Nonlinear process identification using decision theory.
Learning technique incorporating feature selection and decision theory for finite memory nonlinear process identification
SYSTEMS ANALYSIS↗
SEARCH · Search NASA
Results for “Statistical Learning”
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.
Market Survey 2020: Commercial Clinical Decision Support Systems and Wellness Tools
For long-duration, deep space exploration missions, current methods for managing and supporting crew health and medical conditions will be unsuitable. Communication and data transmission lags will necessitate the use of a sophisticated clinical decision support system (CDSS) that will tailor diagnosis and treatment guidance that is context-sensitive for anticipated astronaut health, wellness, and medical conditions. A variety of clinical decision support (CDS) and wellness tools (WT) are currently available in the commercial market and a broad-brush survey of this market can provide an initial impression of the current state of the art which, in turn, can inform the roadmap of NASA deep space CDSS development and associated requirements. Such a survey was undertaken during the first six months of 2020 using directed convenience sampling to obtain information provided by vendors on their websites; both commercially available CDS and WT (such as those used to track and monitor nutrition, exercise, and sleep) were included. Areas assessed were item type (e.g., software/application, device); primary purpose of the item (e.g., diagnostic support, nutrition tracking); additional purposes (if any); reported features, capabilities, and functionality; setting of use (e.g., inpatient, outpatient); intended user (e.g., clinician, patient); location and sources of data/information used or produced by the item; integration with patient electronic health record (EHR); compliance with interoperability ontologies and standards (e.g., Health Level 7 [HL7], Systematized Nomenclature of Medicine – Clinical Terminology [SNOMED-CT]); and whether the item is knowledge-based (derived from research findings) or non-knowledge-based (derived through artificial intelligence, machine learning, advanced probability and statistics), among others. Ninety-seven (97) vendor websites describing 196 CDS and 73 WT (269 total) were reviewed and coded. The primary purpose of the majority of CDS reviewed is diagnosis or diagnosis/treatment/drug decision support—targeted for clinician use— and the primary purpose of the majority of WT reviewed is the monitoring of different health metrics, most often through the use of a biosensor device (e.g., blood pressure)—targeted for patient use. Very few CDS or WT appear to comply with major international interoperability standards or can be integrated with a patient’s EHR data. None consider contextual factors, such as conditions of the physical environment (e.g., CO2 levels). The majority of CDS and WT reviewed are non-knowledge, cloud- or web-based applications or software. Forty-three (43) major findings were identified and the implications those findings have for NASA will be discussed. Example major findings include: CDS-WT capabilities range from diagnosis to treatment applications, CDS-WT may be wearable or non-wearable and are technologically advanced and only a few CDS-WT tools referenced compliance to ensure interoperability, among other findings. Recommendations will also be offered that will help to address ExMC Gap, Medical-701: Enhance medical capabilities within an exploration medical system.
market survey↗
Discovery and Analysis of Rare High-Impact Failure Modes using Adversarial RL-Informed Sampling
Adaptive learning agents have tremendous potential to handle critical tasks currently performed by humans. Unfortunately, due to their complexity, it can be difficult to verify that these learning agents do not have critical failure modes. Standard verification and validation methods often do not apply directly to learning agents and Monte Carlo methods have difficulty covering even a small fraction of the state space, especially in multiagent systems or over long time horizons. To overcome this difficulty, we demonstrate an adaptive stress-testing method based on reinforcement learning of correlations that raise the probability of failure. This approach has three key properties: (1) it is able to find rare failure modes with far greater sample efficiency than Monte Carlo methods, (2) it can estimate the true probability of a failure mode despite the inherent bias in the learning method, and (3) it is capable of learning and resampling compact representations of multimodal failure spaces. These properties are important in practice as we need to find disparate failure modes while accounting for their actual relevance. This is a significant advantage over traditional adaptive stress testing methods that give abstract likelihoods of particular failure instances, but cannot estimate the probability of a broader failure mode. We test our algorithm on a simple problem from the aviation domain where an autonomous aircraft lands in gusty wind conditions. The results suggest that we can find failure modes with far fewer samples than the Monte Carlo approach and simultaneously estimate the probability of failure.
