Search NASA⌕ Search

SEARCH · Search NASA

Results for “adaptive 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.

At least 541 records · Page 30

Adaptive control with an expert system based supervisory level

Adaptive control is presently one of the methods available which may be used to control plants with poorly modelled dynamics or time varying dynamics. Although many variations of adaptive controllers exist, a common characteristic of all adaptive control schemes, is that input/output measurements from the plant are used to adjust a control law in an on-line fashion. Ideally the adjustment mechanism of the adaptive controller is able to learn enough about the dynamics of the plant from input/output measurements to effectively control the plant. In practice, problems such as measurement noise, controller saturation, and incorrect model order, to name a few, may prevent proper adjustment of the controller and poor performance or instability result. In this work we set out to avoid the inadequacies of procedurally implemented safety nets, by introducing a two level control scheme in which an expert system based 'supervisor' at the upper level provides all the safety net functions for an adaptive controller at the lower level. The expert system is based on a shell called IPEX, (Interactive Process EXpert), that we developed specifically for the diagnosis and treatment of dynamic systems. Some of the more important functions that the IPEX system provides are: (1) temporal reasoning; (2) planning of diagnostic activities; and (3) interactive diagnosis. Also, because knowledge and control logic are separate, the incorporation of new diagnostic and treatment knowledge is relatively simple. We note that the flexibility available in the system to express diagnostic and treatment knowledge, allows much greater functionality than could ever be reasonably expected from procedural implementations of safety nets. The remainder of this chapter is divided into three sections. In section 1.1 we give a detailed review of the literature in the area of supervisory systems for adaptive controllers. In particular, we describe the evolution of safety nets from simple ad hoc techniques, up to the use of expert systems for more advanced supervision capabilities.

Sullivan, Gerald A.↗

On the integration of reinforcement learning and approximate reasoning for control

The author discusses the importance of strengthening the knowledge representation characteristic of reinforcement learning techniques using methods such as approximate reasoning. The ARIC (approximate reasoning-based intelligent control) architecture is an example of such a hybrid approach in which the fuzzy control rules are modified (fine-tuned) using reinforcement learning. ARIC also demonstrates that it is possible to start with an approximately correct control knowledge base and learn to refine this knowledge through further experience. On the other hand, techniques such as the TD (temporal difference) algorithm and Q-learning establish stronger theoretical foundations for their use in adaptive control and also in stability analysis of hybrid reinforcement learning and approximate reasoning-based controllers.

Berenji, Hamid R.↗

Artificial neural network implementation of a near-ideal error prediction controller

A theory has been developed at the University of Virginia which explains the effects of including an ideal predictor in the forward loop of a linear error-sampled system. It has been shown that the presence of this ideal predictor tends to stabilize the class of systems considered. A prediction controller is merely a system which anticipates a signal or part of a signal before it actually occurs. It is understood that an exact prediction controller is physically unrealizable. However, in systems where the input tends to be repetitive or limited, (i.e., not random) near ideal prediction is possible. In order for the controller to act as a stability compensator, the predictor must be designed in a way that allows it to learn the expected error response of the system. In this way, an unstable system will become stable by including the predicted error in the system transfer function. Previous and current prediction controller include pattern recognition developments and fast-time simulation which are applicable to the analysis of linear sampled data type systems. The use of pattern recognition techniques, along with a template matching scheme, has been proposed as one realizable type of near-ideal prediction. Since many, if not most, systems are repeatedly subjected to similar inputs, it was proposed that an adaptive mechanism be used to 'learn' the correct predicted error response. Once the system has learned the response of all the expected inputs, it is necessary only to recognize the type of input with a template matching mechanism and then to use the correct predicted error to drive the system. Suggested here is an alternate approach to the realization of a near-ideal error prediction controller, one designed using Neural Networks. Neural Networks are good at recognizing patterns such as system responses, and the back-propagation architecture makes use of a template matching scheme. In using this type of error prediction, it is assumed that the system error responses be known for a particular input and modeled plant. These responses are used in the error prediction controller. An analysis was done on the general dynamic behavior that results from including a digital error predictor in a control loop and these were compared to those including the near-ideal Neural Network error predictor. This analysis was done for a second and third order system.

