SEARCH · Engineering Papers

Results for “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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 397 records · Page 22

All sources NASA NTRS DOE OSTI

DOE OSTI2025

Ansatz-Free Hamiltonian Learning with Heisenberg-Limited Scaling

Learning the unknown interactions that govern a quantum system is crucial for quantum information processing, device benchmarking, and quantum sensing. The problem, known as Hamiltonian learning, is well understood under the assumption that interactions are local, but this assumption may not hold for arbitrary Hamiltonians. Previous methods all require high-order inverse polynomial dependency with precision, unable to surpass the standard quantum limit and reach the gold-standard Heisenberg-limited scaling. Whether Heisenberg-limited Hamiltonian learning is possible without prior assumptions about the interaction structures, a challenge we term ansatz-free Hamiltonian learning , remains an open question. In this work, we present a quantum algorithm to learn arbitrary sparse Hamiltonians without any structure constraints using only black-box queries of the system’s real-time evolution and minimal digital controls to attain Heisenberg-limited scaling in estimation error. Our method is also resilient to state-preparation-and-measurement errors, enhancing its practical feasibility. We numerically demonstrate our ansatz-free protocol for learning physical Hamiltonians and validating analog quantum simulations, benchmarking our performance against the state-of-the-art Heisenberg-limited learning approach. Moreover, we establish a fundamental trade-off between total evolution time and quantum control on learning arbitrary interactions, revealing the intrinsic interplay between controllability and total evolution-time complexity for any learning algorithm. These results pave the way for further exploration into Heisenberg-limited Hamiltonian learning in complex quantum systems under minimal assumptions, potentially enabling new benchmarking and verification protocols.

machine learning↗

DOE OSTI2014

Deformable segmentation of 3D MR prostate images via distributed discriminative dictionary and ensemble learning

Purpose: Automatic prostate segmentation from MR images is an important task in various clinical applications such as prostate cancer staging and MR-guided radiotherapy planning. However, the large appearance and shape variations of the prostate in MR images make the segmentation problem difficult to solve. Traditional Active Shape/Appearance Model (ASM/AAM) has limited accuracy on this problem, since its basic assumption, i.e., both shape and appearance of the targeted organ follow Gaussian distributions, is invalid in prostate MR images. To this end, the authors propose a sparse dictionary learning method to model the image appearance in a nonparametric fashion and further integrate the appearance model into a deformable segmentation framework for prostate MR segmentation. Methods: To drive the deformable model for prostate segmentation, the authors propose nonparametric appearance and shape models. The nonparametric appearance model is based on a novel dictionary learning method, namely distributed discriminative dictionary (DDD) learning, which is able to capture fine distinctions in image appearance. To increase the differential power of traditional dictionary-based classification methods, the authors' DDD learning approach takes three strategies. First, two dictionaries for prostate and nonprostate tissues are built, respectively, using the discriminative features obtained from minimum redundancy maximum relevance feature selection. Second, linear discriminant analysis is employed as a linear classifier to boost the optimal separation between prostate and nonprostate tissues, based on the representation residuals from sparse representation. Third, to enhance the robustness of the authors' classification method, multiple local dictionaries are learned for local regions along the prostate boundary (each with small appearance variations), instead of learning one global classifier for the entire prostate. These discriminative dictionaries are located on different patches of the prostate surface and trained to adaptively capture the appearance in different prostate zones, thus achieving better local tissue differentiation. For each local region, multiple classifiers are trained based on the randomly selected samples and finally assembled by a specific fusion method. In addition to this nonparametric appearance model, a prostate shape model is learned from the shape statistics using a novel approach, sparse shape composition, which can model nonGaussian distributions of shape variation and regularize the 3D mesh deformation by constraining it within the observed shape subspace. Results: The proposed method has been evaluated on two datasets consisting of T2-weighted MR prostate images. For the first (internal) dataset, the classification effectiveness of the authors' improved dictionary learning has been validated by comparing it with three other variants of traditional dictionary learning methods. The experimental results show that the authors' method yields a Dice Ratio of 89.1% compared to the manual segmentation, which is more accurate than the three state-of-the-art MR prostate segmentation methods under comparison. For the second dataset, the MICCAI 2012 challenge dataset, the authors' proposed method yields a Dice Ratio of 87.4%, which also achieves better segmentation accuracy than other methods under comparison. Conclusions: A new magnetic resonance image prostate segmentation method is proposed based on the combination of deformable model and dictionary learning methods, which achieves more accurate segmentation performance on prostate T2 MR images.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2025