reinforcement learning↗
Operations on Graphical Models with Plates
This paper explains how graphical models, for instance Bayesian or Markov networks, can be extended to model problems in data analysis and learning. This provides a unified framework that combines lessons learned from the artificial intelligence, statistical and connectionist communities. This also offers a set of principles for developing a software generator for data analysis, whereby a learning or discovery system can be compiled from specifications. Many of the popular learning algorithms can be compiled in this way from graphical specifications. While in a sense this paper is a multidisciplinary review of learning, the main contribution here is the presentation of the material within the unifying framework of graphical models, and the observation that, as a result, the process of developing learning algorithms can be partly automated.
Buntine, Wray L.↗
Learning classification trees
Algorithms for learning classification trees have had successes in artificial intelligence and statistics over many years. How a tree learning algorithm can be derived from Bayesian decision theory is outlined. This introduces Bayesian techniques for splitting, smoothing, and tree averaging. The splitting rule turns out to be similar to Quinlan's information gain splitting rule, while smoothing and averaging replace pruning. Comparative experiments with reimplementations of a minimum encoding approach, Quinlan's C4 and Breiman et al. Cart show the full Bayesian algorithm is consistently as good, or more accurate than these other approaches though at a computational price.
Buntine, Wray↗
Statistical Engineering
This webinar provides an overview of the International Statistical Engineering Association (ISEA), and it illustrates the practice of statistical engineering at NASA. ISEA was formed to promote the study of how successful data-based problem-solving methods are leveraged to realize innovative opportunities and solve problems sustainably. ISEA is comprised of statisticians, engineers, scientists, and other professionals that exchange ideas and experiences in the development and application of statistical engineering theories. ISEA is building the body of knowledge of the statistical engineering discipline with a particular focus on improving academic preparation for tackling complex problems. Over the past 15 years, the practice of statistical engineering has gained recognition within NASA by spurring innovation and efficiency, and it has demonstrated significant impact. Aerospace research and development benefits from an application-focused statistical engineering perspective to accelerate learning, maximize knowledge, ensure strategic resource investment, and inform data-driven decisions. The second portion of this presentation provides an overview of infusing statistical engineering at NASA through pioneering case studies in aeronautics, space exploration, and atmospheric science.
Peter A Parker↗
Graph Convolutional Network-Strengthened Topic Modeling for Scientific Papers
Machine learning has been woven into statistics to modernize topic modeling over textual documents written in natural language, and scientific paper search and recommendation can consequently offer higher accuracy instead of counting on traditional keyword-based search. However, topic distribution of a paper resulted from existing topic modeling techniques only relies on the statistics of words contained in the paper itself. We argue that community users’ views of a paper may also provide insights at the time of recommendation. For example, if a paper on fake image detection has been cited heavily by machine learning papers, such a feature should be absorbed in the embedding of this paper, so that it can be recommended for future query on machine learning. In this paper, we present a Graph Convolutional Network-strengthened Topic Modeling (GCN-TM) method, which employs GCN technique to refine topic modeling of scientific papers. A citation-oriented knowledge graph is constructed, and topic modeling is mapped to feature embedding of the comprising papers. On top of its own topics carried in its content, each paper learns topics from its neighbors and revise its embedding accordingly. Our empirical studies over real-life scientific literature has proved the necessity and effectiveness of our proposed approach.
