Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data-driven algorithm”

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 181 records · Page 10
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Semi-supervised Learning of Dynamical Systems with Neural Ordinary Differential Equations: A Teacher-Student Model Approach

Modeling dynamical systems is crucial for a wide range of tasks, but it remains challenging due to complex nonlinear dynamics, limited observations, or lack of prior knowledge. Recently, data-driven approaches such as Neural Ordinary Differential Equations (NODE) have shown promising results by leveraging the expressive power of neural networks to model unknown dynamics. However, these approaches often suffer from limited labeled training data, leading to poor generalization and suboptimal predictions. On the other hand, semi-supervised algorithms can utilize abundant unlabeled data and have demonstrated good performance in classification and regression tasks. We propose TS-NODE, the first semi-supervised approach to modeling dynamical systems with NODE. TS-NODE explores cheaply generated synthetic pseudo rollouts to broaden exploration in the state space and to tackle the challenges brought by lack of ground-truth system data under a teacher-student model. TS-NODE employs an unified optimization framework that corrects the teacher model based on the student's feedback while mitigating the potential false system dynamics present in pseudo rollouts. TS-NODE demonstrates significant performance improvements over a baseline Neural ODE model on multiple dynamical system modeling tasks.

Wang, Yu↗
DOE OSTI2023

Water resource recovery modelling 2021 (WRRmod2021 conference)

Our society is transitioning fast into the digital age, spurred by development of cheap and new sensing technology, breakthroughs in computing, and development of efficient algorithms for optimization. This transition is also visible in the field of wastewater treatment and is driving new model developments, especially by exploiting the large data sets available with many utilities. Not surprisingly, WRRmod2021 had featured a strong session on ‘data-driven models and digitalization’ focused on this hot topic. At the same time, engineering practice calls for more robust models for performance evaluation and optimization of both conventional facilities and innovative processes. As a result, the WRRmod2021 program also exhibited sessions on modelling of new process units (e.g., aerobic granular sludge), modelling of the nitrogen cycle, and integrated/plant-wide modelling.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Sequential Inference of Hospitalization Electronic Health Records Using Probabilistic Models

In the dynamic hospital setting, decision support can be a valuable tool for improving patient outcomes. Data-driven inference of future outcomes is challenging in this dynamic setting, where long sequences such as laboratory tests and medications are updated frequently. This is due in part to heterogeneity of data types and mixed-sequence types contained in variable length sequences. In this work we design a probabilistic unsupervised model for multiple arbitrary-length sequences contained in hospitalization Electronic Health Record (EHR) data. The model uses a latent variable structure and captures complex relationships between medications, diagnoses, laboratory tests, neurological assessments, and medications. It can be trained on original data, without requiring any lossy transformations or time binning. Inference algorithms are derived that use partial data to infer properties of the complete sequences, including their length and presence of specific values. We train this model on data from subjects receiving medical care in the Kaiser Permanente Northern California integrated healthcare delivery system. The results are evaluated against held-out data for predicting the length of sequences and presence of Intensive Care Unit (ICU) in hospitalization bed sequences. Our method outperforms a baseline approach, showing that in these experiments the trained model captures information in the sequences that is informative of their future values.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Extending Data-Driven Anomaly Detection Methods to Transient Power Conditions in Nuclear Power Plants

Historically, nuclear power plants have operated predominantly at or near full power, meaning that data driven anomaly detection methods can likely perform well at full power operations. This presents a challenge when the power drops (referred to as a transient) and may result in false alarms due to the lack of historical data at those new power levels. The current approach to handling this challenge is to turn detectors off during transients, which makes it impossible to use the algorithms to detect anomalies during these periods, i.e., causing missed detection.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2024

Evaluating a Commercial Dynamic Line Rating Software with the National PMU Dataset