Mcvey, Eugene S.↗

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

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

Mathematics↗

Generalized Adaptive Artificial Neural Networks

Mathematical model of supervised learning by artificial neural network provides for simultaneous adjustments of both temperatures of neurons and synaptic weights, and includes feedback as well as feedforward synaptic connections. Extension of mathematical model described in "Adaptive Neurons For Artificial Neural Networks" (NPO-17803). Dynamics of neural network represented in new model by less-restrictive continuous formalism.

Tawel, Raoul↗

Machine learning-driven design and self-sensing capabilities of automotive bumper lattices for adaptive impact response

We present a novel approach to design an automotive bumper energy absorber using carbon fiber reinforced polymer composites, optimized to meet conflicting performance requirements for two distinct impact scenarios. The design must satisfy both a low-speed (2.5 mph) pendulum intrusion test, simulating vehicle-to-vehicle collisions, and a high-speed (25 mph) leg flexion test, replicating pedestrian impacts. These tests demand opposing deformation characteristics: high flexibility (deformation < 85 mm) for the former and high stiffness (deformation < 22 mm) for the latter. To address these contradictory requirements, we developed a machine learning (ML) framework for inverse optimization of lattice designs and material selection. Unlike traditional iterative design processes, our ML model directly outputs optimal design parameters and material choices based on target performance inputs. The energy absorber was fabricated using advanced additive manufacturing techniques, including extrusion deposition and digital light processing. The integration of carbon fibers provides multifunctionality to the bumper structure, enabling self-sensing capabilities through changes in electrical resistivity under compression. This electrical response demonstrates high repeatability under multiple cycles at 2% compression and exhibits distinct signatures during crack formation under high deformation. This research offers adaptive performance through innovative design methodologies and smart material integration. The approach has potential applications in various fields requiring adaptive energy absorption and real-time structural health monitoring.

Chawla, Komal [ORNL] (ORCID:0000000190327565)↗

A learning flight control system for the F8-DFBW aircraft

This report contains a complete description of a learning control system designed for the F8-DFBW aircraft. The system is parameter-adaptive with the additional feature that it 'learns' the variation of the control system gains needed over the flight envelope. It, thus, generates and modifies its gain schedule when suitable data are available. The report emphasizes the novel learning features of the system: the forms of representation of the flight envelope and the process by which identified parameters are used to modify the gain schedule. It contains data taken during piloted real-time 6 degree-of-freedom simulations that were used to develop and evaluate the system.

Montgomery, R. C.↗

Challenges and Solutions in Operation and Lay-Up of the Modular Caustic- Side Solvent Extraction Unit - 20078

The Modular Caustic-Side Solvent Extraction (CSSX) Unit (MCU) was designed and constructed to provide interim salt waste processing so that the Liquid Waste Disposition Program (LWDP) could continue waste removal and tank closure until start-up of the Salt Waste Processing Facility (SWPF). MCU processes Clarified Salt Solution (CSS), or salt solution (SS) that has undergone actinide removal/filtration, to produce two waste streams: the cesium laden Strip Effluent (SE), which is incorporated into glass at the Defense Waste Processing Facility (DWPF), and the cesium depleted Decontaminated SS (DSS), which is incorporated into grout at the Saltstone Production Facility (SPF). MCU completed start-up in 2008 with an initial operating life of 3 years and design life of 5 years. Prior to reaching the end of the design life, multiple critical components were upgraded and/or replaced to mitigate risks associated with any delays in SWPF start-up. During the remainder of operation (until June 2019), MCU was challenged to maximize processing with minimal additional modifications to repair/improve aging infrastructure, while maintaining low levels of risk to workers and the environment. Notable challenges during the extended operating life of the facility involved both mechanical and operational issues. Two of the most significant issues were the increased frequency of Process Vessel Ventilation (PVV) High Efficiency Particulate Air (HEPA) filter change-outs and the biofouling of the SE coalescer (SEC). The major impacts of both issues were that replacement of these components 1) resulted in high exposure to workers, and 2) required substantial downtime, thus hindering achievement of processing goals. PVV HEPA filter replacements are required based on dose rate and differential pressure (dP) limits. After introducing a higher-curie feed to MCU, the operating time between change outs was reduced by approximately half. SEC replacements are required based on dP limits. After restarting from an extended outage, the operating life of the SEC was significantly reduced by more than half due to biofouling. As a result of extensive troubleshooting and process improvements, the facility was able to recover performance and extend the operating life of the PVV HEPA filters and the SEC media. Through troubleshooting both of these issues, the facility was also able to learn from prior experience and adapt in order to minimize down time and maintain throughput. Although the facility took advantage of short periods of downtime to perform nonintrusive troubleshooting and minor corrective/preventative maintenance activities, longer outages were periodically required for corrective maintenance involving process cell entry. During these outages, remote tools and specialized shield plates were employed to minimize dose to workers. Mock-ups of non-routine maintenance activities were critical to early identification of potential issues and improvements, which shortened outage duration. Consistent with the facility's operational strategy, evolutions needed for MCU layup were identified and sequenced to ensure safe conditions within the facility, meet pre-defined criteria for lay-up configuration, and protect workers, while minimizing the impact to SWPF integration. De-inventory and flushing of process areas was prioritized so that the total residual inventory was reduced as efficiently as possible. This paper discusses lessons learned and best practices from troubleshooting unique issues, performing complex maintenance activities, and implementing a layup strategy that best supports the overall system mission. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Discriminative Methodology using the Unscented Transform in Special Euclidean Group SE(3) for Autonomous GN&C

