Comparing Deep Neural Network Architectures using Transfer-Learning to Predict Cognitive States
Explore the source record for details and available documents.
machine learning↗
SEARCH · Engineering Papers
Results for “state machine”
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.
Distributed heterogeneous compute infrastructure for the study of additive manufacturing systems
We present the current status of a large-scale computing framework to address the need of the multidisciplinary effort to study chemical dynamics. Specifically, we are enabling scientists to process and store experimental data, run large-scale computationally expensive high-fidelity physical simulation, and analyze these results using the state-of-the-art data analytics tools, machine learning, and uncertainty quantification methods using heterogeneous computing resources, such as CPU and GPU cluster. The framework can integrate or abstract out multiple domains based on roles. In order to develop this framework, we have leveraged an existing framework coupled with in-house heterogeneous computing resources. We present the results of using this framework on a single metadata triggered workflow to accelerate an additive manufacturing use case.
heterogeneous computing, hpc, workflows, reproduci↗
Fundamental limit of jet tagging
Identifying the origin of high-energy hadronic jets (jet tagging) has been a critical benchmark problem for machine learning in particle physics. Jets are ubiquitous at colliders and are complex objects that serve as prototypical examples of collections of particles to be categorized. Over the last decade, machine learning-based classifiers have replaced classical observables as the state of the art in jet tagging. Increasingly complex machine learning models are leading to increasingly more effective tagger performance. Our goal is to address the question of convergence—are we getting close to the fundamental limit on jet tagging or is there still potential for computational, statistical, and physical insights for further improvements? We address this question using state-of-the-art generative models to create a realistic, synthetic dataset with a known jet tagging optimum. Various state-of-the-art taggers are deployed on this dataset, showing that there is a significant gap between their performance and the optimum. Our dataset and software are made public to provide a benchmark task for future developments in jet tagging and other areas of particle physics.
Artificial intelligence↗
Quantum circuit fidelity estimation using machine learning
The computational power of real-world quantum computers is limited by errors. When using quantum computers to perform algorithms which cannot be efficiently simulated classically, it is important to quantify the accuracy with which the computation has been performed. In this work, we introduce a machine learning-based technique to estimate the fidelity between the state produced by a noisy quantum circuit and the target state corresponding to ideal noise-free computation. Our machine learning model is trained in a supervised manner, using smaller or simpler circuits for which the fidelity can be estimated using other techniques like direct fidelity estimation and quantum state tomography. Here we demonstrate that, for simulated random quantum circuits with a realistic noise model, the trained model can predict the fidelities of more complicated circuits for which such methods are infeasible. In particular, we show that the trained model may make predictions for circuits with higher degrees of entanglement than were available in the training set and that the model may make predictions for non-Clifford circuits even when the training set included only Clifford-reducible circuits. This empirical demonstration suggests classical machine learning may be useful for making predictions about beyond-classical quantum circuits for some non-trivial problems.
97 MATHEMATICS AND COMPUTING↗
Multiport high-pressure synchrotron x-ray microscopy cell (CRADA Final Report)
The LBNL team developed a multiport cell allowing for fluid flow, with electrical contacts for sample bias, and with reference electrodes. The cell is currently installed and operating in beam-line 11.0.2 of the ALS, the Berkeley Synchrotron. The cell is closed by a membrane consisting of several stacked layers of graphene (1 to 3). The graphene membrane is supported on a 100 nm thick Si3N4 membrane covered with a 20 nm thin gold film for improved adhesion of the graphene and to ensure electrical conductivity. The Si3N4 film is perforated with a periodic array of holes 1 to 2 micrometers diameter. The cell can withhold an internal pressure (gas or liquid) of up to 10 atmospheres for in situ/operando studies using X-ray Photoelectron emission (XPS), and X-ray Absorption Spectroscopy (XAS). It is also ideal for tip-enhanced nano-Infrared Spectroscopy studies of molecules at the interface between the graphene and the liquid or gas filling the cell. Materials studied in the LBNL laboratory are in the form of nanoparticles, or thin films (few nm) deposited on the side of the graphene membrane facing the solution or gas inside the cell.