To accelerate the development of data-driven applications for power systems, the Department of Energy (DOE) supported the collection and curation of a synchrophasor dataset spanning two years of observations from transmission utilities across the US. This National PMU Dataset (NPDS) was anonymized and distributed to awardees of a DOE research grant under nondisclosure agreements (NDAs) but has also been retained at PNNL to enable further research. Agreements with data contributors prevent the data from being shared outside the organization. However, establishing a blind research validation methodology is envisioned to maximize the value proposition of the NPDS. In this validation strategy, researchers may share algorithms/software (potentially as executables to protect intellectual property) with PNNL, and PNNL will share feedback about the software’s performance on subsets of the NPDS. Such a blind methodology ensures that sensitive information about critical infrastructure remains protected, but the value of the NPDS can be extended to research beyond PNNL. Through iterative feedback, the algorithms may be tweaked to address real-world artifacts. As the NPDS data is temporally and geographically diverse, it may capture features absent in smaller datasets used during the development of the algorithm under test. This report presents lessons learned from applying the blind validation methodology to LineID™, a synchrophasor-based dynamic line rating software developed by Topolonet Corporation. Improvements made to the software through iterative feedback, limitations of the validation methodology, as well as how the limitations of the NPDS affected the evaluation process are discussed. Observations indicate that the proposed validation methodology can be valuable for evaluating other tools in the future.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Model Development and Analysis of a High-Fidelity Neutron Transport Sensor: The Quadrupole Detector Concept for Measurement of the Neutron Flux Gradient

Accurate reconstruction of the neutron flux distribution within a reactor core is essential for safe and efficient reactor operation. Traditional power shape synthesis in Light Water Reactors relies on hundreds of in-core detectors. However, this approach becomes impractical for Advanced Reactors and Microreactors due to limited space and harsh environments. To address this challenge, we propose a data-driven methodology that combines high-fidelity modeling with real-time ex-core sensor measurements, enabling the reconstruction of core power distribution while minimizing the reliance on intrusive in-core instrumentation. This project began in FY24 and achieved two initial milestones: (1) the definition of a three-year development plan for a Digital Twin framework and (2) the development of high-fidelity neutronics models of the Purdue University Reactor One (PUR-1) using both MCNP6 and OpenMC. The PUR-1 reactor, a zero-power facility, was selected due to its suitability for neutronics-focused modeling and the availability of experimental data for validation. Both models were benchmarked using neutron flux measurements obtained from irradiated gold foils, which were strategically placed within the core during a dedicated campaign in July 2024. This report marks the continuation and completion of those foundational tasks. The OpenMC model has been refined (improved geometric accuracy, expanded cross-section libraries, and refined sampling) and validated using additional experimental data. An updated sensor design—based on quadrupole configuration—was designed to measure both ex-core flux and its spatial gradient. These measurements will serve as inputs to a neural network-based reconstruction algorithm. Finally, the methodology was demonstrated on a two-dimensional test case representative of the heterogeneous material composition of the PUR-1 reactor core. A neural network implementation of the Kirchhoff-Helmholtz integral equation was employed to solve the boundary value problem using peripheral sensor measurements. The preliminary results confirm the strong potential of the proposed approach for accurate and minimally invasive neutron flux reconstruction.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2025

Data Structure Alchemy

In an increasingly more data-driven world, the project set out to uncover the first principles of data-structure design, chart the immense design space they form, and build automation that can synthesize an optimal structure, or even a whole storage engine, for any given workload, hardware platform, and cost target. Data structures are at the center of every computational system and are directly responsible for its performance. Two core technical thrusts were defined: 1) Mapping design spaces for key data-centric abstractions (filters, hash functions, storage-engine layouts, neural-network topologies, blockchain protocols, image layouts, etc.). 2) Developing search & synthesis algorithms, initially analytical cost models, later neural-guided bi-level optimisers that navigate sextillions of candidate designs in seconds and materialise the best one as ready‐to-run code. This report distills the key insights, accomplishments, and impact.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Web-Based Tools for Data-Informed Remedy Optimization: Software Theory and User Guide

