arXiv subjects

Zhehui Wang

Publications and source records attributed to Zhehui Wang.

At least 37 records · Page 2Linked to original sources

arXiv2024

Physics-informed Meta-instrument for eXperiments (PiMiX) with applications to fusion energy

Data-driven methods (DDMs), such as deep neural networks, offer a generic approach to integrated data analysis (IDA), integrated diagnostic-to-control (IDC) workflows through data fusion (DF), which includes multi-instrument data fusion (MIDF), multi-experiment data fusion (MXDF), and simulation-experiment data fusion (SXDF). These features make DDMs attractive to nuclear fusion energy and power plant applications, leveraging accelerated workflows through machine learning and artificial intelligence. Here we describe Physics-informed Meta-instrument for eXperiments (PiMiX) that integrates X-ray (including high-energy photons such as $γ$-rays from nuclear fusion), neutron and others (such as proton radiography) measurements for nuclear fusion. PiMiX solves multi-domain high-dimensional optimization problems and integrates multi-modal measurements with multiphysics modeling through neural networks. Super-resolution for neutron detection and energy resolved X-ray detection have been demonstrated. Multi-modal measurements through MIDF can extract more information than individual or uni-modal measurements alone. Further optimization schemes through DF are possible towards empirical fusion scaling laws discovery and new fusion reactor designs.

physics.data-an↗

arXiv2023

Efficient Spiking Neural Networks with Radix Encoding

Spiking neural networks (SNNs) have advantages in latency and energy efficiency over traditional artificial neural networks (ANNs) due to its event-driven computation mechanism and replacement of energy-consuming weight multiplications with additions. However, in order to reach accuracy of its ANN counterpart, it usually requires long spike trains to ensure the accuracy. Traditionally, a spike train needs around one thousand time steps to approach similar accuracy as its ANN counterpart. This offsets the computation efficiency brought by SNNs because longer spike trains mean a larger number of operations and longer latency. In this paper, we propose a radix encoded SNN with ultra-short spike trains. In the new model, the spike train takes less than ten time steps. Experiments show that our method demonstrates 25X speedup and 1.1% increment on accuracy, compared with the state-of-the-art work on VGG-16 network architecture and CIFAR-10 dataset.

cs.NE↗

arXiv2023

Picosecond X-ray Imaging of Shockwaves with Non-Rankine-Hugoniot Behavior

The first-known observation of plasma-induced cavitation bubbles and expanding shockwaves in liquid during plasma initiation timescales reveals deviation from expected Rankine-Hugoniot shock behavior due to coupled shock-cavitation dynamics, imaged using megahertz-framerate picosecond X-ray imaging. The imaging target features an inexpensive benchtop-scale pulsed plasma device used to generate well-timed spark discharges in ambient liquid heptane at an unprecedented repetition rate ($>$3/min) compared with more commonly used dynamic targets. These shockwaves are relatively weak (Mach number $\leq$ 1.4) compared with X-ray-imaged shockwaves in prior literature, advancing the resolution and sensitivity limits of this high-speed imaging diagnostic. Phase contrast imaging (PCI) has facilitated enhanced quantitative analysis of the expanding shocks in this work, via comparison to thermodynamic models and a Fresnel-Kirchhoff diffraction model.

physics.plasm-ph↗

arXiv2023

Ultrafast Radiographic Imaging and Tracking: An overview of instruments, methods, data, and applications

Ultrafast radiographic imaging and tracking (U-RadIT) use state-of-the-art ionizing particle and light sources to experimentally study sub-nanosecond dynamic processes in physics, chemistry, biology, geology, materials science and other fields. These processes, fundamental to nuclear fusion energy, advanced manufacturing, green transportation and others, often involve one mole or more atoms, and thus are challenging to compute by using the first principles of quantum physics or other forward models. One of the central problems in U-RadIT is to optimize information yield through, e.g. high-luminosity X-ray and particle sources, efficient imaging and tracking detectors, novel methods to collect data, and large-bandwidth online and offline data processing, regulated by the underlying physics, statistics, and computing power. We review and highlight recent progress in: a.) Detectors; b.) U-RadIT modalities; c.) Data and algorithms; and d.) Applications. Hardware-centric approaches to U-RadIT optimization are constrained by detector material properties, low signal-to-noise ratio, high cost and long development cycles of critical hardware components such as ASICs. Interpretation of experimental data, including comparisons with forward models, is frequently hindered by sparse measurements, model and measurement uncertainties, and noise. Alternatively, U-RadIT make increasing use of data science and machine learning algorithms, including experimental implementations of compressed sensing. Machine learning and artificial intelligence approaches, refined by physics and materials information, may also contribute significantly to data interpretation, uncertainty quantification, and U-RadIT optimization.