Kalman filtering techniques usually compare a known dynamic model with some process noise to an updated, noisy measurement. Both sources of noise are generally assumed to be Gaussian and are filtered to obtain a more accurate estimate of the state of the system of interest. However, these methods assume that a relatively accurate dynamical model already exists. By including discriminative methodology to interpret differences in the predictive model and the measured states over time, an incorrect model can be corrected overtime, allowing a spacecraft to learn about its environment and adapt to changing conditions. In this paper, a discriminative unscented Kalman filter (DUKF) is introduced on the special Euclidean group SE(3) formulation framework. A spacecraft is initialized near a small body with no knowledge of its parameters, and using several consecutive conditional models the environment is studied and the filter is updated. The case study of a spacecraft in a slightly elliptical, inclined, relatively tight orbit around the irregular moon Phobos is simulated and discussed. The accuracy of this approach is quantifying the differences between the true state, the UKF with a poor model, and the DUKF.

GN&C↗

Distributed Inference with Sparse and Quantized Communication

Here, we consider the problem of distributed inference where agents in a network observe a stream of private signals generated by an unknown state, and aim to uniquely identify this state from a finite set of hypotheses. We focus on scenarios where communication between agents is costly, and takes place over channels with finite bandwidth. To reduce the frequency of communication, we develop a novel event-triggered distributed learning rule that is based on the principle of diffusing low beliefs on each false hypothesis. Building on this principle, we design a trigger condition under which an agent broadcasts only those components of its belief vector that have adequate innovation, to only those neighbors that require such information. We prove that our rule guarantees convergence to the true state exponentially fast almost surely despite sparse communication, and that it has the potential to significantly reduce information flow from uninformative agents to informative agents. Next, to deal with finite-precision communication channels, we propose a distributed learning rule that leverages the idea of adaptive quantization. We show that by sequentially refining the range of the quantizers, every agent can learn the truth exponentially fast almost surely, while using just 1 bit to encode its belief on each hypothesis. For both our proposed algorithms, we rigorously characterize the trade-offs between communication-efficiency and the learning rate.

42 ENGINEERING↗

Machine Learning Derived Dynamic Operating Reserve Requirements in High-Renewable Power Systems

Accurately forecasting wind and solar power output poses challenges for deeply decarbonized electricity systems. Grid operators must commit resources to provide reserves to ensure reliable operations in the face of forecast errors, a process which can increase fuel consumption and emissions. We apply neural network-based machine learning to expand the usefulness of median point forecast data by creating probabilistic distributions of short-term uncertainty in demand, wind, and solar forecasts that adapt to prevailing grid conditions. Machine learning derived estimates of forecast errors compare favorably to estimates based on incumbent methods. Reserves derived from machine learning are usually smaller than values derived using incumbent methods, which enables fuel savings during most hours. Machine learning reserves are generally larger than incumbent reserves during times of higher forecast error, potentially improving system reliability. Performance is tested using multi-stage production simulation modeling of the California Independent System Operator (CAISO) system. Machine learning reserves provide production cost and greenhouse gas (GHG) emission reductions of approximately 0.3% relative to historical 2019 requirements. Savings in the 2030 timeframe are highly dependent on battery storage capacity. At lower levels of battery capacity, savings of 0.4% from machine learning reserves are shown. Significant quantities of battery storage are expected to be added to meet California's resource adequacy needs and GHG reduction targets. Addition of these batteries saturate reserve needs and results in minimal within-hour balancing costs in 2030.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Dyn$\mathrm{AMO}$: Multi-agent reinforcement learning for dynamic anticipatory mesh optimization with applications to hyperbolic conservation laws