This report documents the development and application of two web-based decision-support tools for pump-and-treat (P&T) groundwater remediation systems: PTOLEMY (Pump-and-Treat Optimized Location Evaluation to Maximize Yields) and OPTIMA (Optimization for Pump-and-Treat Implementation, Management, & Assessment). These tools enhance remedy design and management by leveraging advanced computational methods – specifically deep learning and multi-objective optimization – within a user-friendly platform. By integrating data-driven models with established hydrogeological knowledge, PTOLEMY and OPTIMA enable more efficient evaluation of well placement and operational strategies, helping site managers balance multiple remediation objectives under complex conditions. Both tools are implemented as modules within the SOCRATES (Suite Of Comprehensive Rapid Analysis Tools for Environmental Sites) web platform, which provides data access, visualization, and analytics to support remedy optimization across sites in the U.S. Department of Energy Office of Environmental Management complex. PTOLEMY is a rapid screening module designed to identify promising locations for new extraction wells. It employs a multi-channel three-dimensional convolutional neural network (MC3D-CNN) trained on high-fidelity simulation data to predict the relative performance (in terms of contaminant mass recovery) of potential well sites. Through an interactive web interface, PTOLEMY visualizes the probability of high performance across a site, highlighting areas where an extraction well is likely to yield above-threshold contaminant removal over a multi-year period. PTOLEMY’s map-based displays and exportable results support transparent communication of screening analyses. By focusing attention on the most favorable candidate locations, the tool augments traditional engineering judgment and physics-based modeling, providing a data informed basis for subsequent detailed evaluations. OPTIMA is a multi objective optimization module designed to find wellfield layouts and operating schedules that meet various cleanup goals. It quickly evaluates thousands of candidate setups – combinations of well locations, timing, and rates – and returns a small set of best trade-off options for comparison. At its core, OPTIMA uses a U-Net-based surrogate model – a deep-learning emulator of a groundwater flow and transport simulator – to dramatically accelerate scenario evaluations. Coupling this fast surrogate with the NSGA-II (Non-dominated Sorting Genetic Algorithm II) evolutionary algorithm, OPTIMA explores a wide decision space of well locations and schedules to identify Pareto-optimal solutions that trade off key objectives (e.g., minimizing cleanup time, maximizing contaminant mass removal, and minimizing plume extent). The tool outputs a family of optimal configurations and visualizes their trade-offs (Pareto frontiers of cleanup metrics and maps of optimized well placements). Site managers can use these results to understand the range of viable strategies and to select candidate designs for more detailed verification. OPTIMA is currently under active development and not yet fully released; this guide provides early documentation to support planning and gather user feedback.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Toward Drilling the Perfect Geothermal Well: An International Research Coordination Network for Geothermal Drilling Optimization Supported by Deep Machine Learning and Cloud Based Data Aggregation

The EDGE project, supported by the U.S. Department of Energy Geothermal Technologies Office under award DE-EE0008793, established a data-driven framework for improving the efficiency, cost-effectiveness, and reliability of geothermal well drilling. The project focused on developing scalable data infrastructure, advanced machine learning and probabilistic models, and integrated analytics tools to support continuous drilling optimization. A central objective was to reduce geothermal drilling costs by up to seventy percent while minimizing the risk of well failure through predictive diagnostics and adaptive planning. Over the project period, a comprehensive data repository was designed and deployed, incorporating records from over one hundred geothermal wells across varied geological settings. This repository supported both structured and unstructured data and adhered to FAIR data principles, enabling provenance tracking, quality control, and standardized metadata. The project introduced automated ingestion pipelines and a cloud-hosted platform that facilitated access to raw, processed, and derived datasets. This infrastructure served as the foundation for model development and analysis. Machine learning workflows were developed to predict key drilling metrics including rate of penetration, non-productive time, and total drilling costs. Self-organizing maps and dimensionality reduction methods were used to uncover operational patterns and outliers, while supervised learning algorithms such as random forests and deep neural networks were applied to forecast performance outcomes. The models were validated on heterogeneous datasets from both U.S. and Icelandic fields, demonstrating variable but significant predictive accuracy. The results indicated that finer temporal resolution, inclusion of lithological data, and consistency in operational annotations could substantially improve model performance. The project also implemented process mining techniques to reconstruct state-transition models from drilling event logs. These models enabled the identification of deviations from optimal workflows and provided insights into recurring failure modes. Analysis of non-productive time highlighted the impact of equipment failures, geological challenges, and human factors, offering opportunities for targeted mitigation strategies. The EDGE Dashboard was developed as a web-based expert system integrating data visualization, model outputs, and user-driven queries. It provided an accessible interface for operators to explore historical data, evaluate predicted outcomes, and compare drilling scenarios. Initial feedback from project partners suggested that the dashboard could serve as a foundation for more advanced advisory and optimization tools. Overall, the EDGE project demonstrated the feasibility and value of applying modern data science techniques to geothermal drilling. It delivered a set of interoperable tools and models that can support more efficient, lower-risk well development. The findings point toward a viable path for transitioning from advisory analytics to semi-autonomous drilling systems, contingent on continued collaboration, expanded datasets, and field validation. The project results have immediate relevance for drilling operations, data management practices, and future geothermal R&D efforts aimed at achieving reliable, cost-competitive geothermal energy at scale.