36 MATERIALS SCIENCE↗
Applying Machine Learning to the Classification of DC-DC Converters. NA-22 Final Report
Tools are now available that enable measurement electromagnetic radiation (EMR) from active electronics in an item. This radiation may be intended WIFI or cellular network links, for example or unintended such as the switching noise generated by DC-to-DC converters. It would be extremely valuable to have the capability to discriminate between the low-voltage DC-to-DC converters or other digital noise prevalent in most modern electronics, versus the high-voltage DC-to-DC converters used in utility firesets. Previous work performed under a Sandia Laboratory Directed Research and Development (LDRD) project on Charge State Detection using a deep neural network has been continued in this effort. A state-of-the-art supervised machine learning algorithm has not only been extended to discriminate between low and high voltage converters but has been validated in determining a converters make and model.
42 ENGINEERING↗
An approach to verification and validation of a reliable multicasting protocol
This paper describes the process of implementing a complex communications protocol that provides reliable delivery of data in multicast-capable, packet-switching telecommunication networks. The protocol, called the Reliable Multicasting Protocol (RMP), was developed incrementally using a combination of formal and informal techniques in an attempt to ensure the correctness of its implementation. Our development process involved three concurrent activities: (1) the initial construction and incremental enhancement of a formal state model of the protocol machine; (2) the initial coding and incremental enhancement of the implementation; and (3) model-based testing of iterative implementations of the protocol. These activities were carried out by two separate teams: a design team and a V&V team. The design team built the first version of RMP with limited functionality to handle only nominal requirements of data delivery. In a series of iterative steps, the design team added new functionality to the implementation while the V&V team kept the state model in fidelity with the implementation. This was done by generating test cases based on suspected errant or off-nominal behaviors predicted by the current model. If the execution of a test was different between the model and implementation, then the differences helped identify inconsistencies between the model and implementation. The dialogue between both teams drove the co-evolution of the model and implementation. Testing served as the vehicle for keeping the model and implementation in fidelity with each other. This paper describes (1) our experiences in developing our process model; and (2) three example problems found during the development of RMP.
Callahan, John R.↗
An Approach to Verification and Validation of a Reliable Multicasting Protocol
This paper describes the process of implementing a complex communications protocol that provides reliable delivery of data in multicast-capable, packet-switching telecommunication networks. The protocol, called the Reliable Multicasting Protocol (RMP), was developed incrementally using a combination of formal and informal techniques in an attempt to ensure the correctness of its implementation. Our development process involved three concurrent activities: (1) the initial construction and incremental enhancement of a formal state model of the protocol machine; (2) the initial coding and incremental enhancement of the implementation; and (3) model-based testing of iterative implementations of the protocol. These activities were carried out by two separate teams: a design team and a V&V team. The design team built the first version of RMP with limited functionality to handle only nominal requirements of data delivery. In a series of iterative steps, the design team added new functionality to the implementation while the V&V team kept the state model in fidelity with the implementation. This was done by generating test cases based on suspected errant or offnominal behaviors predicted by the current model. If the execution of a test was different between the model and implementation, then the differences helped identify inconsistencies between the model and implementation. The dialogue between both teams drove the co-evolution of the model and implementation. Testing served as the vehicle for keeping the model and implementation in fidelity with each other. This paper describes (1) our experiences in developing our process model; and (2) three example problems found during the development of RMP.
Callahan, John R.↗
High-Performance Semiempirical Excited-State Molecular Dynamics Powered by Graphics Processing Units
Here, this Letter introduces excited-state molecular dynamics in PYSEQM, a GPU-accelerated semiempirical quantum chemistry engine implemented in PyTorch. The new module enables Born–Oppenheimer molecular dynamics (BOMD) using configuration-interaction singles and random phase approximation for excited states, allowing long trajectories and large statistical ensembles to be simulated efficiently on a single GPU. We also implement an extended Lagrangian excited-state BOMD (XL-ESMD) scheme that propagates auxiliary electronic variables, enabling relaxed ground and excited-state convergence thresholds without compromising energy conservation. The excited-state BOMD implementation scales smoothly from small chromophores to a nearly 900-atom dendrimer (taking 6.5 s per MD step). PYSEQM also supports batched execution, allowing many geometries or trajectories to be evaluated in a single GPU launch, substantially increasing throughput and making ensemble-based protocols routine. As a demonstration, we compute absorption, emission, and infrared spectra from trajectories propagated on the ground and first excited states. The XL-ESMD scheme yields identical spectra at significantly lower computational cost, establishing the role of extended Lagrangian based dynamics for efficient excited-state BOMD simulations. Beyond raw performance, PYSEQM’s PyTorch foundation provides automatic differentiation for forces, efficient GPU batching, and seamless interfacing with machine learning models. These capabilities position PYSEQM as a practical platform for machine learning-augmented excited-state dynamics and lay the foundation for future data-driven nonadiabatic excited-state dynamics modeling of ultrafast spectroscopic probes.