Jia Zhang↗
Interpretable Machine Learning for Molecular Biosignatures: a Novel Single-Sample Feature Importance Method That Is Sensitive To Statistical Interactions
Isotope ratio mass spectrometry (IRMS) of volatiles (e.g., CO 2 ) promises to be a powerful tool for potential biosignature detection for future missions to ocean worlds (OW) such as Europa and Enceladus. Machine learning (ML) methods for IRMS data could enable science autonomy by onboard prediction of seawater chemistry and biosignature presence. However, ML models are likely to be complex and involve statistical interactions between features (variables), which can make predictions seem opaque and enigmatic. For ML predictions as significant as extraterrestrial biosignatures, we must place extraordinary confidence in models. It is therefore essential that these models make interpretable predictions (i.e., human-understandable) and include false-prediction diagnostics. We achieve high accuracy and interpretability in ML biosignature and seawater chemistry models for OW through a nearest-neighbors feature selection tool that detects statistical interactions between predictors, constructs interaction networks for visualization of selected features working together to make a prediction, and reports single-sample feature importance scores for false-detection diagnostics. Here we develop a novel single-sample nearest-neighbors projected distance regression(ssNPDR) feature selection method that improves upon existing single-sample algorithms through the inclusion of statistical interactions while providing false-prediction diagnostics for ML models.
geochemistry↗
The Statistical Emulators of GGCMI Phase 2: Responses of Year-to-Year Variation of Crop Yield to CO 2 , Temperature, Water and Nitrogen Perturbations
Understanding the impact of climate change on year-to-year variation of crop yield is critical to global food stability and security. While crop model emulators are believed to be lightweight tools to replace the models, few emulators have been developed to capture such interannual variation of crop yield in response to climate variability. In this study, we developed a statistical emulator with a machine learning algorithm to reproduce the response of year-to-year variation of four crop yields to CO 2 (C), temperature (T), water (W), and nitrogen (N) perturbations defined in the Global Gridded Crop Model Intercomparison Project (GGCMI) phase 2. The emulators were able to explain more than 52 % of the variance of simulated yield and performed well in capturing the year-to-year variation of global average and gridded crop yield over current croplands in the baseline. With the changes in CO 2 –temperature–water–nitrogen (CTWN) perturbations, the emulators could reproduce the year-to-year variation of crop yield well over most current cropland. The variation of R and the mean absolute error was small under the single CTWN perturbations and dual-factor perturbations. These emulators thus provide statistical response surfaces of yield, including both its mean and interannual variability, to climate factors. They could facilitate spatiotemporal downscaling of crop model simulation, projecting the changes in crop yield variability in the future and serving as a lightweight tool for multi-model ensemble simulation. The emulators enhanced the flexibility of crop yield estimates and expanded the application of large-ensemble simulations of crop yield under climate change.
Climate change↗
Curating a Standardized Dataset for Statistical Biosignature Classification
In recent years, machine learning has been explored as a toolkit for planetary science and operations [Helbert, Azari]. Machine learning has been used to improve our understanding of possible biosignatures and mineral signatures to improve science return on future missions [Warren-Rhodes, Cleaves].
Biosignatures↗
NASA DEVELOP’s Communication Strategy for Sharing Applied Science Outcomes
NASA’s DEVELOP Program conducts rapid feasibility projects that enable the future workforce and current decision makers to collaborate and build capacity to use Earth science data to enhance environmental management. The program communicates its results and applications to a broad audience through multiple methods: social media, website blog posts, print materials, short project videos, an “Ambassador Corps,” scientific and policy conference presentations, community engagement activities, end-of-project presentations, project “hand-offs,” traditional publications (both gray literature and peer-reviewed), and a project archive gallery. This presentation will describe the various methods employed by DEVELOP to communicate the program’s scientific outputs, target audiences, general statistics, community response, and lessons learned.