Synergistic learning with multi-task DeepONet for efficient PDE problem solving

Multi-task learning (MTL) is an inductive transfer mechanism designed to leverage useful information from multiple tasks to improve generalization performance compared to single-task learning. It has been extensively explored in traditional machine learning to address issues such as data sparsity and overfitting in neural networks. In this work, we apply MTL to problems in science and engineering governed by partial differential equations (PDEs). However, implementing MTL in this context is complex, as it requires task-specific modifications to accommodate various scenarios representing different physical processes. To this end, we present a multi-task deep operator network (MT-DeepONet) to learn solutions across various functional forms of source terms in a PDE and multiple geometries in a single concurrent training session. We introduce modifications in the branch network of the vanilla DeepONet to account for various functional forms of a parameterized coefficient in a PDE. Additionally, we handle parameterized geometries by introducing a binary mask in the branch network and incorporating it into the loss term to improve convergence and generalization to new geometry tasks. Our approach is demonstrated on three benchmark problems: (1) learning different functional forms of the source term in the Fisher equation; (2) learning multiple geometries in a 2D Darcy Flow problem and showcasing better transfer learning capabilities to new geometries; and (3) learning 3D parameterized geometries for a heat transfer problem and demonstrate the ability to predict on new but similar geometries. Finally, our MT-DeepONet framework offers a novel approach to solving PDE problems in engineering and science under a unified umbrella based on synergistic learning that reduces the overall training cost for neural operators.

42 ENGINEERING↗

DOE OSTI2022

The roles of learning mechanisms in services: Evidence from US residential solar installations

Non-hardware costs are majority of the cost of producing solar photovoltaic (PV) electricity. Here we use matched data on patents and over 125,000 residential PV installations to estimate the effects of three learning mechanisms in reducing PV costs: learning by doing, searching, and interacting. While previous work in this area has focused predominantly on learning by doing, we find that learning by searching and interacting are also significant mechanisms to facilitate non-hardware cost reductions. Including these two mechanisms reduces the effect of learning by doing in explaining non-hardware cost reductions by 43%. Our results suggest that prior work may overemphasize the role of learning by doing and the policies that help generate learning by doing. Analysis of the supplier-network between installers and their suppliers shows that concentrated supplier networks are associated with lower non-hardware costs, although there are key differences between installer-panel and installer-inverter manufacturer networks. An important implication is that policies for reducing non-hardware costs need to take a more complete view of how different learning mechanisms engender cost reductions. They should particularly consider the important role of learning in supplier networks in cost reductions—an effect that until now has largely been missing in analyses of solar non-hardware costs.

14 SOLAR ENERGY↗

DOE OSTI2022

A robust deep learning workflow to predict multiphase flow behavior during geological C O 2 sequestration injection and Post-Injection periods

Simulation of multiphase flow in porous media is essential to manage the geologic CO 2 sequestration (GCS) process, and physics-based simulation approaches usually take prohibitively high computational cost due to the nonlinearity of the coupled physics. This paper contributes to the development and evaluation of a deep learning workflow that accurately and efficiently predicts the temporal-spatial evolution of pressure and CO 2 plumes during injection and post-injection periods of GCS operations. Based on a Fourier Neural Operator, the deep learning workflow takes input variables or features including rock properties, well operational controls and time steps, and predicts the state variables of pressure and CO 2 saturation. To further improve the predictive fidelity, separate deep learning models are trained for CO 2 injection and post-injection periods due to the difference in primary driving force of fluid flow and transport during these two phases. We also explore different combinations of features to predict the state variables. We use a realistic example of CO 2 injection and storage in a 3D heterogeneous saline aquifer, and apply the deep learning workflow that is trained from physics-based simulation data and emulate the physics process. Through this numerical experiment, we demonstrate that using two separate deep learning models to distinguish post-injection from injection period generates the most accurate prediction of pressure, and a single deep learning model of the whole GCS process including the cumulative injection volume of CO 2 as a deep learning feature, leads to the most accurate prediction of CO 2 saturation. For the post-injection period, it is key to use cumulative CO 2 injection volume to inform the deep learning models about the total carbon storage when predicting either pressure or saturation. The deep learning workflow not only provides high predictive fidelity across temporal and spatial scales, but also offers a speedup of 250 times compared to full physics reservoir simulation, and thus will be a significant predictive tool for engineers to manage the long-term process of GCS.