15 GEOTHERMAL ENERGY↗
DOE OSTI2025

CRCNS22 Learning Rules in the Hippocampus and their Mapping to Neuromorphic Systems (Final Technical Report)

Large scale biologically-realistic computational models are key to investigating the interplay between structure and function in nervous systems, thus paving the way to new clinical methods and neuro-inspired computing solutions. This project focuses on the hippocampus, in particular the CA3-CA1 regions, due to their role in associative learning and memory, pattern separation and completion, and spatial navigation. Investigations into the neuronal organization and learning rule(s) of this circuit can shed light into how declarative memories are formed, stored, recalled and forgotten and inform computational, experimental and clinical neuroscience work. Our project aims at developing a novel data-driven methodology supported by a broad heterogeneous base of neuroscience experimental knowledge and inspired from advances in computer science and engineering. Specifically, this work will benchmark existing and new learning rules within a full-scale spiking neural network simulation of the CA3-CA1 region. The model will be based on an open-source repository, called the Hippocampome, which contains neuronal morphologies, firing patterns, synapse probabilities, and most other required parameters for all known neuron types in the rodent hippocampal formation. The model will be first trained in a supervised fashion for associative memory tasks using backpropagation through time traditionally used in computer science, enhanced with a new technique called the surrogate gradient method. This optimization method will be used to obtain a global loss minimization, but it is not biologically inspired as it assumes the use of data not locally available to the synapses. However, we propose its use as a benchmarking tool, to compare the training performance of local biologically plausible and hardware-mappable learning rules at scale. New rules or combinations will be proposed and tested as needed, based on the obtained results. Progress in this area will also drive the development of novel hardware-mappable algorithms for continual lifelong learning and categorization of new events from few presented examples. This project goes beyond the existing state-of-the-art by looking at large scale realistic neuronal circuits as networks trainable via global optimization methods such as surrogate gradient descent. The objective function of the brain that supports learning is largely unknown, but it is likely that it operates through local learning rules. Studying network trajectories around local minima as proposed in this work represents a useful strategy for understanding whether a network is training by using a specific (set of) learning rule(s). Starting from a completely untrained network is a challenging test since it is difficult to determine how the learning rule affects the trajectory of the network. This interdisciplinary project will help understand what rule governs learning in these regions or if multiple learning rules are involved. The work will develop a robust methodology to measure if the network is converging to the target solution, oscillating around it, or diverging away.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Collaborative Research: Enabling multi-scale studies of magnetic reconnection with interpretable data-driven models

The development of accurate reduced descriptions and improved closures for magnetic reconnection is an important and a long‐standing challenge in plasma physics. The four‐fluid approach, and associated closures, that were investigated have the potential to improve the accuracy of plasma fluid models, capturing physical effects which would otherwise require a kinetic description. If successful, this approach could have an important impact for the modeling of laboratory and space plasmas. The major goals of this project were to develop new machine learning (ML) tools based on sparse and symbolic regression techniques, and to extract interpretable and generalizable reduced models (e.g., in the form of partial differential equations - PDEs) from data generated by first principles plasma simulations. Preserving interpretability of such data‐driven models is key to addressing the long‐standing theoretical and numerical challenges. Prior proof‐of‐principle studies have demonstrated the enormous potential of this approach, by recovering the well‐established hierarchy of plasma equations (from Vlasov to MHD) from data produced by particle‐in‐cell (PIC) simulations. Our goal in this project was to extend and apply these new tools to construct better kinetic closures for magnetic reconnection; to derive better models of particle injection and acceleration by this fundamental plasma process; and to use this understanding to accelerate the development of multi‐scale plasma algorithms. While our immediate focus was on the problem of magnetic reconnection, the tools that were will developed are general and applicable to other areas of plasma physics, and more broadly to many‐body phenomena. We anticipate that the development of these multi‐scale models will have a significant impact across different areas of plasma science, from fusion to space and astrophysical plasmas.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2023

A Data-Driven Framework for Direct Local Tensile Property Prediction of Laser Powder Bed Fusion Parts

This article proposes a generalizable, data-driven framework for qualifying laser powder bed fusion additively manufactured parts using part-specific in situ data, including powder bed imaging, machine health sensors, and laser scan paths. To achieve part qualification without relying solely on statistical processes or feedstock control, a sequence of machine learning models was trained on 6299 tensile specimens to locally predict the tensile properties of stainless-steel parts based on fused multi-modal in situ sensor data and a priori information. A cyberphysical infrastructure enabled the robust spatial tracking of individual specimens, and computer vision techniques registered the ground truth tensile measurements to the in situ data. The co-registered 230 GB dataset used in this work has been publicly released and is available as a set of HDF5 files. The extensive training data requirements and wide range of size scales were addressed by combining deep learning, machine learning, and feature engineering algorithms in a relay. The trained models demonstrated a 61% error reduction in ultimate tensile strength predictions relative to estimates made without any in situ information. Lessons learned and potential improvements to the sensors and mechanical testing procedure are discussed.

