE OSTI2023

Transferable Adversarial Attack on 3D Object Tracking in Point Cloud

3D point cloud object tracking has recently witnessed considerable progress relying on deep learning. Such progress, however, mainly focuses on improving tracking accuracy. The risk, especially considering that deep neural network is vulnerable to adversarial perturbations, of a tracker being attacked is often neglected and rarely explored. In order to attract attentions to this potential risk and facilitate the study of robustness in point cloud tracking, we introduce a novel transferable attack network (TAN) to deceive 3D point cloud tracking. Specifically, TAN consists of a 3D adversarial generator, which is trained with a carefully designed multi-fold drift (MFD) loss. The MFD loss considers three common grounds, including classification, intermediate feature and angle drifts, across different 3D point cloud tracking frameworks for perturbation generation, leading to high transferability of TAN for attack. In our extensive experiments, we demonstrate the proposed TAN is able to not only drastically degrade the victim 3D point cloud tracker, \ie, P2B, but also effectively deceive other unseen state-of-the-art approaches such as BAT and M^2Track, posing a new threat to 3D point cloud tracking.

97 MATHEMATICS AND COMPUTING↗