58 GEOSCIENCES↗

DOE OSTI2019

Memristor design rules for dynamic learning and edge processing applications

The ability to dynamically learn and adapt to changes in the environment is one of the hallmarks of biological systems. In this work, we identify the subset of the design space of memristive materials that is optimal for dynamic learning applications. Using an architecture inspired on the learning center of the insect brain, we implement a model system consisting of a discrete implementation of spiking neurons where dynamic learning takes place on a set of plastic synapses formed by memristor pairs in a crossbar array. Using two separate benchmarks, one comprising the dynamic learning of the Modified National Institute of Standards and Technology dataset and another one targeting one shot learning, we have identified the key properties that memristive materials should have to be optimal dynamic learners. The results obtained show that a fine degree of control of the memristor internal state is key to achieve high classification accuracy during dynamic learning but that within this optimal region learning is extremely robust both to device variability and to errors in the writing of the internal state, in all cases allowing for 2 sigma variations greater than 40% without significant loss of accuracy, hence overcoming two of the perceived limitations of memristors. By additionally requiring that learning takes place concurrently to information processing, we are able to derive a set constraints to the memristor dynamics.

36 MATERIALS SCIENCE↗

DOE OSTI2016

SU-A-BRA-02: Making the Most of a One Hour Lecture with Alternative Teaching Methodologies: Implementing Project-Based and Flipped Learning

Vic Montemayor - No one has been more passionate about improving the quality and effectiveness of the teaching of Medical Physics than Bill Hendee. It was in August of 2008 that the first AAPM Workshop on Becoming a Better Teacher of Medical Physics was held, organized and run by Bill Hendee. This was followed up in July of 2010 with a summer school on the same topic, again organized by Bill. There has been continued interest in alternate approaches to teaching medical physics since those initial gatherings. The momentum established by these workshops is made clear each year in the annual Innovation in Medical Physics Education session, which highlights work being done in all forms of medical physics education, from one-on-one residencies or classroom presentations to large-scale program revisions and on-line resources for international audiences. This symposium, presented on behalf of the Education Council, highlights the work of three finalists from past Innovation in Education sessions. Each will be presenting their approaches to and innovations in teaching medical physics. It is hoped that audience members interested in trying something new in their teaching of medical physics will find some of these ideas and approaches readily applicable to their own classrooms. Rebecca Howell - The presentation will discuss ways to maximize classroom learning, i.e., increasing the amount of material covered while also enhancing students’ understanding of the broader implications of the course topics. Specifically, the presentation will focus on two teaching methodologies, project based learning and flip learning. These teaching methods will be illustrated using an example of graduate medical physics course where both are used in conjunction with traditional lectures. Additionally, the presentation will focus on our experience implementing these methods including challenges that were overcome. Jay Burmeister - My presentation will discuss the incorporation of active learning techniques into a traditional medical physics classroom course. I will describe these techniques and how they were implemented as well as student performance before and after implementation. Student feedback indicated that these course changes improved their ability to actively assimilate the course content, thus improving their understanding of the material. Shahid Naqvi - My talk will focus on ways to help students visualize crucial concepts that lie at the core of radiation physics. Although particle tracks generated by Monte Carlo simulations have served as an indispensable visualization tool, students often struggle to resolve the underlying physics from a simultaneous jumble of tracks. We can clarify the physics by “coding” the tracks, e.g., by coloring the tracks according to their “starting” or “crossing” regions. The regionally-coded tracks when overlaid with dose distributions help the students see the elusive connection between dose, kerma and electronic disequilibrium. Tracks coded according to local energy or energy-loss rate can illustrate the need for stopping power corrections in electron beams and explain the Bragg peak in a proton beam. Coding tracks according to parent interaction type and order can clarify the often misunderstood distinction between primary and scatter dose. The students can thus see the “whole” simultaneously with the “sum of the parts,” which enhances their physical insight and creates a sustainable foundation for further learning. After the presentations the speakers and moderator will be open to questions and discussion with the audience members. Learning Objectives: Be able to explain Project-Based Learning and how can it be incorporated into a Medical Physics classroom. Be able to explain Flipped Learning and how can it be incorporated into a Medical Physics classroom. Be able to explain active-learning strategies for the teaching of Medical Physics. Be able to explain how Monte Carlo simulations can be used to deepen a student’s understanding of radiation physics and dosimetry.

