SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
We describe a neural network learning algorithm that implements differential learning in a generalized backpropagation framework. The algorithm regulates model complexity during the learning procedure, generating the best low-complexity approximation to the Bayer-optimal classifier allowed by the training sample.
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.
The Department of Energy's (DOE) Geothermal Data Repository (GDR) team has integrated a Large Language Model (LLM) with the metadata and supporting documents associated with GDR datasets to create an Artificially Intelligent (AI) research assistant. By leveraging work done to make GDR metadata machine-readable and an open-source LLM integration model called the Energy Language Model, developed by the National Renewable Energy Laboratory, AskGDR serves as a virtual research assistant to GDR users. It provides answers to a variety of user-provided questions using natural language processing and generative machine learning. Users can get answers to questions about specific datasets, including inquiries about the equipment, assumptions and methodologies used in the origination of the data; or more abstract questions, such as the applicability of data to specific research fields. AskGDR improves the discoverability of geothermal data by helping guide users to datasets beyond simple keyword searches. It enables users to find data based on properties of the data, discover information contained within supporting documents, and explore data from projects related to their research objectives.
This study examined the improvement of microscopy segmentation accuracy by transfer learning from a large dataset of microscopy images called MicroNet. Many neural network encoder architectures, including VGG, Inception, and ResNet, were trained on over 100,000 labelled microscopy images from 54 classes. These pre-trained encoders were then embedded into multiple segmentation architectures including U-Net and DeepLabV3+ to evaluate segmentation performance on newly created benchmark microscopy datasets. Compared to ImageNet pre-training, models pre-trained on MicroNet generalized better to out-of-distribution micrographs taken under different imaging and sample conditions and were more accurate with less training data. When training with only a single Ni-superalloy image, pre-training on MicroNet produced a 72.2 percent reduction in relative segmentation error. These results suggest that transfer learning from large in-domain datasets generate models with learned feature representations that are more useful for downstream tasks and will likely improve any microscopy image analysis technique that can leverage pre-trained encoders.
This study examined the improvement of microscopy segmentation intersection over union accuracy by transfer learning from a large dataset of microscopy images called MicroNet. Many neural network encoder architectures were trained on over 100,000 labeled microscopy images from 54 material classes. These pre-trained encoders were then embedded into multiple segmentation architectures including UNet and DeepLabV3+ to evaluate segmentation performance on created benchmark microscopy datasets. Compared to ImageNet pre-training, models pre-trained on MicroNet generalized better to out-of-distribution micrographs taken under different imaging and sample conditions and were more accurate with less training data. When training with only a single Ni-superalloy image, pre-training on MicroNet produced a 72.2% reduction in relative intersection over union error. These results suggest that transfer learning from large in-domain datasets generate models with learned feature representations that are more useful for downstream tasks and will likely improve any microscopy image analysis technique that can leverage pre-trained encoders.
Deep generative models have become ubiquitous due to their ability to learn and sample from complex distributions. Despite the proliferation of various frameworks, the relationships among these models remain largely unexplored, a gap that hinders the development of a unified theory of AI learning. In this work, we address two central challenges: clarifying the connections between different deep generative models and deepening our understanding of their learning mechanisms. We focus on Restricted Boltzmann Machines (RBMs), a class of generative models known for their universal approximation capabilities for discrete distributions. By introducing a reciprocal space formulation for RBMs, we reveal a connection between these models, diffusion processes, and systems of coupled bosons. Our analysis shows that at initialization, the RBM operates at a saddle point, where the local curvature is determined by the singular values of the weight matrix, whose distribution follows the Marc̆enko-Pastur law and exhibits rotational symmetry. During training, this rotational symmetry is broken due to hierarchical learning, where different degrees of freedom progressively capture features at multiple levels of abstraction. This leads to a symmetry breaking in the energy landscape, reminiscent of Landau’s theory. This symmetry breaking in the energy landscape is characterized by the singular values and the weight matrix eigenvector matrix. We derive the corresponding free energy in a mean-field approximation. We show that in the limit of infinite size RBM, the reciprocal variables are Gaussian distributed. Our findings indicate that in this regime, there will be some modes for which the diffusion process will not converge to the Boltzmann distribution. To illustrate our results, we trained replicas of RBMs with different hidden layer sizes using the MNIST dataset. Our findings not only bridge the gap between disparate generative frameworks but also shed light on the fundamental processes underpinning learning in deep generative models.