37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
Applications of neural networks to the studies of phase transitions of two-dimensional Potts models
We study the phase transitions of two-dimensional (2D) Q-states Potts models on the square lattice, using the first principles Monte Carlo (MC) simulations as well as the techniques of neural networks (NN). We demonstrate that the ideas from NN can be adopted to study these considered phase transitions efficiently. In particular, even with a simple NN constructed in this investigation, we are able to obtain the relevant information of the nature of these phase transitions, namely whether they are first order or second order. Our results strengthen the potential applicability of machine learning in studying various states of matters. Subtlety of applying NN techniques to investigate many-body systems is briefly discussed as well.
75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
Advances in imaging, scattering, spectroscopy, and machine learning-aided approaches for multiscale characterization of cementitious systems
Recent progress in methods used in the multiscale characterization of cementitious systems is reviewed, focusing on advances in imaging, scattering, and spectroscopy. The review includes relevant applications and developments in machine learning and other data analytics approaches to enhance characterization. Developments in imaging using light and electron microscopy as well as x-ray (i.e., from synchrotron) methods are summarized. Updates include scanning electron microscopy (SEM), transmission electron microscopy (TEM), tomography, and holography. A critical overview of spectroscopy (e.g., MAS NMR, Raman) and scattering (e.g., neutron, x-ray, synchrotron x-ray) methods is provided, and the intersection of these with imaging is developed (e.g., Raman imaging). Additionally, the paper summarizes recent developments in and implementations of state-of-the-art machine-learning algorithms and data analytics methods for automated, systematic, and/or quantitative analyses of image data sets. The review considers but is not limited to the application of these methods for investigating the hydration and microstructure development of cement phases, low-carbon-footprint cements (e.g., limestone calcined clay cements, LC3), environmental interactions (e.g., ASR) and model systems. Finally, the present work provides a critical presentation of advances in characterization methods that link together the composition and multiscale structure of cementitious materials.
42 ENGINEERING↗
Power System Frequency Dynamics Modeling, State Estimation, and Control using Neural Ordinary Differential Equations (NODEs) and Soft Actor-Critic (SAC) Machine Learning Approaches
With the global energy transition of the electric power system, grid control, supervision, and protection is becoming more challenging. With the increasing integration of renewable energy sources (RES), the system dynamics are changing, causing traditional power system dynamic modeling with swing equation-based modeling approaches to fail. Additionally, the converter-dominated power grid is decreasing the system inertia, making the power system more fragile to the frequency swings. This paper first investigates and compares the application of a model-based Kalman filter state estimation approach with (i) a model-free machine learning approach --- neural ordinary differential equations (NODEs) --- and (ii) a data-driven system identification (SysId) approach to model and infer critical state values of the power system frequency dynamics. Then a model predictive control (MPC) framework is compared to a model-free Soft Actor-Critic (SAC) reinforcement learning (RL) control algorithm in providing efficient fast frequency response (FFR) to the power system frequency dynamics. The approaches are compared in terms of their performance goals as well as their per-timestep computational efficiency. Furthermore, the comparative study for state estimation shows that for the model-free requirement, both NODEs and SysId can provide accurate state estimates; however, with increasing model complexity, NODEs can be a better choice for model identification. Similarly, the results from the FFR comparative study show that the SAC RL-based FFR, once trained, outperforms MPC with better control signals and faster computation time, making the SAC RL-based FFR better option for providing FFR to the power system.