36 MATERIALS SCIENCE↗
DOE OSTI2023

Adaptively Learned Modeling for a Digital Twin of Hydropower Turbines with Application to a Pilot Testing System

In the development of a digital twin (DT) for hydropower turbines, dynamic modeling of the system (e.g., penstock, turbine, speed control) is crucial, along with all the necessary data interface, virtualization, and dashboard designs. Since the DT must mimic the actual dynamics of the hydropower turbine accurately, adaptive learning is required to train these dynamic models online so that the models in the DT can effectively follow the representation of the actual hydropower turbine dynamics accurately and reliably. This study presents an adaptive learning method for obtaining the hydropower turbine models for DT development of hydropower systems using the recursive least squares algorithm. To simplify the formulation, the hydropower turbine under consideration was assumed to operate near a fixed operating point, where the system dynamics can be well represented by a set of linear differential equations with constant parameters. In this context, the well-known six-coefficient model for the Francis turbine was formulated as the starting point to obtain input and output models for the turbine. Then, an adaptive learning mechanism was developed to learn model parameters using real-time data from a hydropower turbine testing system. This led to semi-physical modeling, in which first principles and data-driven modeling are integrated to produce dynamic models for DT development. Applications to a pilot system at the Norwegian University of Science and Technology (NTNU) were made, and the models learned adaptively using the data collected from the university’s pilot system. Desired modeling and validation results were obtained.

13 HYDRO ENERGY↗
DOE OSTI2023

QUOTAS: A New Research Platform for the Data-driven Discovery of Black Holes

We present QUOTAS, a novel research platform for the data-driven investigation of supermassive black hole (SMBH) populations. While SMBH data—observations and simulations—have grown in complexity and abundance, our computational environments and tools have not matured commensurately to exhaust opportunities for discovery. To explore the BH, host galaxy, and parent dark matter halo connection—in this pilot version—we assemble and colocate the high-redshift, z > 3 quasar population alongside simulated data at the same cosmic epochs. As a first demonstration of the utility of QUOTAS, we investigate correlations between observed Sloan Digital Sky Survey (SDSS) quasars and their hosts with those derived from simulations. Leveraging machine-learning algorithms (ML), to expand simulation volumes, we show that halo properties extracted from smaller dark-matter-only simulation boxes successfully replicate halo populations in larger boxes. Next, using the Illustris-TNG300 simulation that includes baryonic physics as the training set, we populate the larger LEGACY Expanse dark-matter-only box with quasars, and show that observed SDSS quasar occupation statistics are accurately replicated. First science results from QUOTAS comparing colocated observational and ML-trained simulated data at z3 are presented. QUOTAS demonstrates the power of ML, in analyzing and exploring large data sets, while also offering a unique opportunity to interrogate theoretical assumptions that underpin accretion and feedback models. QUOTAS and all related materials are publicly available at the Google Kaggle platform. (The full data set—observational data and simulation data—are available at: https://www.kaggle.com/ and the codes are available at:https://www.kaggle.com/datasets/quotasplatform/quotas)

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2026

Leveraging machine learning to enhance aerosol classification using Single-Particle Mass Spectrometry

