Machine-Learned, Finite Temperature Fermi-Operator Expansions Suitable for GPUs and AI-Hardware
Explore the source record for details and available documents.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
SEARCH · Engineering Papers
Results for “Machine Learning Hardware”
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.
A Quantum-Assisted Algorithm for Sampling Applications in Machine Learning
An increase in the efficiency of sampling from Boltzmann distributions would have a significant impact in deep learning and other machine learning applications. Recently, quantum annealers have been proposed as a potential candidate to speed up this task, but several limitations still bar these state-of-the-art technologies from being used effectively. One of the main limitations is that, while the device may indeed sample from a Boltzmann-like distribution, quantum dynamical arguments suggests it will do so with an instance-dependent effective temperature, different from the physical temperature of the device. Unless this unknown temperature can be unveiled, it might not be possible to effectively use a quantum annealer for Boltzmann sampling. In this talk, we present a strategy to overcome this challenge with a simple effective-temperature estimation algorithm. We provide a systematic study assessing the impact of the effective temperatures in the learning of a kind of restricted Boltzmann machine embedded on quantum hardware, which can serve as a building block for deep learning architectures. We also provide a comparison to k-step contrastive divergence (CD-k) with k up to 100. Although assuming a suitable fixed effective temperature also allows to outperform one step contrastive divergence (CD-1), only when using an instance-dependent effective temperature we find a performance close to that of CD-100 for the case studied here. We discuss generalizations of the algorithm to other more expressive generative models, beyond restricted Boltzmann machines.
Perdomo-Ortiz, Alejandro↗
Comparison of CNN-Based Image Classification Approaches for Implementation of Low-Cost Multispectral Arcing Detection
Camera-based sensing has benefited in recent years from developments in machine learning data processing methods, as well as improved data collection options such as Unmanned Aerial Vehicles (UAV) mounted sensors. However, cost considerations, both for the initial purchase of sensors as well as updates, maintenance, or potential replacement if damaged, can limit adoption of more expensive sensing options for some applications. To evaluate more affordable options with less expensive, more available, and more easily replaceable hardware, we examine the use of machine learning-based image classification with custom datasets, utilizing deep learning based-image classification and the use of ensemble models for sensor fusion. Utilizing the same models for each camera to reduce technical overhead, we showed that for a very representative training dataset, camera-based detection can be successful for detection of electrical arcing. We also use multiple validation datasets, based on conditions expected to be of varying difficulty, to evaluate custom data. These results show that ensemble models of different data sources can mitigate risks from gaps in training data, though the system will be less redundant for those cases unless other precautions are taken. We found that with good quality custom datasets, data fusion models can be utilized without specialization in design to the specific cameras utilized, allowing for less specialized, more accessible equipment to be utilized as multispectral camera components. This approach can provide an alternative to expensive sensing equipment for applications in which lower-cost or more easily replaceable sensing equipment is desirable.
convolutional neural networks↗
wa-hls4ml: A Benchmark and Surrogate Models for hls4ml Resource and Latency Estimation
As machine learning (ML) is increasingly implemented in hardware to address real-time challenges in scientific applications, the development of advanced toolchains has significantly reduced the time required to iterate on various designs. These advancements have solved major obstacles, but also exposed new challenges. For example, processes that were not previously considered bottlenecks, such as hardware synthesis, are becoming limiting factors in the rapid iteration of designs. To mitigate these emerging constraints, multiple efforts have been undertaken to develop an ML-based surrogate model that estimates resource usage of ML accelerator architectures. We introduce wa-hls4ml, a benchmark for ML accelerator resource and latency estimation, and its corresponding initial dataset of over 680,000 fully connected and convolutional neural networks, all synthesized using hls4ml and targeting Xilinx FPGAs. The benchmark evaluates the performance of resource and latency predictors against several common ML model architectures, primarily originating from scientific domains, as exemplar models, and the average performance across a subset of the dataset. Additionally, we introduce GNN- and transformer-based surrogate models that predict latency and resources for ML accelerators. We present the architecture and performance of the models and find that the models generally predict latency and resources for the 75% percentile within several percent of the synthesized resources on the synthetic test dataset.