97 MATHEMATICS AND COMPUTING↗
Mechanical Purcell filters for microwave quantum machines
In circuit quantum electrodynamics, measuring the state of a superconducting qubit introduces a loss channel, which can enhance spontaneous emission through the Purcell effect, thus decreasing the qubit lifetime. This decay can be mitigated by performing the measurement through a Purcell filter, which strongly suppresses signal propagation at the qubit transition frequency. If the filter is also well-matched at the readout cavity frequency, it will protect the qubit from decoherence channels without sacrificing measurement bandwidth. In this work, we propose and analyze design for a mechanical Purcell filter, which we also fabricate and characterize at room temperature. The filter is composed of an array of nanomechanical resonators in thin-film lithium niobate, connected in a ladder topology, with series and parallel resonances arranged to produce a bandpass response. Their modest footprint, steep band edges, and lack of cross talk make these filters an appealing alternative to analogous electromagnetic versions currently used in microwave quantum machines.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
Probabilistic machine learning for battery health diagnostics and prognostics—review and perspectives
Abstract Diagnosing lithium-ion battery health and predicting future degradation is essential for driving design improvements in the laboratory and ensuring safe and reliable operation over a product’s expected lifetime. However, accurate battery health diagnostics and prognostics is challenging due to the unavoidable influence of cell-to-cell manufacturing variability and time-varying operating circumstances experienced in the field. Machine learning approaches informed by simulation, experiment, and field data show enormous promise to predict the evolution of battery health with use; however, until recently, the research community has focused on deterministic modeling methods, largely ignoring the cell-to-cell performance and aging variability inherent to all batteries. To truly make informed decisions regarding battery design in the lab or control strategies for the field, it is critical to characterize the uncertainty in a model’s predictions. After providing an overview of lithium-ion battery degradation, this paper reviews the current state-of-the-art probabilistic machine learning models for health diagnostics and prognostics. Details of the various methods, their advantages, and limitations are discussed in detail with a primary focus on probabilistic machine learning and uncertainty quantification. Last, future trends and opportunities for research and development are discussed.
25 ENERGY STORAGE↗
An approach to verification and validation of a reliable multicasting protocol: Extended Abstract
This paper describes the process of implementing a complex communications protocol that provides reliable delivery of data in multicast-capable, packet-switching telecommunication networks. The protocol, called the Reliable Multicasting Protocol (RMP), was developed incrementally using a combination of formal and informal techniques in an attempt to ensure the correctness of its implementation. Our development process involved three concurrent activities: (1) the initial construction and incremental enhancement of a formal state model of the protocol machine; (2) the initial coding and incremental enhancement of the implementation; and (3) model-based testing of iterative implementations of the protocol. These activities were carried out by two separate teams: a design team and a V&V team. The design team built the first version of RMP with limited functionality to handle only nominal requirements of data delivery. This initial version did not handle off-nominal cases such as network partitions or site failures. Meanwhile, the V&V team concurrently developed a formal model of the requirements using a variant of SCR-based state tables. Based on these requirements tables, the V&V team developed test cases to exercise the implementation. In a series of iterative steps, the design team added new functionality to the implementation while the V&V team kept the state model in fidelity with the implementation. This was done by generating test cases based on suspected errant or off-nominal behaviors predicted by the current model. If the execution of a test in the model and implementation agreed, then the test either found a potential problem or verified a required behavior. However, if the execution of a test was different in the model and implementation, then the differences helped identify inconsistencies between the model and implementation. In either case, the dialogue between both teams drove the co-evolution of the model and implementation. We have found that this interactive, iterative approach to development allows software designers to focus on delivery of nominal functionality while the V&V team can focus on analysis of off nominal cases. Testing serves as the vehicle for keeping the model and implementation in fidelity with each other. This paper describes (1) our experiences in developing our process model; and (2) three example problems found during the development of RMP. Although RMP has provided our research effort with a rich set of test cases, it also has practical applications within NASA. For example, RMP is being considered for use in the NASA EOSDIS project due to its significant performance benefits in applications that need to replicate large amounts of data to many network sites.
