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.

10 records
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

QuantifyML: How good is my machine learning model?

This paper presents an approach, QuantifyML, which employs model counting to assess the learnability and robustness of machine learning models. Typically the efficacy of machine learning models is determined by computing their accuracy statistically on test data sets. However, this may be misleading, if the test data is not representative of the problem that is being studied. Further, two different models may have the same accuracy on a given data set, measured statistically, but may be very different in their behavior on unseen data. Also, models with high accuracy could have poor adversarial robustness. In QuantifyML, our goal is to precisely quantify the extent to which machine learning models have learned and generalized from the given data. In QuantifyML, a trained model is translated into a C program, which is fed to the CBMC model checking tool to produce a formula in Conjunctive Normal Form (CNF), which in turn is analyzed with state-of-the-art model counters to efficiently obtain precise counts w.r.t different outputs. QuantifyML enables i) evaluating the learnability of models by comparing the counts for the outputs to ground truth, expressed as logical predicates (if available), ii) comparing the performance of different models that may be built with different machine learning algorithms (e.g., decision-trees vs. neural networks), and iii) quantifying the robustness of trained models around given inputs. Our evaluation demonstrates these applications of QuantifyML on decision trees and neural networks trained to learn relational properties of graphs, for which we know the ground truth, and to perform image classification, for which we do not have the ground truth, but we can quantify local robustness.

Deep Neural Networks↗
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↗
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)↗
NASA NTRS2020

CyberGAN: Generating High-fidelity Cybersecurity Data With Generative Adversarial Networks

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.

Zhang, Yuening↗
NASA NTRSDate not supplied

An Inverse Chance-constrained Approach to the Calibration of Robust Models

This paper proposes a strategy to calibrate computational models according to uncertain input-output data. To this end, uncertainty in the data is first quantified by creating adversarial data sets. Samples drawn from such sets are then mapped from the input-output space to the parameter space using an inverse mapping. This mapping minimizes the collective output spread of an ensemble of point predictions while satisfying a set of individual data-matching requirements. The distribution of the resulting parameter points, which often exhibits strong parameter dependencies, is then modeled using sliced-normals. The chance-constrained formulation used to learn this distribution enables the analyst to trade-off a greater likelihood for most of the data against a lower likelihood for some of the data thereby relaxing the conservatism of the calibrated model. This formulation not only neglects the worst-performing quantiles of each adversarial distribution but also eliminates the potentially serious effects that outliers might have on the resulting model. This calibration approach not only has a considerably lower computational cost than the standard forward approach but it also allows for the identification of suitable distribution classes, which in turn yield better calibrated models.

Calibration↗
NASA NTRSDate not supplied

SafeDNN: Understanding and Verifying Neural Networks

The SafeDNN project at NASA Ames explores analysis techniques and tools to ensure that systems that use Deep Neural Networks (DNN) are safe, robust and interpretable. Research directions we are pursuing include: symbolic execution for DNN analysis, label-guided clustering to automatically identify input regions that are robust, parallel and compositional approaches to improve formal SMT-based verification, property inference and automated program repair for DNNs, adversarial training and detection, probabilistic reasoning for DNNs. In this talk I will highlight some of the research advances from SafeDNN, that were already published.

Corina Pasareanu↗
NASA NTRS2011

Process Algebra Approach for Action Recognition in the Maritime Domain

The maritime environment poses a number of challenges for autonomous operation of surface boats. Among these challenges are the highly dynamic nature of the environment, the onboard sensing and reasoning requirements for obeying the navigational rules of the road, and the need for robust day/night hazard detection and avoidance. Development of full mission level autonomy entails addressing these challenges, coupled with inference of the tactical and strategic intent of possibly adversarial vehicles in the surrounding environment. This paper introduces PACIFIC (Process Algebra Capture of Intent From Information Content), an onboard system based on formal process algebras that is capable of extracting actions/activities from sensory inputs and reasoning within a mission context to ensure proper responses. PACIFIC is part of the Behavior Engine in CARACaS (Cognitive Architecture for Robotic Agent Command and Sensing), a system that is currently running on a number of U.S. Navy unmanned surface and underwater vehicles. Results from a series of experimental studies that demonstrate the effectiveness of the system are also presented.

process algebras↗
NASA NTRSDate not supplied

NEUROSPF: A Tool For the Symbolic Analysis of Neural Networks

This paper presents NEUROSPF, a tool for the symbolic analysis of neural networks. Given a trained neural network model, the tool extracts the architecture and model parameters and translates them into a Java representation that is amenable for analysis using the Symbolic PathFinder symbolic execution tool. Notably, NEUROSPF encodes specialized peer classes for parsing the model’s parameters, thereby enabling efficient analysis. With NEUROSPF the user has the flexibility to specify either the inputs or the network internal parameters as symbolic, promoting the application of program analysis and testing approaches from software engineering to the field of machine learning. For instance, NEUROSPF can be used for coverage-based testing and test generation, finding adversarial examples and also constraint-based repair of neural networks, thus improving the reliability of neural networks and of the applications that use them.

neural networks↗
NASA NTRSDate not supplied

Combining Data with Physical Knowledge for Uncertainty Quantification in Certification and Reliability Analysis

Unifying empirical data with predictive models can enable engineering cost-savings through certification by analysis and reliability-based design. Both concepts require rigorous uncertainty quantification (UQ) and robust understanding and treatment of relevant physics. Combining sampling-based UQ algorithms with high-fidelity simulations creates a computational bottleneck that is often alleviated through the use of machine learning (ML). ML can be used to create computationally efficient surrogates for simulations of complex or high-dimensional physical interactions (e.g., multi-phase interactions associated with melt pools in laser powder bed fusion or spatially-dependent material properties in functionally graded materials). However, negative side effects of ML may include a lack of interpretability and negative correlation between event rarity and simulation accuracy due to a lack of training data. As such, it is important to infuse ML algorithms with physics-based guardrails to provide confidence in their predictions. This talk will provide a brief review of recent NASA research at this intersection of physics-based simulation, ML, and UQ with a focus on certification and reliability analysis.

uncertainty quantification↗