07 ISOTOPE AND RADIATION SOURCES↗

DOE OSTI2016

SU-A-BRA-04: Incorporating Active Learning Into Medical Physics Education

07 ISOTOPE AND RADIATION SOURCES↗

DOE OSTI2020

The importance of round-robin validation when assessing machine-learning-based vertical extrapolation of wind speeds

The extrapolation of wind speeds measured at a meteorological mast to wind turbine rotor heights is a key component in a bankable wind farm energy assessment and a significant source of uncertainty. Industry-standard methods for extrapolation include the power-law and logarithmic profiles. The emergence of machine-learning applications in wind energy has led to several studies demonstrating substantial improvements in vertical extrapolation accuracy in machine-learning methods over these conventional power-law and logarithmic profile methods. In all cases, these studies assess relative model performance at a measurement site where, critically, the machine-learning algorithm requires knowledge of the rotor-height wind speeds in order to train the model. This prior knowledge provides fundamental advantages to the site-specific machine-learning model over the power-law and log profiles, which, by contrast, are not highly tuned to rotor-height measurements but rather can generalize to any site. Furthermore, there is no practical benefit in applying a machine-learning model at a site where winds at the heights relevant for wind energy production are known; rather, its performance at nearby locations (i.e., across a wind farm site) without rotor-height measurements is of most practical interest. To more fairly and practically compare machine-learning-based extrapolation to standard approaches, we implemented a round-robin extrapolation model comparison, in which a random-forest machine-learning model is trained and evaluated at different sites and then compared against the power-law and logarithmic profiles. We consider 20 months of lidar and sonic anemometer data collected at four sites between 50 and 100 km apart in the central United States. We find that the random forest outperforms the standard extrapolation approaches, especially when incorporating surface measurements as inputs to include the influence of atmospheric stability. When compared at a single site (the traditional comparison approach), the machine-learning improvement in mean absolute error was 28 % and 23 % over the power-law and logarithmic profiles, respectively. Using the round-robin approach proposed here, this improvement drops to 20 % and 14 %, respectively. These latter values better represent practical model performance, and we conclude that round-robin validation should be the standard for machine-learning-based wind speed extrapolation methods.

17 WIND ENERGY↗

DOE OSTI2026

Enhancing segmentation fairness through curriculum learning and progressive loss: a centralized and federated perspective on radiograph analysis