Callahan, John R.↗
Predicting the Functional State of Protein Kinases Using Interpretable Graph Neural Networks
Kinases are a family of proteins that function as molecular switches, regulating several essential cellular activities such as cell proliferation. Dysfunctional kinases are implicated in several types of cancers and hence they are actively pursued as drug targets. Given the vast number of complex kinase structures that are available in the protein data bank (PDB), there is a necessity to develop methodologies that can identify structurally important moieties of the kinases in an automated fashion, for such techniques can be instrumental in identifying novel drug targets. In this work, we develop a graph neural network (GNN) based deep learning framework for classifying the functionally active and inactive states of a large set of eukaryotic protein kinases, making use of their 3D structure from the PDB. We show that GNN based machine learning models can classify protein states with an accuracy greater than 97%. We further use the GNN models to automatically identify regions of the kinases that are important for its function. For this purpose, Gradient-weighted Class Activation Mapping (Grad-CAM) was implemented on the protein graphs. Remarkably, Grad-CAM consistently identifies the highly conserved DFG motif as the most important part of the protein across the entire kinome, without any prior input. Other regions of the hydrophobic core such as the HRD motif were also identified by the interpretable GNN framework, consistent with the literature. We discuss the significance of each of these regions in detail.
Ashwin Ravichandran↗
Machine-Learning-Based Adaptive Thinning of CrIS Radiances to Improve Global Tropical Cyclone Analysis and Forecasts
This work is focused on optimizing the assimilation of hyperspectral infrared (IR) radiances from the Cross-track Infrared Sounder (CrIS) with the goal of improving the representation of tropical cyclones (TCs) in global analyses and forecasts. Current operational assimilation systems rely on subsampling IR radiances on a regular thinning grid. A new and improved adaptive methodology based on machine learning (ML) recognizes TCs from geostationary satellite imagery and is implemented in the Goddard Earth Observing System (GEOS) model and data assimilation framework. The ML methodology is extensively trained on existing TC data sets and creates for each TC a dynamic mask, based on the evolving shape and life cycle of that specific event. Once a TC mask is created, a switch is then activated in the data assimilation system to alter the thinning, ingesting more CrIS radiances within the moving mask, thus increasing the TC sampling. After the TC dissipates, the assimilation of CrIS radiances reverts to normal data density. Results of TC segmentation provided by a state-of-the-art generative machine learning model known as the Denoising Diffusion Probabilistic Model (DDPM) are compared to the previously used U-Net model. The new approach surpasses the performance of the previously developed one. The methodology is applied to both clear-sky and cloud-cleared radiances. Benefits from the latter methodology, particularly in improving the structure of TCs and the intensity forecasts, are presented.
Oreste Reale↗
Search for CP violation in t$\overline{\textrm{t}}$H and tH production in multilepton channels in proton-proton collisions at $\sqrt{s}$ = 13 TeV
The charge-parity (CP) structure of the Yukawa interaction between the Higgs (H) boson and the top quark is measured in a data sample enriched in the t$\overline{t}$H and tH associated production, using 138 fb -1 of data collected in proton-proton collisions at $\sqrt{s}$ = 13 TeV by the CMS experiment at the CERN LHC. The study targets events where the H boson decays via H → WW or H → ττ and the top quarks decay via t → Wb: the W bosons decay either leptonically or hadronically, and final states characterized by the presence of at least two leptons are studied. Machine learning techniques are applied to these final states to enhance the separation of CP -even from CP -odd scenarios. Two-dimensional confidence regions are set on $κ$ t and $\widetilde{k}$t, which are respectively defined as the CP -even and CP -odd top-Higgs Yukawa coupling modifiers. No significant fractional CP -odd contributions, parameterized by the quantity |$f^{Htt}_{CP}$| are observed; the parameter is determined to be |$f^{Htt}_{CP}$| = 0.59 with an interval of (0.24, 0.81) at 68% confidence level. The results are combined with previous results covering the H → ZZ and H → γγ decay modes, yielding two- and one-dimensional confidence regions on $κ$ t and $\widetilde{k}$t, while |$f^{Htt}_{CP}$| is determined to be |$f^{Htt}_{CP}$| = 0.28 with an interval of |$f^{Htt}_{CP}$| < 0.55 at 68% confidence level, in agreement with the standard model CP -even prediction of |$f^{Htt}_{CP}$| = 0.
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