Here we introduce DynAMO, a reinforcement learning paradigm for Dynamic Anticipatory Mesh Optimization. Adaptive mesh refinement is an effective tool for optimizing computational cost and solution accuracy in numerical methods for partial differential equations. However, traditional adaptive mesh refinement approaches for time-dependent problems typically rely only on instantaneous error indicators to guide adaptivity. As a result, standard strategies often require frequent remeshing to maintain accuracy. In the DynAMO approach, multi-agent reinforcement learning is used to discover new local refinement policies that can anticipate and respond to future solution states by producing meshes that deliver more accurate solutions for longer time intervals. By applying DynAMO to discontinuous Galerkin methods for the linear advection and compressible Euler equations in two dimensions, we demonstrate that this new mesh refinement paradigm can outperform conventional threshold-based strategies while also generalizing to different mesh sizes, remeshing and simulation times, and initial conditions.

97 MATHEMATICS AND COMPUTING↗

Machine Learning for Automated Weld Quality Monitoring and Control

Resistance Spot Welding (RSW) is a critical process in the automotive industry, valued for its cost-effectiveness, short cycle time, and robustness. However, achieving consistent high-quality joints remains challenging due to the complex interplay of various factors, like materials, processes, and manufacturing uncertainties, etc. Under the collaborative project between Oak Ridge National Laboratory (ORNL) and General Motors (GM), we have developed a robust and expansible machine learning (ML) framework aimed at enhancing quality control in RSW. By harnessing the power of machine learning, we have developed the ability to ensure every aspect of the welding process, from the initial process design stage to the final weld joint quality. The framework operates by analyzing a variety of data streams, including in-line process signals, process parameters, materials, and postprocessed weld joint data. Through this analysis, the models have been trained to detect deviations from optimal quality standards, leveraging their ability to identify signature data patterns and anomalies within in-line signals and construct complex correlations between these signals and weld quality parameters. Meanwhile, the machine learning framework is designed to adapt to a variety of materials, including high strength steels and aluminum alloys, etc. Its flexible architecture facilitates the incorporation of diverse data sources and features, enabling precise modeling and prediction across a broad range of material properties and weld quality variables. The expansible ML frameworks represent a promising transformation in weld quality monitoring and control, empowering industry to achieve high levels of efficiency, consistency, and reliability in manufacturing.