Hawks, Benjamin [Fermilab] (ORCID:0000000157000288↗
wa-hls4ml: A Benchmark and Surrogate Models for hls4ml Resource and Latency Estimation
As machine learning (ML) is increasingly implemented in hardware to address real-time challenges in scientific applications, the development of advanced toolchains has significantly reduced the time required to iterate on various designs. These advancements have solved major obstacles, but also exposed new challenges. For example, processes that were not previously considered bottlenecks, such as hardware synthesis, are becoming limiting factors in the rapid iteration of designs. To mitigate these emerging constraints, multipleefforts have been undertaken to develop an ML-based surrogate model that estimates resource usage of synthesized ML accelerator architectures. We introduce wa-hls4ml, a benchmark for ML accelerator resource and latency estimation, and its corresponding initial dataset of over 680 000 fully connected and convolutional neural networks, all synthesized using hls4ml and targeting Xilinx FPGAs. The benchmark evaluates the performance of resource and latency predictors against several common ML model architectures, primarily originating from scientific domains, as exemplar models, and the average performance across a subset of the dataset. Additionally, we introduce GNN- and transformer-based surrogate models that predict latency and resources for ML accelerators. We present the architecture and performance of the models and find that the models generally predict latency and resources for the 75% percentile within several percent of the synthesized resources on the synthetic test dataset.
Hawks, Benjamin G. [Fermilab]↗
ICE-RASSOR: Intelligent Capabilities Enhanced
NASA’s Regolith Advanced Surface Systems Operations Robot (RASSOR) is principally designed to mine and deliver regolith for In-Situ Resource Utilization (ISRU) processing. RAS-SOR’s design enables it to efficiently collect and deposit regolith, return collected material for processing, and myriad related ISRU activities. To reliably perform these operations on the lunar sur-face, RASSOR software and sensory systems need to be robust and maximize the information extracted from on-board sensing. Herein, we present preliminary findings from the Intelligent Capabilities Enhanced RASSOR project. We apply supervised learning using real data to estimate the soil mass collected without the need for mass flow rate monitors or other explicate sensing techniques. We also create a reduced-order simulation environment to develop autonomous trenching controllers via reinforcement learning and proto-type state estimation architectures. Our initial results suggest that excavated regolith mass can be inferred within 2.9% RMS error of full scale, and reinforcement learning for autonomous operations has learned viable trenching strategies and helped identify desirable sensing capabilities, arrangements, and considerations. Future work includes regolith mass estimation during dynamic operation, expanding our simulation to more complex environments, and transfer learning from simulation to hardware.
machine learning↗
Optimization-Based Parametric Design via High-Fidelity Simulation: Overview + Examples
Design-Build-Test approaches for developing spaceflight hardware are prohibitively time and cost intensive and often lead to suboptimal mechanism designs. Approaches that couple machine learning and high-fidelity physics simulation could eliminate the need for hardware prototyping and dramatically accelerate the engineering design cycle, ultimately reducing cost. This talk presents a modular NASA-developed toolchain to optimize hardware mechanisms in a virtual environment using numerical optimization and multi-body physics simulation and includes example applications related to rigid wheel design for autonomous rovers and computational fluid dynamics.
optimization↗
Parametric Mechanism Design Through Numerical Optimization and Physics Simulation
Design-Build-Test approaches for developing spaceflight hardware are prohibitively time and cost intensive and often lead to suboptimal mechanism designs. Approaches that couple machine learning and high-fidelity physics simulation could eliminate the need for hardware prototyping and dramatically accelerate the engineering design cycle, ultimately reducing cost. This work presents a modular NASA-developed toolchain to optimize hardware mechanisms in a virtual environment using numerical optimization and multi-body physics simulation. The toolchain enables multi-objective optimization, generates parametric CAD files that can be further post-processed by an end user, and can be expanded to optimize full systems and non-mechanical parameters such as feedback control variables. We demonstrate the toolchain through an independently verifiable design problem that optimizes wheel radius to achieve a desired linear velocity in a rigid-body physics environment when the wheel rotates at a constant angular speed, and then post-process the parametric CAD file of the optimal design generated by the tool before ultimately manufacturing it via 3D printing. We end with a discussion of how the toolchain can incorporate other analysis tools, including finite element analysis, computational fluid dynamics, and granular media simulations.