While quantum state tomography is notoriously hard, most states hold little interest to practically minded tomographers. Given that states and unitaries appearing in nature are of bounded gate complexity, it is natural to ask if efficient learning becomes possible. In this work, we prove that to learn a state generated by a quantum circuit with G two-qubit gates to a small trace distance, a sample complexity scaling linearly in G is necessary and sufficient. We also prove that the optimal query complexity to learn a unitary generated by G gates to a small average-case error scales linearly in G . While sample-efficient learning can be achieved, we show that under reasonable cryptographic conjectures, the computational complexity for learning states and unitaries of gate complexity G must scale exponentially in G . We illustrate how these results establish fundamental limitations on the expressivity of quantum machine-learning models and provide new perspectives on no-free-lunch theorems in unitary learning. Together, our results answer how the complexity of learning quantum states and unitaries relate to the complexity of creating these states and unitaries. Published by the American Physical Society 2024
In this paper, we present a new machine learning (ML) workflow with unsupervised learning techniques to identify domains within atomic force microscopy (AFM) images obtained from polymer films.
Machine and deep learning (ML/DL) have transformed our approach to Earth observation and system modeling (EOSM), unifying both in view of ML/DL models as a form of data assimilation (DA). Trained on diverse Earth observation records, detailed physical models, or hybrids of both in physics-informed machine learning, ML/DL may improve upon existing Earth system model (ESM) formulations while creating entirely new classes of models. One important application of deep learning is observation synthesis, allowing ESM developers to prepare for the increased spatial, temporal, and spectral/polar resolution of proposed future observing systems. This may involve the retrospective application of learning algorithms to existing observational records, with or without physical radiative transfer models, to co-inform mission planning and ESM development while providing a degree of data continuity for new missions.
Low temperature plasmas (LTPs) are an enabling technology behind reducing device dimensions and the continuation of Moore’s Law. It is estimated that 40-45% of all process steps necessary to manufacture semiconductor devices involve LTPs. However, challenges in plasma process design and continuous incorporation of novel materials for new device architectures are pushing the limits of what is possible with current plasma technology. For example, creating higher aspect ratio structures and etching features at the atomic scale both require finer control of the ion energy/velocity at wafer surfaces. To support these types of future innovations in the plasma processing systems that Sandia and the DOE rely upon, we have developed novel diagnostics, simulations, and machine learning capabilities to discover, characterize, and predict plasma phenomena affecting the ion energy/velocity distribution function (IEDF). These efforts also supported research program development and external collaboration with industry and academia through Sandia’s Plasma Research Facility (PRF). This report will focus on the following topics and accomplishments of this three year LDRD project, briefly summarized.
Integrating monitoring data to efficiently update reservoir pressure and CO 2 plume distribution forecasts presents a significant challenge in geological carbon storage (GCS) applications. Inverse modeling techniques are commonly used to fuse observational data and refine reservoir model parameters, thereby improving state variable forecasts. However, these techniques often rely on linear or Gaussian assumptions, which can limit their effectiveness in accurately predicting state variables. Moreover, simulating large-scale three-dimensional (3D) GCS problems is computationally expensive, making iterative runs in inverse problems prohibitive. To address these challenges, we propose a conditional generative model utilizing the score-based diffusion method for real-time 3D pressure and saturation field distribution predictions. Our approach involves solving the score function with a mini-batch-based Monte Carlo estimator to generate labeled data. This data is subsequently employed to train a fully connected neural network, enabling it to learn the conditional sample generator within a supervised learning framework. This method enables the rapid generation of a large ensemble of predictions, facilitating comprehensive uncertainty quantification of state variables. Here we applied our method to forecast the dynamic 3D distributions of pressure and saturation fields over a 30-year injection period. The statistical assessment with low root mean square error (RMSE) values demonstrates that our method can accurately predict the spatiotemporal distributions of both pressure and saturation fields. Moreover, the developed conditional generative model shows high computational efficiency by generating 100 ensemble forecasts of 3D state variables in less than 10 min. The consistency between ensemble averages and ground truth values further illustrates the model’s capability to capture state variable dynamics during the CO 2 plume injection process. Notably, the ground truth values fall within the ensemble forecasts, indicating that our uncertainty quantification effectively captures variability and potential noise in the observations. Thus, the developed conditional generative model proves to be a more efficient, accurate, and practical tool for GCS applications, facilitating timely risk analysis and informed decision-making.
Predicting high-fidelity ground motions for future earthquakes is crucial for seismic hazard assessment and infrastructure resilience. Conventional empirical simulations suffer from sparse sensor distribution and geographically localized earthquake locations, while physics-based methods are computationally intensive and require accurate representations of Earth structures and earthquake sources. We propose an artificial intelligence (AI) spectrogram generator, Conditional Generative Modeling for Ground Motion (CGM-GM). CGM-GM leverages earthquake magnitudes and geographic coordinates of earthquakes and sensors as inputs, when postprocessed with phase information, capturing spatially continuous Fourier amplitude spectra (FAS) as well as properties such as P and S arrivals, and waveform durations, without explicit physics constraints. This is achieved through a probabilistic autoencoder that extracts latent distributions in the time-frequency domain and variational sequential models for prior and posterior distributions. We evaluate the performance of CGM-GM using small-magnitude earthquake records from the San Francisco Bay Area, a region with high seismic risks. Here, we report that CGM-GM demonstrates potential for complementing physics-based simulations and non-ergodic empirical ground motion models, as well as shows promise in seismology and beyond.