Capacity Building↗
CLAIRE: Enabling Heterogeneous Communication Network Optimization for Robust and Resilient Operations
In this paper, we present the capabilities of the CLAIRE System which provides resilient communications for NASA in presence of interference and congestion for a heterogeneous multi-vendor network. CLAIRE increases mission science data return to improve resource efficiencies and ensures resilience in the unpredictable space environment for NASA missions and communication networks. CLAIRE provides technology / waveform agnostic cognitive control plane that is instantiated at the Application Layer (APP) so that it can ride on NASA’s HDTN bundle protocol or any other protocol stack that is used by the network. The cognitive control plane is instantiated using Heartbeats (HTBTs). CLAIRE is assisted by Wideband UHF-Ka Band RF Sensing that leverages advances in the Direct Digital Transceiver (DDTRX) technology. The Wideband RF Sensing is driven by statistical signal processing and machine learning algorithms. Interference is mitigated using Dynamic Spectrum Access (DSA). Finally, CLAIRE addresses congestion using spectrum aware packet forwarding algorithm. CLAIRE provides an extensible protocol that allows passing of RF spectrum situational awareness, cross-layer sensing, delay tolerant networking and dynamic spectrum access information that can help with network optimization. Cross-Layer Sensing (CLS) and CLAIRE Decision Engine (CDE) enable spectrum and delay aware packet forwarding and Dynamic Spectrum Access during cases of severe interference.
cognitive communications↗
Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project
The DebriSat project is a collaboration effort with the NASA Orbital Debris Program Office, the U.S. Space Force Space Systems Command Center, The Aerospace Corporation, and the University of Florida. To date, over 200,000 fragments from this ground-based, hypervelocity impact experiment have been collected, and processing is underway to determine their physical characteristics, such as material, shape, color, characteristic length, and average cross-sectional area. The x-ray process is primarily used to identify the location of the fragments and estimated size for extraction, so that these physical characteristics can be assessed. This paper proposes a machine learning-based approach to characterize materials from x-ray images of debris fragments embedded in soft-catch foam used in the DebriSat project. The novel methodology discussed in this paper will highlight the use of x-ray imagery data to characterize these fragments without extraction or a human-in-the-loop. Both supervised and unsupervised machine learning techniques are utilized with this approach to infer the physical parameters of the fragments embedded in the soft-catch foam panels used in the impact experiment based on x-ray images of the foam panels. Additionally, 3D reconstructions of the extracted fragments are created with images taken from two different angles using the structure from motion (SfM) method. The characteristic lengths and shape from the 3D reconstruction, alongside the physical characteristics of the debris, are used in the inference of the material type. To develop and test the approach, a dataset of x-ray images of debris fragments of varying sizes and materials is collected. Supervised learning methods such as convolutional neural networks (CNNs), support vector machines (SVM), decision trees, and random forest classifiers are used due to the high-dimensional feature spaces of the debris and nonlinear decision boundaries for material categorization. Given the limited pre-labeled data of embedded debris materials smaller than 10 mm, unsupervised machine learning techniques such as clustering algorithms and autoencoders are used, in addition to supervised learning methods. The clustering algorithms group similar fragments together based on their physical properties, and autoencoders reduce the dimensionality of the x ray images and extract relevant features. The performance of the proposed approach's is analyzed using a range of statistical methods, including confusion matrices, receiver operating characteristic curves, and precision-recall curves. The results are compared with those obtained using a baseline approach that relies on manual identification and classification of debris fragments. To evaluate the effectiveness of different machine learning methods, statistical tests such as t-tests, ANOVA, and cross-validation are performed, comparing the performance of CNNs, SVMs, clustering algorithms, and autoencoders. Additional analysis needs to be conducted to identify any sources of bias or variability that may affect the results, such as variations in imaging conditions or fragmentation patterns. Other topics explored are limitations, refinements, and the potential use of semi-supervised learning techniques, such as self-training to label unlabeled datasets and co-training using x-ray images taken from two different angles as two different models.