Optimization↗
An Optimization-Based Toolchain for Parametric Mechanism Design
Design-Build-Test approaches for developing spaceflight hardware are prohibitively time and cost intensive and often lead to suboptimal mechanism designs. Approaches that couple machine learning and high-fidelity physics simulation could eliminate the need for hardware prototyping and dramatically accelerate the engineering design cycle, ultimately reducing cost. This work presents a modular NASA-developed toolchain to optimize hardware mechanisms in a virtual environment using numerical optimization and multi-body physics simulation. The toolchain enables multi-objective optimization, generates parametric CAD files that can be further post-processed by an end user, and can be expanded to optimize full systems and non-mechanical parameters such as feedback control variables. We demonstrate the toolchain through an independently verifiable design problem that optimizes wheel radius to achieve a desired linear velocity in a rigid-body physics environment when the wheel rotates at a constant angular speed, and then post-process the parametric CAD file of the optimal design generated by the tool before ultimately manufacturing it via 3D printing. We end with a discussion of how the toolchain can incorporate other analysis tools, including finite element analysis, computational fluid dynamics, and granular media simulations.
Optimization↗
Quantum Hardware-Enabled Molecular Dynamics via Transfer Learning
The ability to perform ab initio molecular dynamics simulations using potential energy surfaces provided by quantum computers would open the door to virtually exact dynamics for a variety of chemical and biochemical systems, with impacts on catalysis and biophysics. Nonetheless, performing molecular dynamics on surfaces produced by quantum hardware has been hampered by the noisy energies typically produced by quantum computers and challenges associated with computing gradients and scaling to large systems interest. A recent set of advances in machine learning, known as transfer learning, provides a new path forward for molecular dynamics simulations on quantum hardware. Transfer learning offers a workaround, where one first trains models on larger, less accurate classical datasets and then refines them on smaller, more accurate quantum datasets. We explore this approach by training machine learning models to predict a molecule's potential energy based on its geometric structure using Behler-Parrinello neural networks. When successfully trained, the model enables energy gradient predictions necessary for dynamic simulations. To reduce the quantum resources needed, the model is initially trained with data derived from classical density functional theory and subsequently refined with a smaller dataset obtained from a variational quantum eigensolver optimization of the unitary coupled cluster ansatz. We show that this approach significantly reduces the size of the needed quantum training dataset while capturing the high accuracies needed within quantum chemistry simulations. The success of this two-step training method opens more opportunities to apply machine learning models on quantum data, a significant stride towards efficient quantum-classical hybrid computational models.
quantum computing↗
From observation to replication: machine-learning-driven quantification and replication of fine-scale fish kinematics and behavior
Long-term quantification of fish behavior is essential for aquatic ecology, wildlife telemetry, and biomechanical device development. However, the observation duration required to obtain reliable behavioral and kinematic metrics remains unclear, and few tools exist to physically reproduce natural swimming motion for controlled experimentation. We address these challenges by developing a generalizable framework that models behavioral reliability (Spearman–Brown reliability index) as a function of observation duration and derives metric-specific monitoring thresholds. Using juvenile white sturgeon (Acipenser transmontanus) as a case study, we demonstrate that the minimum duration needed for reliable estimates varies substantially across kinematic features: to exceed a reliability of 0.8, total distance traveled requires 12 days, average curvature (mm?¹) 15 days, tail-beat frequency (Hz) 8 days, and average speed (body length/s) 17 days. We further bridge digital analysis and physical testing by developing a hardware-in-the-loop simulator that reconstructs machine-learning-derived swimming kinematics with high fidelity (correlation coefficient 0.98–0.99, RMSE 1.22–1.27 mm over a 5-minute segment). This platform enables realistic, repeatable motion stimuli for evaluating aquatic sensing technologies and bio-integrated devices under controlled conditions. Together, these contributions provide a scalable approach for designing long-term behavioral studies and a data-driven connection between ecological observation and robotic experimentation.