physics.ins-det↗

arXiv2023

PATCorrect: Non-autoregressive Phoneme-augmented Transformer for ASR Error Correction

Speech-to-text errors made by automatic speech recognition (ASR) systems negatively impact downstream models. Error correction models as a post-processing text editing method have been recently developed for refining the ASR outputs. However, efficient models that meet the low latency requirements of industrial grade production systems have not been well studied. We propose PATCorrect-a novel non-autoregressive (NAR) approach based on multi-modal fusion leveraging representations from both text and phoneme modalities, to reduce word error rate (WER) and perform robustly with varying input transcription quality. We demonstrate that PATCorrect consistently outperforms state-of-the-art NAR method on English corpus across different upstream ASR systems, with an overall 11.62% WER reduction (WERR) compared to 9.46% WERR achieved by other methods using text only modality. Besides, its inference latency is at tens of milliseconds, making it ideal for systems with low latency requirements.

cs.CL↗

arXiv2023

Ultrafast CMOS image sensors and data-enabled super-resolution for multimodal radiographic imaging and tomography

We summarize recent progress in ultrafast Complementary Metal Oxide Semiconductor (CMOS) image sensor development and the application of neural networks for post-processing of CMOS and charge-coupled device (CCD) image data to achieve sub-pixel resolution (thus $super$-$resolution$). The combination of novel CMOS pixel designs and data-enabled image post-processing provides a promising path towards ultrafast high-resolution multi-modal radiographic imaging and tomography applications.

physics.ins-det↗

arXiv2022

Needs, trends, and advances in scintillators for radiographic imaging and tomography

Scintillators are important materials for radiographic imaging and tomography (RadIT), when ionizing radiations are used to reveal internal structures of materials. Since its invention by Röntgen, RadIT now come in many modalities such as absorption-based X-ray radiography, phase contrast X-ray imaging, coherent X-ray diffractive imaging, high-energy X- and $γ-$ray radiography at above 1 MeV, X-ray computed tomography (CT), proton imaging and tomography (IT), neutron IT, positron emission tomography (PET), high-energy electron radiography, muon tomography, etc. Spatial, temporal resolution, sensitivity, and radiation hardness, among others, are common metrics for RadIT performance, which are enabled by, in addition to scintillators, advances in high-luminosity accelerators and high-power lasers, photodetectors especially CMOS pixelated sensor arrays, and lately data science. Medical imaging, nondestructive testing, nuclear safety and safeguards are traditional RadIT applications. Examples of growing or emerging applications include space, additive manufacturing, machine vision, and virtual reality or `metaverse'. Scintillator metrics such as light yield and decay time are correlated to RadIT metrics. More than 160 kinds of scintillators and applications are presented during the SCINT22 conference. New trends include inorganic and organic scintillator heterostructures, liquid phase synthesis of perovskites and $μ$m-thick films, use of multiphysics models and data science to guide scintillator development, structural innovations such as photonic crystals, nanoscintillators enhanced by the Purcell effect, novel scintillator fibers, and multilayer configurations. Opportunities exist through optimization of RadIT with reduced radiation dose, data-driven measurements, photon/particle counting and tracking methods supplementing time-integrated measurements, and multimodal RadIT.

physics.ins-det↗

arXiv2022

Contour Extraction of Inertial Confinement Fusion Images By Data Augmentation

X-Ray radiographs are one of the primary results from inertial confinement fusion (ICF) experiments. Issues such as scarcity of experimental data, high levels of noise in the data, lack of ground truth data, and low resolution of data limit the use of machine/deep learning for automated analysis of radiographs. In this work we combat these roadblocks to the use of machine learning by creating a synthetic radiograph dataset resembling experimental radiographs. Accompanying each synthetic radiograph are corresponding contours of each capsule shell shape, which enables neural networks to train on the synthetic data for contour extraction and be applied to the experimental images. Thus, we train an instance of the convolutional neural network U-Net to segment the shape of the outer shell capsule using the synthetic dataset, and we apply this instance of U-Net to a set of radiographs taken at the National Ignition Facility. We show that the network extracted the outer shell shape of a small number of capsules as an initial demonstration of deep learning for automatic contour extraction of ICF images. Future work may include extracting outer shells from all of the dataset, applying different kinds of neural networks, and extraction of inner shell contours as well.

physics.plasm-ph↗

arXiv2022

A Resource-efficient Spiking Neural Network Accelerator Supporting Emerging Neural Encoding