Advancing automated classification of atmospheric aerosols from Single-Particle Mass Spectrometry (SPMS) data remains challenging due to overlapping ion signatures, compositional diversity, and limited labeled data. This study evaluates supervised and semi-supervised learning frameworks to enhance aerosol identification by jointly leveraging labeled and unlabeled spectra. Four models were compared: a supervised Support Vector Machine (SVM), a self-training SVM, a stacked autoencoder classifier, and a stacked autoencoder trained using a temporal-ensembling Mean Teacher approach. All models achieved high and stable accuracies (90.0 %–91.1 %), surpassing previous results on the same dataset (87 %) and matching the performance of state-of-the-art deep learning methods. Despite small global metric differences (≤ 1 %), semi-supervised variants yielded up to 5 %–10 % improvements for compositionally rare particle types – such as soot (0.77 % of spectra, F1-score: 0.93–0.97) and hazelnut pollen (0.98 % of spectra, F1-score: 0.97–1.00) – equating to roughly ∼ 187 additional correctly classified spectra. These gains are scientifically significant, as such rare particles exert disproportionate influence on radiative absorption and ice nucleation processes; their improved detection reduces modeled uncertainties in aerosol absorption optical depth and mixed-phase cloud ice nucleation rates. The models' residual misclassifications (≈ 9 %) largely arise from true spectral overlap among chemically adjacent species (e.g., Na- vs. K-feldspar, coated vs. uncoated feldspars), reflecting physical compositional continuity rather than algorithmic error. Collectively, these findings demonstrate that leveraging unlabeled data to learn robust spectral representations and refine classification enhances both fidelity and interpretability, bridging data-driven analysis with aerosol–climate process understanding.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Adaptable Data Driven Model Predictive Control for Heat Pipe Microreactors

To establish a technical basis for self-regulating microreactors, a model predictive control (MPC) system is investigated to proactively respond to anomalies and disturbances in anticipation of potential deviations from operating setpoints. Due to the difficulty of developing a physics-based surrogate model that can accurately match plant data in various operating conditions, machine learning algorithms are used in MPC, which allow for learning from both simulation and operation data, thus efficiently describing the targeted transient with arbitrary accuracy. However, one of the biggest concerns in applying ML algorithms like artificial neural networks (ANNs) is that the predictive capabilities of ANN are limited by training data. If there are gaps between the training and target domain, the accuracy of an ANN can degrade significantly when it is used to predict unseen data. To improve the predictive capability of ANN and enable a confident use of data-driven MPCs outside the training data, this study proposes an adaptive data-driven MPC framework. The system will monitor the discrepancy between plant responses and surrogate predictions, fine-tune the ANN-based surrogate when a large discrepancy is detected, and continue MPC operation with updated surrogates. The framework is demonstrated on a point kinetic model for microreactors. The hyperparameters of the update strategy, including layers to update, error thresholds, learning rate discount, and number of data points used for fine-tuning, are optimized so the simulated microreactor is able to follow changes in setpoint with the smallest of deviations.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗
DOE OSTI2024

Extreme Risk Mitigation in Reinforcement Learning using Extreme Value Theory

Risk-sensitive reinforcement learning (RL) has garnered significant attention in recent years due to the growing interest in deploying RL agents in real-world scenarios. A critical aspect of risk awareness involves modelling highly rare risk events (rewards) that could potentially lead to catastrophic outcomes. These infrequent occurrences present a formidable challenge for data-driven methods aiming to capture such risky events accurately. While risk-aware RL techniques do exist, they suffer from high variance estimation due to the inherent data scarcity. Our work proposes to enhance the resilience of RL agents when faced with very rare and risky events by focusing on refining the predictions of the extreme values predicted by the state-action value distribution. To achieve this, we formulate the extreme values of the state-action value function distribution as parameterized distributions, drawing inspiration from the principles of extreme value theory (EVT). We propose an extreme value theory based actor-critic approach, namely, Extreme Valued Actor-Critic (EVAC) which effectively addresses the issue of infrequent occurrence by leveraging EVT-based parameterization. Importantly, we theoretically demonstrate the advantages of employing these parameterized distributions in contrast to other risk-averse algorithms. Our evaluations show that the proposed method outperforms other risk averse RL algorithms on a diverse range of benchmark tasks, each encompassing distinct risk scenarios.

Wang, Yu↗
DOE OSTI2024

A Distributed Model Identification Algorithm for Multi-Agent Systems: Preprint

In this study, we investigate agent-based approach for system model identification with emphasis on power distribution system applications. Departing from conventional practices of relying on historical data for offline model identification, we adopt online update approach utilizing real-time data by employing the latest data points for gradient computation. This methodology offers advantages including a large reduction in the communication network's bandwidth requirements by minimizing the data exchanged at each iteration and enabling the model to adapt in real-time to disturbances. Furthermore, we extend our model identification process from linear frameworks to more complex non-linear convex models. This extension is validated through numerical studies demonstrating improved control performance for a synthetic IEEE test case.

data-driven control↗