Bias in medical image segmentation can lead to unequal performance across demographic subgroups, raising concerns about fairness and reliability in clinical AI systems. While deep learning models have achieved high segmentation accuracy, ensuring equitable performance across race and gender remains a significant challenge, particularly in privacy-sensitive healthcare environments. This study investigates fairness-aware medical image segmentation for hip and knee radiographs using deep learning models evaluated in both centralized and Federated Learning (FL) settings. We introduce Curriculum Learning (CL) strategies and Progressive Loss (PL) functions to regulate sample difficulty during training. In addition, we propose two novel fairness-oriented federated learning algorithms, Federated Intersection over Union (FedIoU) and Federated Intersection over Union with Outlier Analysis (FedIoUoutlier). Experiments are conducted using multiple segmentation backbones and simulated multi-site data partitions derived from the Osteoarthritis Initiative dataset. Model performance is evaluated using Intersection over Union (IoU), IoU standard deviation, Skewed Error Ratio (SER), and Min-Max Disparity across race and gender subgroups. Statistical significance was verified using paired t-tests to compare per-sample IoU performance against baseline configurations. Across both hip and knee segmentation tasks, curriculum learning and progressive loss strategies consistently improved segmentation accuracy and reduced demographic performance disparities in centralized training. In federated settings, fairness-aware aggregation further enhanced performance. Notably, FedIoUoutlier combined with balanced curriculum learning and tiered progressive loss achieved the highest mean IoU while yielding the lowest SER and Min-Max Disparity, indicating improved fairness without sacrificing accuracy. In several configurations, federated models matched or exceeded the performance of optimized centralized models, with statistically significant improvements in per-sample IoU over baseline configurations. The results demonstrate that structured training strategies and fairness-aware federated aggregation can jointly improve accuracy, stability, and demographic fairness in medical image segmentation. By integrating curriculum learning, progressive loss, and novel FL algorithms, this work provides a practical pathway toward equitable and privacy-preserving AI systems for medical imaging.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Explore Spatio‐Temporal Learning of Large Sample Hydrology Using Graph Neural Networks

Abstract Streamflow forecasting over gauged and ungauged basins play a vital role in water resources planning, especially under the changing climate. Increased availability of large sample hydrology data sets, together with recent advances in deep learning techniques, has presented new opportunities to explore temporal and spatial patterns in hydrological signatures for improving streamflow forecasting. The purpose of this study is to adapt and benchmark several state‐of‐the‐art graph neural network (GNN) architectures, including ChebNet, Graph Convolutional Network (GCN), and GraphWaveNet, for end‐to‐end graph learning. We explicitly represent river basins as nodes in a graph, learn the spatiotemporal nodal dependencies, and then use the learned relations to predict streamflow simultaneously across all nodes in the graph. The efficacy of the developed GNN models is investigated using the Catchment Attributes and MEteorology for Large‐sample Studies (CAMELS) data set under two settings, fixed graph topology (transductive learning), and variable graph topology (inductive learning), with the latter applicable to prediction in ungauged basins (PUB). Results indicate that GNNs are generally robust and computationally efficient, achieving similar or better performance than a baseline model trained using the long short‐term memory (LSTM) network. Further analyses are conducted to interpret the graph learning process at the edge and node levels and to investigate the effect of different model configurations. We conclude that graph learning constitutes a viable machine learning‐based method for aggregating spatiotemporal information from a multitude of sources for streamflow forecasting

Sun, Alexander Y.↗

DOE OSTI2022

Scientific machine learning benchmarks

Deep learning has transformed the use of machine learning technologies for the analysis of large experimental datasets. In science, such datasets are typically generated by large-scale experimental facilities, and machine learning focuses on the identification of patterns, trends and anomalies to extract meaningful scientific insights from the data. In upcoming experimental facilities, such as the Extreme Photonics Application Centre (EPAC) in the UK or the international Square Kilometre Array (SKA), the rate of data generation and the scale of data volumes will increasingly require the use of more automated data analysis. Furthermore, at present, identifying the most appropriate machine learning algorithm for the analysis of any given scientific dataset is a challenge due to the potential applicability of many different machine learning frameworks, computer architectures and machine learning models. Historically, for modelling and simulation on high-performance computing systems, these issues have been addressed through benchmarking computer applications, algorithms and architectures. Extending such a benchmarking approach and identifying metrics for the application of machine learning methods to open, curated scientific datasets is a new challenge for both scientists and computer scientists. Here, we introduce the concept of machine learning benchmarks for science and review existing approaches. As an example, we describe the SciMLBench suite of scientific machine learning benchmarks.