Spiking neural networks (SNNs) recently gained momentum due to their low-power multiplication-free computing and the closer resemblance of biological processes in the nervous system of humans. However, SNNs require very long spike trains (up to 1000) to reach an accuracy similar to their artificial neural network (ANN) counterparts for large models, which offsets efficiency and inhibits its application to low-power systems for real-world use cases. To alleviate this problem, emerging neural encoding schemes are proposed to shorten the spike train while maintaining the high accuracy. However, current accelerators for SNN cannot well support the emerging encoding schemes. In this work, we present a novel hardware architecture that can efficiently support SNN with emerging neural encoding. Our implementation features energy and area efficient processing units with increased parallelism and reduced memory accesses. We verified the accelerator on FPGA and achieve 25% and 90% improvement over previous work in power consumption and latency, respectively. At the same time, high area efficiency allows us to scale for large neural network models. To the best of our knowledge, this is the first work to deploy the large neural network model VGG on physical FPGA-based neuromorphic hardware.

cs.AR↗

arXiv2022

Machine Learning for Detection of 3D Features using sparse X-ray data

In many inertial confinement fusion experiments, the neutron yield and other parameters cannot be completely accounted for with one and two dimensional models. This discrepancy suggests that there are three dimensional effects which may be significant. Sources of these effects include defects in the shells and shell interfaces, the fill tube of the capsule, and the joint feature in double shell targets. Due to their ability to penetrate materials, X-rays are used to capture the internal structure of objects. Methods such as Computational Tomography use X-ray radiographs from hundreds of projections in order to reconstruct a three dimensional model of the object. In experimental environments, such as the National Ignition Facility and Omega-60, the availability of these views is scarce and in many cases only consist of a single line of sight. Mathematical reconstruction of a 3D object from sparse views is an ill-posed inverse problem. These types of problems are typically solved by utilizing prior information. Neural networks have been used for the task of 3D reconstruction as they are capable of encoding and leveraging this prior information. We utilize half a dozen different convolutional neural networks to produce different 3D representations of ICF implosions from the experimental data. We utilize deep supervision to train a neural network to produce high resolution reconstructions. We use these representations to track 3D features of the capsules such as the ablator, inner shell, and the joint between shell hemispheres. Machine learning, supplemented by different priors, is a promising method for 3D reconstructions in ICF and X-ray radiography in general.

cs.CV↗

arXiv2022

2022 Review of Data-Driven Plasma Science

Data science and technology offer transformative tools and methods to science. This review article highlights latest development and progress in the interdisciplinary field of data-driven plasma science (DDPS). A large amount of data and machine learning algorithms go hand in hand. Most plasma data, whether experimental, observational or computational, are generated or collected by machines today. It is now becoming impractical for humans to analyze all the data manually. Therefore, it is imperative to train machines to analyze and interpret (eventually) such data as intelligently as humans but far more efficiently in quantity. Despite the recent impressive progress in applications of data science to plasma science and technology, the emerging field of DDPS is still in its infancy. Fueled by some of the most challenging problems such as fusion energy, plasma processing of materials, and fundamental understanding of the universe through observable plasma phenomena, it is expected that DDPS continues to benefit significantly from the interdisciplinary marriage between plasma science and data science into the foreseeable future.

physics.plasm-ph↗

arXiv2022

Plasma Image Classification Using Cosine Similarity Constrained CNN

Plasma jets are widely investigated both in the laboratory and in nature. Astrophysical objects such as black holes, active galactic nuclei, and young stellar objects commonly emit plasma jets in various forms. With the availability of data from plasma jet experiments resembling astrophysical plasma jets, classification of such data would potentially aid in investigating not only the underlying physics of the experiments but the study of astrophysical jets. In this work we use deep learning to process all of the laboratory plasma images from the Caltech Spheromak Experiment spanning two decades. We found that cosine similarity can aid in feature selection, classify images through comparison of feature vector direction, and be used as a loss function for the training of AlexNet for plasma image classification. We also develop a simple vector direction comparison algorithm for binary and multi-class classification. Using our algorithm we demonstrate 93% accurate binary classification to distinguish unstable columns from stable columns and 92% accurate five-way classification of a small, labeled data set which includes three classes corresponding to varying levels of kink instability.