Hwang, SungJoo↗
Geant4 Event Biasing and Fast Simulation
Geant4 offers advanced event biasing techniques to significantly accelerate simulations involving rare events. Various biasing methods, such as leading particle selection, cross-section biasing, radioactive decay enhancement, and bremsstrahlung splitting, enable efficient event sampling, though they require careful handling. Additionally, Geant4 provides a Fast Simulation Interface, allowing the replacement of standard processes in specific region and for selected particles, enabling faster execution or external code integration. Applications of fast simulation include electromagnetic shower modeling in calorimeters, machine learning inference, and offloading tasks to specialized hardware like GPUs, making Geant4 a powerful tool for computationally demanding simulations.
72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
Hls4ml Synthesis Testing
HLS4ml (high level synthesis for machine learning) Is a Python package used to translate commonly used open-source machine learning models into HLS. This is useful in machine learning applications on FPGAs. Machine learning algorithms are only as fast as the hardware that they are used on, and some applications require high speed without sacrificing accuracy. In these situations, an FPGA is a good choice since it is faster than a CPU or a GPU, but programming an FPGA is difficult. This is where HLS4ml can be used to simplify the process, as a well-known learning model can be converted to HLS and more easily deployed onto an FPGA. There are many use cases for a machine learning algorithm running on an FPGA. For example, detectors in a particle accelerator cannot keep every event that they detect, and so a computer must decide which events to keep and which to discard. Using an FPGA with a machine learning algorithm would be a good way to keep as many events as possible.
Swanson, Caiden↗
hls4ml
hls4ml (high level synthesis for machine learning) Is a Python package used to translate commonly used open-source machine learning models into HLS. This is useful in machine learning applications on FPGAs. Machine learning algorithms are only as fast as the hardware that they are used on, and some applications require high speed without sacrificing accuracy. In these situations, an FPGA is a good choice since it is faster than a CPU or a GPU, but programming an FPGA is difficult. This is where hls4ml can be used to simplify the process, as a well-known learning model can be converted to HLS and more easily deployed onto an FPGA. There are many use cases for a machine learning algorithm running on an FPGA. For example, detectors in a particle accelerator cannot keep every event that they detect, and so a computer must decide which events to keep and which to discard. Using an FPGA with a machine learning algorithm would be a good way to keep as many events as possible.
Swanson, Caiden↗
Medics: Medical Decision Support System for Long-Duration Space Exploration
The Autonomous Medical Operations (AMO) group at NASA Ames is developing a “medical decision support system” to enable astronauts on long-duration exploration missions to operate autonomously. The system will support clinical actions by providing medical interpretation advice and procedural recommendations during emergent care and clinical work performed by crew. The current state of development of the system, called MedICS (Medical Interpretation Classification and Segmentation) includes two separate aspects: a set of machine learning diagnostic models trained to analyze organ images and patient health records, and an interface to ultrasound diagnostic hardware and to medical repositories. Three sets of images of different organs and medical records were utilized for training machine learning models for various analyses, as follows: 1. Pneumothorax condition (collapsed lung). The trained model provides a positive or negative diagnosis of the condition. 2. Carotid artery occlusion. The trained model produces a diagnosis of 5 different occlusion levels (including “normal”). 3. Ocular retinal images. The model extracts optic disc pixels (image segmentation). This is a precursor step for advanced autonomous fundus clinical evaluation algorithms to be implemented in FY20. 4. Medical health records. The model produces a differential diagnosis for any particular individual, based on symptoms and other health and demographic information. A probability is calculated for each of 25 most common conditions. The same model provides the likelihood of survival. All results are provided with a confidence level. Item 1 images were provided by the US Army and were part of a data set for the clinical treatment of injured battlefield soldiers. This condition is relevant to possible space mishaps, due to pressure management issues. Item 2 images were provided by Houston Methodist Hospital, and item 3 health records were acquired from the MIT laboratory of computational physiology. The machine learning technology utilized is deep multilayer networks (Deep Learning), and new models will continue to be produced, as relevant data is made available and specific health needs of astronaut crews are identified. The interfacing aspects of the system include a GUI for running the different models, and retrieving and storing data, as well as support for integration with an augmented reality (AR) system deployed at JSC by Tietronix Software Inc. (HoloLens). The AR system provides guidance for the placement of an ultrasound transducer that captures images to be sent to the MedICS system for diagnosis. The image captured and the associated diagnosis appear in the technician’s AR visual display.