42 ENGINEERING↗

DOE OSTI2023

Use of machine learning to analyze chemistry card sort tasks

Education researchers are deeply interested in understanding the way students organize their knowledge. Card sort tasks, which require students to group concepts, are one mechanism to infer a student’s organizational strategy. However, the limited resolution of card sort tasks means they necessarily miss some of the nuance in a student’s strategy. Here in this work, we propose new machine learning strategies that leverage a potentially richer source of student thinking: free-form written language justifications associated with student sorts. Using data from a university chemistry card sort task, we use vectorized representations of language and unsupervised learning techniques to generate qualitatively interpretable clusters, which can provide unique insight in how students organize their knowledge. We compared these to machine learning analysis of the students’ sorts themselves. Machine learning-generated clusters revealed different organizational strategies than those built into the task; for example, sorts by difficulty or even discipline. There were also many more categories generated by machine learning for what we would identify as more novice-like sorts and justifications than originally built into the task, suggesting students’ organizational strategies converge when they become more expert-like. Finally, we learned that categories generated by machine learning for students’ justifications did not always match the categories for their sorts, and these cases highlight the need for future research on students’ organizational strategies, both manually and aided by machine learning. In sum, the use of machine learning to analyze results from a card sort task has helped us gain a more nuanced understanding of students’ expertise, and demonstrates a promising tool to add to existing analytic methods for card sorts.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

Versatile feature learning with graph convolutions and graph structures

Graphs represent real world relationships, and graph embedding projects nodes in a graph to a latent space that can help simplify downstream tasks. Recent development of graph convolutions in deep learning significantly improves the performance of many learning tasks on graphs. Unfortunately, prior embedding methods either do not embed graphs with node features, or fail to produce high-quality embeddings for downstream learning tasks that result in large performance gap in comparison to direct learning on graphs.We present a versatile and effective embedding method, Conv2Vec, for embedding graphs with or without node features. It is based on graph convolutions with objective functions motivated by concepts and structures from classical graph algorithms. Conv2Vec produce high-quality embedding for both plain graphs and graphs with node features for downstream tasks.We evaluate the embeddings generated by Conv2Vec with a transductive node classification task. With the generated embeddings and very simple machine learning approaches, we are able to achieve accuracies similar to those achieved by direct learning with graph convolutions. Interestingly, if we strip the node features from the graph and thus learning an embedding has to rely entirely on the graph topology, node classification with our embedding significantly outperforms direct learning with various graph convolutions. This suggests that structures from classical graph algorithms may play an important role in learning on graphs.

Cong, Guojing↗

DOE OSTI2021

Evolutionary NAS in Light of Model Stability for Accurate Continual Learning

Continual learning, the capability to learn new knowledge from streaming data without forgetting the previous knowledge, is a critical requirement for dynamic learning systems, especially for emerging edge devices such as self-driving cars and drones. However, continual learning is still facing the catastrophic forgetting problem. Previous work illustrate that model performance on continual learning is not only related to the learning algorithms but also strongly dependent on the inherited model, i.e., the model where continual learning starts. The better stability of the inherited model, the less catastrophic forgetting and thus, the inherited model should be elaborately selected. Inspired by this finding, we develop an evolutionary neural architecture search (ENAS) algorithm that emphasizes the Stability of the inherited model, namely ENAS-S. ENAS-S aims to find optimal architectures for accurate continual learning on edge devices. On CIFAR-10 and CIFAR-100, we present that ENAS-S achieves competitive architectures with lower catastrophic forgetting and smaller model size when learning from a data stream, as compared with handcrafted DNNs.

Du, Xiaocong↗

DOE OSTI2019

Computationally Efficient Training of Deep Neural Networks via Transfer Learning

Transfer learning is a highly successful approach to train deep neural networks for customized image recognition tasks. Transfer learning leverages deep learning models that are previously trained on massive datasets and re-trains the network for a novel image recognition dataset. Typically, the advantage of transfer learning has been measured in sample efficiency, but instead, we investigate the computational efficiency of transfer learning. A good pre-trained model provides features that can be used as input to a new classifier (usually the top layers of a neural network). We show that if a good pre-trained model is selected, that training a new classifier can be much more computationally efficient than training a deep neural network without transfer learning. The first step in transfer learning is to select a pre-trained model to use as a feed-forward network for generating features. Here this selection process either uses human intuition/convenience or a methodical but computationally intensive validation loop. Instead, we would prefer a method to select the pre-trained model that will produce the best transfer results for the least amount of computation. To this end, we provide a computationally efficient metric for the fit between a pre-trained model and a novel image recognition task. The better the fit, the less computation will be needed to re-train the pre-trained model for the novel task. As machine learning becomes ubiquitous, highly-accurate trained models will proliferate and computationally efficient transfer learning methods will enable rapid development of new image recognition models.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