physics.plasm-ph↗

arXiv2022

Efficiency Studies of Fast Neutron Tracking using MCNP

Fast neutron identification and spectroscopy is of great interest to nuclear physics experiments. Using the neutron elastic scattering, the fast neutron momentum can be measured. (Wang and Morris, 2013) introduced the theoretical concept that the initial fast neutron momentum can be derived from up to three consecutive elastic collisions between the neutron and the target, including the information of two consecutive recoil ion tracks and the vertex position of the third collision or two consecutive elastic collisions with the timing information. Here we also include the additional possibility of measuring the deposited energies from the recoil ions. In this paper, we simulate the neutron elastic scattering using the Monte Carlo N-Particle Transport Code (MCNP) and study the corresponding neutron detection and tracking efficiency. The corresponding efficiency and the scattering distances are simulated with different target materials, especially natural silicon (92.23$\%$ $^{28}$Si, 4.67$\%$ $^{29}$Si, and 3.1$\%$ $^{30}$Si) and helium-4 ($^4$He). The timing of collision and the recoil ion energy are also investigated, which are important characters for the detector design. We also calculate the ion travelling range for different energies using the software, "The Stopping and Range of Ions in Matter (SRIM)", showing that the ion track can be most conveniently observed in $^4$He unless sub-micron spatial resolution can be obtained in silicon.

physics.ins-det↗

arXiv2022

Stability of Bernstein type theorem for the minimal surface equation

Let $ Ω\subsetneq \mathbf{R}^n\,(n\geq 2)$ be an unbounded convex domain. We study the minimal surface equation in $Ω$ with boundary value given by the sum of a linear function and a bounded uniformly continuous function in $ \mathbf{R}^n$. If $ Ω$ is not a half space, we prove that the solution is unique. If $ Ω$ is a half space, we prove that graphs of all solutions form a foliation of $Ω\times\mathbf{R}$. This can be viewed as a stability type theorem for Edelen-Wang's Bernstein type theorem in \cite{EW2021}. We also establish a comparison principle for the minimal surface equation in $Ω$.

math.AP↗

arXiv2021

Optimizing for In-memory Deep Learning with Emerging Memory Technology

In-memory deep learning computes neural network models where they are stored, thus avoiding long distance communication between memory and computation units, resulting in considerable savings in energy and time. In-memory deep learning has already demonstrated orders of magnitude higher performance density and energy efficiency. The use of emerging memory technology promises to increase the gains in density, energy, and performance even further. However, emerging memory technology is intrinsically unstable, resulting in random fluctuations of data reads. This can translate to non-negligible accuracy loss, potentially nullifying the gains. In this paper, we propose three optimization techniques that can mathematically overcome the instability problem of emerging memory technology. They can improve the accuracy of the in-memory deep learning model while maximizing its energy efficiency. Experiments show that our solution can fully recover most models' state-of-the-art accuracy, and achieves at least an order of magnitude higher energy efficiency than the state-of-the-art.

cs.LG↗

arXiv2021

E3NE: An End-to-End Framework for Accelerating Spiking Neural Networks with Emerging Neural Encoding on FPGAs

Compiler frameworks are crucial for the widespread use of FPGA-based deep learning accelerators. They allow researchers and developers, who are not familiar with hardware engineering, to harness the performance attained by domain-specific logic. There exists a variety of frameworks for conventional artificial neural networks. However, not much research effort has been put into the creation of frameworks optimized for spiking neural networks (SNNs). This new generation of neural networks becomes increasingly interesting for the deployment of AI on edge devices, which have tight power and resource constraints. Our end-to-end framework E3NE automates the generation of efficient SNN inference logic for FPGAs. Based on a PyTorch model and user parameters, it applies various optimizations and assesses trade-offs inherent to spike-based accelerators. Multiple levels of parallelism and the use of an emerging neural encoding scheme result in an efficiency superior to previous SNN hardware implementations. For a similar model, E3NE uses less than 50% of hardware resources and 20% less power, while reducing the latency by an order of magnitude. Furthermore, scalability and generality allowed the deployment of the large-scale SNN models AlexNet and VGG.

cs.NE↗

arXiv2021

A Bernstein Type Theorem for Minimal Graphs over Convex Domains

Given any $n \geq 2$, we show that if $Ω\subsetneq \mathbb{R}^n$ is an open convex domain (e.g. a half-space), and $u : Ω\to \mathbb{R}$ is a solution to the minimal surface equation which agrees with a linear function on $\partial Ω$, then $u$ must itself be linear.

math.AP↗

arXiv2021

Entangled-photon interferometry for plasmas

Sub-picosecond coincidence timing from nonlocal intensity interference of entangled photons allows quantum interferometry for plasmas. Using a warm plasma dispersion relation, we correlate phase measurement sensitivity with different plasma properties or physics mechanisms over 6 orders of magnitude. Due to $N^α$ ($α\leq -1/2$) scaling with the photon number $N$, quantum interferometry using entangled light can probe small signals in plasmas not previously accessible. As an example, it is predicted that plasmas will induce shifts to a Gaussian dip, a well-known quantum optics phenomenon that is yet to be demonstrated for plasmas.

quant-ph↗