The resilience of safety-critical systems is gaining importance due to the rise in cyber and physical threats, especially within critical infrastructure. Traditional static resilience metrics may not capture dynamic system states, leading to inaccurate assessments and ineffective responses to cyber threats. This work aims to develop a data-driven, adaptive method for resilience metric learning. We propose a data-driven approach using inverse reinforcement learning (IRL) to learn a single, adaptive resilience metric. The method infers a reward function from expert control actions. Unlike previous approaches using static weights or fuzzy logic, this work applies adversarial inverse reinforcement learning (AIRL), training a generator and discriminator in parallel to learn the reward structure and derive an optimal policy. The proposed approach is evaluated on multiple scenarios: optimal communication network rerouting, power distribution network reconfiguration, and cyber–physical restoration of critical loads using the IEEE 123-bus system. The adaptive, learned resilience metric enables faster critical load restoration in comparison to conventional RL approaches.
No abstract available
The complexity of cause-and-effect relationships between spaceflight hazards and resulting health conditions clouds understanding of the totality of human system risk in space. In response, NASA has introduced Directed Acyclic Graphs (causal diagrams) into the human systems risk management process. These diagrams allow for a common understanding of the mechanisms that lead from unique hazards of spaceflight to the health outcomes important to agencies and astronauts. However, the paucity of available biomedical data from spaceflight creates a need for methods of validating causal models that can accommodate data from spaceflight model analogs. Here we outline one approach utilizing open-access rodent bone datasets from the Ames Life Sciences Data Archive. The properties of directed acyclic graphs themselves can provide an epistemological and statistical framework for validation of a priori causal representations of human system risk in space flight. The assumed causal connections on the graph creates sets of logical implications: variables that – if the causal diagram is correct – should be correlated, as well as sets that should be conditionally independent. By testing these implied correlations and conditional independencies both statistically and heuristically, we can provide evidence for or against specific causal pathways on the causal diagram. In addition to validation of expert-generated causal diagrams, machine learning techniques can learn the most likely structure of a causal diagram from a given dataset. Comparison with and reconciliation between machine-learned causal diagrams and expert-generated diagrams is another technique for challenging assumptions and improving our understanding of causal mechanisms. Accurately representing complex causation is essential to systemic understanding of human health risks in space travel. Having a robust system of validating causal diagrams helps us arrive at more accurate representations of causal systems. This process will be integral to developing the countermeasures necessary for extended exploration of the moon and Mars.
Providing safety guarantees for autonomous systems is difficultas these systems operate in complex environments that require the use of learning-enabled components, such as deep neural networks (DNNs) for visual perception. DNNs are hard to analyze due to their size (they can have thousands or millions of parameters), lack of formal specifications (DNNs are typically learnt from labeled data, in the absence of any formal or informal requirements), and sensitivity to small changes in the environment. We present an assume-guarantee style compositional approach for the formal verification of system-level safety properties of such autonomous systems. Our insight is that we can analyze the system in the absence of the DNN perception components by automatically synthesizing assumptions on the DNN behaviour that guarantee the satisfaction of the required safety properties. The synthesized assumptions are the weakest in the sense that they characterize the output sequences of all the possible DNNs that, plugged into the autonomous system, guarantee the required safety properties. The assumptions can be leveraged as run-time monitors over a deployed DNN to guarantee the safety of the overall system; they can also be mined to extract local specifications for use during training and testing of DNNs. We illustrate our approach on a case study taken from the autonomous airplanes domain that uses a complex DNN for perception
A supervised learning task involves constructing a mapping from input data (normally described by several features) to the appropriate outputs. Within supervised learning, one type of task is a classification learning task, in which each output is one or more classes to which the input belongs. In supervised learning, a set of training examples---examples with known output values---is used by a learning algorithm to generate a model. This model is intended to approximate the mapping between the inputs and outputs. This model can be used to generate predicted outputs for inputs that have not been seen before. For example, we may have data consisting of observations of sunspots. In a classification learning task, our goal may be to learn to classify sunspots into one of several types. Each example may correspond to one candidate sunspot with various measurements or just an image. A learning algorithm would use the supplied examples to generate a model that approximates the mapping between each supplied set of measurements and the type of sunspot. This model can then be used to classify previously unseen sunspots based on the candidate's measurements. This chapter discusses methods to perform machine learning, with examples involving astronomy.