A Data-Agnostic, Continuous Machine Learning Framework for Application in High Energy Physics and Beyond: Phase 1 Final Scientific/Technical Report

This Phase 1 effort has focused on the development of continual learning frameworks for use in machine learning, specifically in the applied context of High Energy Physics (HEP). Machine learning (ML) is a transformative technology by which computers, typically through the use of neural networks, are able to perform tasks with proficiency that rivals or surpasses that of human users. Model Degradation & Catastrophic Forgetting are two undesired phenomena which can occur in ML where the performance of a model degrades when either deployed on novel data streams, or trained on novel data which are sufficiently different than the data the models were initially trained on. A natural example where these sorts of effects can be observed is in the performance of detectors in harsh environments, where the detector signature may change over the lifetime of the detector as it ages and deteriorates — precisely what occurs in the experiments conducted in HEP. Real world HEP data is therefore an excellent test-ground and use-case for Continual Learning paradigms, which are techniques used in ML to counteract these problems. Ensemble learning is one such technique, where multiple smaller models are trained on subsets of the overall data and are ensembled together during inference. The intuition behind this technique is that, although there are shifts in the distributions which govern the incoming data streams, these shifts are not expected to be homogeneous or global. If a sufficient diversity in solutions within the various sub-models has been achieved, then at least one sub-model is expected to retain its performance within the overall ensemble. One further strength of this approach is that the architectures of the various models do not need to be identical, and in fact even different modalities of data can naturally be combined in this way. This work focused on applying ensemble learning techniques to derive results using two main datasets, anomaly detection in HEP data & time-series forecasting in semiconductor manufacturing data. Semiconductor manufacturing involves data with surprising similarity to that of HEP (e.g. wafer maps look very similar to digi-occupancy maps) and Cerium Lab’s prominence within the semiconductor industry makes semiconductor manufacturing a natural opportunity for commercialization of this work. Our efforts have led to two strong results. The first is that we evaluated the proposed ensembling techniques using previously proposed machine learning architectures for use in anomaly detection, namely AutoEncoder based models and their derivatives. We also developed new architectures which have not been evaluated in this context before. In fact, this work marks the first use of Vision Transformers for anomaly detection in HEP. Second, we demonstrated that ensemble learning significantly improves model performance in scenarios prone to degradation, validating its effectiveness across both HEP and semiconductor datasets. These results further support ensemble learning as a powerful strategy for mitigating catastrophic forgetting and maintaining robust performance in evolving data environments.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2020

Reinforcement learning based automated history matching for improved hydrocarbon production forecast

History matching aims to find a numerical reservoir model that can be used to predict the reservoir performance. An engineer and model calibration (data inversion) method are required to adjust various parameters/properties of the numerical model in order to match the reservoir production history. In this study, we develop deep neural networks within the reinforcement learning framework to achieve automated history matching that will reduce engineers’ efforts, human bias, automatically and intelligently explore the parameter space, and remove the need of large set of labeled training data. To that end, a fast-marching-based reservoir simulator is encapsulated as an environment for the proposed reinforcement learning. The deep neural-network-based learning agent interacts with the reservoir simulator within reinforcement learning framework to achieve the automated history matching. Reinforcement learning techniques, such as discrete Deep Q Network and continuous Deep Deterministic Policy Gradients, are used toth, used to train the learning agents. The continuous actions enable the Deep Deterministic Policy Gradients to explore more states at each iteration in a a learning episode; consequently, a better history matching is achieved using this algorithm as compared to Deep Q Network. For simplified dual-target composite reservoir models, the best history-matching performances of the discrete and continuous learning methods in terms of normalized root mean square errors are 0.0447 and 0.0038, respectively. Furthermore, our study shows that continuous action space achieved by the deep deterministic policy gradient drastically outperforms deep Q network.

42 ENGINEERING↗