Colombano, Silvano↗
A CHIL Validation of Machine Learning-Assisted Methods for Real-Time Controls of Solar PV for Grid Services
Recent research has highlighted the potential for solar to act as a zero-marginal-cost and zero-emission flexibility resource on the bulk power system when operated with advanced control systems. To increase the performance of these systems, leading technologies, including machine learning (ML) and hierarchical inverter set point allocation, have been proposed; however, these technologies lack comprehensive validation under real-world application scenarios. This paper addresses this gap by designing and developing a controller-hardware-in-the-loop framework to evaluate the performance of different flexible solar technologies in responding to automatic generation control signals in a closed-loop fashion. Simulation results indicate the superior performance of an ML-based approach compared to the conventional reference-control grouping-based approach, showcasing its potential to support grid stability and operational efficiency.
14 SOLAR ENERGY↗
HDBind: encoding of molecular structure with hyperdimensional binary representations
Traditional methods for identifying “hit” molecules from a large collection of potential drug-like candidates rely on biophysical theory to compute approximations to the Gibbs free energy of the binding interaction between the drug and its protein target. These approaches have a significant limitation in that they require exceptional computing capabilities for even relatively small collections of molecules. Increasingly large and complex state-of-the-art deep learning approaches have gained popularity with the promise to improve the productivity of drug design, notorious for its numerous failures. However, as deep learning models increase in their size and complexity, their acceleration at the hardware level becomes more challenging. Hyperdimensional Computing (HDC) has recently gained attention in the computer hardware community due to its algorithmic simplicity relative to deep learning approaches. The HDC learning paradigm, which represents data with high-dimension binary vectors, allows the use of low-precision binary vector arithmetic to create models of the data that can be learned without the need for the gradient-based optimization required in many conventional machine learning and deep learning methods. This algorithmic simplicity allows for acceleration in hardware that has been previously demonstrated in a range of application areas (computer vision, bioinformatics, mass spectrometery, remote sensing, edge devices, etc.). To the best of our knowledge, our work is the first to consider HDC for the task of fast and efficient screening of modern drug-like compound libraries. We also propose the first HDC graph-based encoding methods for molecular data, demonstrating consistent and substantial improvement over previous work. We compare our approaches to alternative approaches on the well-studied MoleculeNet dataset and the recently proposed LIT-PCBA dataset derived from high quality PubChem assays. We demonstrate our methods on multiple target hardware platforms, including Graphics Processing Units (GPUs) and Field Programmable Gate Arrays (FPGAs), showing at least an order of magnitude improvement in energy efficiency versus even our smallest neural network baseline model with a single hidden layer. Our work thus motivates further investigation into molecular representation learning to develop ultra-efficient pre-screening tools. We make our code publicly available at https://github.com/LLNL/hdbind.
59 BASIC BIOLOGICAL SCIENCES↗
Neuromorphic intermediate representation: A unified instruction set for interoperable brain-inspired computing
Abstract Spiking neural networks and neuromorphic hardware platforms that simulate neuronal dynamics are getting wide attention and are being applied to many relevant problems using Machine Learning. Despite a well-established mathematical foundation for neural dynamics, there exists numerous software and hardware solutions and stacks whose variability makes it difficult to reproduce findings. Here, we establish a common reference frame for computations in digital neuromorphic systems, titled Neuromorphic Intermediate Representation (NIR). NIR defines a set of computational and composable model primitives as hybrid systems combining continuous-time dynamics and discrete events. By abstracting away assumptions around discretization and hardware constraints, NIR faithfully captures the computational model, while bridging differences between the evaluated implementation and the underlying mathematical formalism. NIR supports an unprecedented number of neuromorphic systems, which we demonstrate by reproducing three spiking neural network models of different complexity across 7 neuromorphic simulators and 4 digital hardware platforms. NIR decouples the development of neuromorphic hardware and software, enabling interoperability between platforms and improving accessibility to multiple neuromorphic technologies. We believe that NIR is a key next step in brain-inspired hardware-software co-evolution, enabling research towards the implementation of energy efficient computational principles of nervous systems. NIR is available atneuroir.org
Science & Technology - Other Topics↗