99 GENERAL AND MISCELLANEOUS↗

Knowledge gaps for neuromorphic ionic computing

BACKGROUND Neuromorphic computing, inspired by the human brain’s ability to process information efficiently, represents a transformative approach to computation. In this Review, we explore the emerging field of neuromorphic ionic computing, which leverages ionic conduction and coupling to mimic neural processes, and identify critical knowledge gaps that must be addressed to realize its full potential. A central theme of the discussion is energy efficiency, a challenge that is both a limitation and an opportunity for this technology. Although complementary metal-oxide semiconductor (CMOS)–based neuromorphic technologies have made strides in scaling to billions of neurons and are increasingly applied in artificial intelligence and numerical computing, they remain orders of magnitude behind the human brain in terms of connectivity and energy efficiency. Neuromorphic ionic computing promises to overcome these limitations by leveraging the distinct architectural and operational principles of the brain. Our brains achieve this energy efficiency by combining several key features: using the same network elements to store and process information; using an incredibly complex and massively interconnected three-dimensional (3D) network of locally active elements that enables sparsity, robustness in the presence of noise, adaptation, and life-long learning; computing at comparatively low voltage and frequency; and last, taking advantage of a plethora of ions and small molecules as information carriers. Here, we propose that ionic computing systems can take advantage of similar features to achieve substantial gains in energy efficiency. ADVANCES Since the first reports of neuromorphic ionic behavior in nanofluidic channels, we have witnessed an explosion of reports that used ionic devices to produce synaptomimetic behaviors. However, achieving the goals of ionic computing requires not only implementation of much more sophisticated device functionality but also overcoming fundamental barriers in materials science, device architecture, and system integration. Current ionic devices, even those incorporating state-of-the-art materials, still suffer from limited functionality and stability, which restrict their performance and increase energy demands. Developing new materials with enhanced ionic properties is essential to overcome these limitations. Similarly, the design of neuromorphic devices must evolve to leverage the particular advantages of ionic processes. Existing architectures often follow a single-information-carrier logic of conventional electronics or are constructed of mesoscale fluidics, failing to capitalize on the energy-efficient mechanisms inherent to ionic systems or implement the multiple-information-carrier paradigm. Current neuromorphic chips focus on large-scale networks of analog memory elements based on mechanisms such as charge trap (flash), filamentary, phase change, or spin, which are built on top of a network of artificial CMOS neurons. Although such prototype networks have achieved impressive performance, it is difficult to envision how they can implement the key features such as massive connectivity, sophisticated plasticity, adaptability, sparsity, and “multichromatic” computing. Although small-scale devices have demonstrated promising results, integrating them, maintaining energy efficiency, and implementing temperature control as systems grow in complexity and size to computationally relevant scale remain major hurdles. Furthermore, interfacing neuromorphic ionic devices with existing computing technologies presents technical and conceptual challenges that will require innovative approaches that combine insights from neuroscience, materials science, and engineering. OUTLOOK Despite these challenges, the potential impact of neuromorphic ionic computing is profound with potential applications ranging from artificial intelligence to robotics and beyond. We also argue that neuromorphic ionic computing systems should not, at least in the beginning, compete with CMOS technologies but rather should focus on applications that require extreme energy efficiency with chemical and/or biological compatibility, such as biomedical applications (for example, brain-computer interfaces), environmental monitoring, and agricultural and food applications. Ultimately, this Review highlights the crucial role of interdisciplinary collaboration in advancing the field. Neuromorphic ionic computing is not merely a technological innovation; it represents a substantial step toward sustainable computation, aligning with the growing demand for energy-conscious solutions in a world that is increasingly reliant on data and computation.

Neuromorphic↗

Decentralized Distributed Proximal Policy Optimization (DD-PPO) for High Performance Computing Scheduling on Multi-User Systems

Resource allocation in High Performance Computing (HPC) environments presents a complex and multifaceted challenge for job scheduling algorithms. Beyond the efficient allocation of system resources, schedulers must account for and optimize multiple performance metrics, including job wait time and system throughput. Traditional heuristic-based scheduling algorithms increasingly struggle and lack the efficiency needed to meet the demands and address the complexity and scale of modern HPC systems. Consequently, recent research efforts have focused on leveraging advancements in Artificial Intelligence (AI) and Deep Learning (DL), particularly Reinforcement Learning (RL), to develop more adaptable and intelligent scheduling strategies. Previous RL-based scheduling approaches have explored a range of algorithms, from Deep Q-Networks (DQN) to Proximal Policy Optimization (PPO), and more recently, hybrid methods that integrate Graph Neural Networks (GNNs) with RL techniques. However, a common limitation across these methods is their reliance on relatively small datasets, with few methods being evaluated using large-scale, multi-million-job trace datasets representative of real-world HPC workloads. Moreover, existing RL schedulers face scalability issues due to centralized policy updates, which hinder training efficiency and performance when applied to large datasets. This study introduces a novel RL-based scheduler utilizing Decentralized Distributed Proximal Policy Optimization (DD-PPO) algorithm, which supports large-scale distributed training across multiple workers without requiring parameter synchronization at every step. By eliminating reliance on centralized updates to a shared policy, the DD-PPO scheduler enhances scalability, training efficiency, and sample utilization. Experimental validation using a large real-world dataset containing over 11.5 million job traces collected from petascale HPC systems over six years assesses the influence of dataset scale on training effectiveness and compares DD-PPO performance to traditional and advanced scheduling approaches. The experimental results demonstrate improved scheduling performance in comparison to both heuristic-based schedulers and existing RL-based scheduling algorithms.

AI↗