Saik Anam Siam↗
Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project
The DebriSat project is a collaboration effort with the NASA Orbital Debris Program Office, the U.S. Space Force Space Systems Command Center, The Aerospace Corporation, and the University of Florida. To date, over 200,000 fragments from this ground-based, hypervelocity impact experiment have been collected, and processing is underway to determine their physical characteristics, such as material, shape, color, characteristic length, and average cross-sectional area. The x-ray process is primarily used to identify the location of the fragments and estimated size for extraction, so that these physical characteristics can be assessed. This paper proposes a machine learning-based approach to characterize materials from x-ray images of debris fragments embedded in soft-catch foam used in the DebriSat project. The novel methodology discussed in this paper will highlight the use of x-ray imagery data to characterize these fragments without extraction or a human-in-the-loop. Both supervised and unsupervised machine learning techniques are utilized with this approach to infer the physical parameters of the fragments embedded in the soft-catch foam panels used in the impact experiment based on x-ray images of the foam panels. Additionally, 3D reconstructions of the extracted fragments are created with images taken from two different angles using the structure from motion (SfM) method. The characteristic lengths and shape from the 3D reconstruction, alongside the physical characteristics of the debris, are used in the inference of the material type. To develop and test the approach, a dataset of x-ray images of debris fragments of varying sizes and materials is collected. Supervised learning methods such as convolutional neural networks (CNNs), support vector machines (SVM), decision trees, and random forest classifiers are used due to the high-dimensional feature spaces of the debris and nonlinear decision boundaries for material categorization. Given the limited pre-labeled data of embedded debris materials smaller than 10 mm, unsupervised machine learning techniques such as clustering algorithms and autoencoders are used, in addition to supervised learning methods. The clustering algorithms group similar fragments together based on their physical properties, and autoencoders reduce the dimensionality of the x ray images and extract relevant features. The performance of the proposed approach's is analyzed using a range of statistical methods, including confusion matrices, receiver operating characteristic curves, and precision-recall curves. The results are compared with those obtained using a baseline approach that relies on manual identification and classification of debris fragments. To evaluate the effectiveness of different machine learning methods, statistical tests such as t-tests, ANOVA, and cross-validation are performed, comparing the performance of CNNs, SVMs, clustering algorithms, and autoencoders. Additional analysis needs to be conducted to identify any sources of bias or variability that may affect the results, such as variations in imaging conditions or fragmentation patterns. Other topics explored are limitations, refinements, and the potential use of semi-supervised learning techniques, such as self-training to label unlabeled datasets and co-training using x-ray images taken from two different angles as two different models.
Saik Anam Siam↗
Investigation of Error Patterns in Geographical Databases
The objective of the research conducted in this project is to develop a methodology to investigate the accuracy of Airport Safety Modeling Data (ASMD) using statistical, visualization, and Artificial Neural Network (ANN) techniques. Such a methodology can contribute to answering the following research questions: Over a representative sampling of ASMD databases, can statistical error analysis techniques be accurately learned and replicated by ANN modeling techniques? This representative ASMD sample should include numerous airports and a variety of terrain characterizations. Is it possible to identify and automate the recognition of patterns of error related to geographical features? Do such patterns of error relate to specific geographical features, such as elevation or terrain slope? Is it possible to combine the errors in small regions into an error prediction for a larger region? What are the data density reduction implications of this work? ASMD may be used as the source of terrain data for a synthetic visual system to be used in the cockpit of aircraft when visual reference to ground features is not possible during conditions of marginal weather or reduced visibility. In this research, United States Geologic Survey (USGS) digital elevation model (DEM) data has been selected as the benchmark. Artificial Neural Networks (ANNS) have been used and tested as alternate methods in place of the statistical methods in similar problems. They often perform better in pattern recognition, prediction and classification and categorization problems. Many studies show that when the data is complex and noisy, the accuracy of ANN models is generally higher than those of comparable traditional methods.
Dryer, David↗
Learning probability spaces for classification and recognition of patterns with or without supervision
Learning probability spaces for classification and recognition of patterns with or without supervision
PROBABILITY DISTRIBUTION↗
Learning process mapping heuristics under stochastic sampling overheads
A statistical method was developed previously for improving process mapping heuristics. The method systematically explores the space of possible heuristics under a specified time constraint. Its goal is to get the best possible heuristics while trading between the solution quality of the process mapping heuristics and their execution time. The statistical selection method is extended to take into consideration the variations in the amount of time used to evaluate heuristics on a problem instance. The improvement in performance is presented using the more realistic assumption along with some methods that alleviate the additional complexity.
Ieumwananonthachai, Arthur↗
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↗