Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Network inference”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
NASA NTRSDate not supplied

Methods for Determining Subsets of High Impact, Probabilistically Dependent Medical Conditions Represented in a Directed Graph

One of the longest standing questions in network theory is how a component influences other parts in the system, and how that role is affected when restricting the navigation through the network. The Katz score, one of many centrality measures created for this purpose, takes into account all possible walks through the network, penalizing each additional step in a walk by a scalar called the Katz parameter. This centrality measure often covers an infinite number of walks with infinite length. In this paper we identify the maximum path length which has influence on the Katz score. We ultimately provide guidance when deciding which Katz parameter to use as it depends on the path length of interest. We show how changing the Katz parameter affects the ranking of the vertices in some synthetic graphs as well as NASA's expert informed network of medical dependencies called the Susceptibility Inference Network (SIN).

Hunter Rehm↗
DOE OSTI2024

Generic Multi-Layer Perceptron Inference Accelerator on FPGA (vneuron) v1.0

We have designed and implemented a neural network inference compute engine (vneuron) that can be deployed in the fabric of any FPGA without using special hardware accelerator primitive. The "vneuron" is purely written in verilog, and supports scalable neural network structure with fully connected layers and ReLU activation ( Multi-Layer Perceptron architecture) with 16 bits of precision. We have demonstrated it on an Xilinx Artix 7 FPGA for a 16-input, 8-output MLP with 3 layer, 1600 parameters. It takes 40 DSP48E and 40 BRAM18, and takes 131 clock cycles for computing (1048 ns when clocked at 125MHz). We include PyTorch quantization from a given floating point model, and provide behavioral verification simulation in the disclosed software package.

Du, Qiang↗
DOE OSTI2021

Low-Power Deep Learning Inference using the SpiNNaker Neuromorphic Platform

n this presentation we will discuss recent results on using the SpiNNaker neuromorphic platform (48-chip model) for deep learning neural network inference. We use the Sandia Labs developed Whet stone spiking deep learning library to train deep multi-layer perceptrons and convolutional neural networks suitable for the spiking substrate on the neural hardware architecture. By using the massively parallel nature of SpiNNaker, we are able to achieve, under certain network topologies, substantial network tiling and consequentially impressive inference throughput. Such high-throughput systems may have eventual application in remote sensing applications where large images need to be chipped, scanned, and processed quickly. Additionally, we explore complex topologies that push the limits of the SpiNNaker routing hardware and investigate how that impacts mapping software-implemented networks to on-hardware instantiations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Analysis and mitigation of parasitic resistance effects for analog in-memory neural network acceleration

To support the increasing demands for efficient deep neural network processing, accelerators based on analog in-memory computation of matrix multiplication have recently gained significant attention for reducing the energy of neural network inference. However, analog processing within memory arrays must contend with the issue of parasitic voltage drops across the metal interconnects, which distort the results of the computation and limit the array size. This work analyzes how parasitic resistance affects the end-to-end inference accuracy of state-of-the-art convolutional neural networks, and comprehensively studies how various design decisions at the device, circuit, architecture, and algorithm levels affect the system's sensitivity to parasitic resistance effects. Here, a set of guidelines are provided for how to design analog accelerator hardware that is intrinsically robust to parasitic resistance, without any explicit compensation or re-training of the network parameters.

97 MATHEMATICS AND COMPUTING↗
NASA NTRS2020

Precipitation and Aridity Constraints on Early Mars from Globally-Distributed Paleolakes

The widespread occurrence of fluvio-lacustrine features on Mars support long-lived flow and accumulation of water in a warmer, wetter past. However, martian climate models have been unable to recreate the necessary conditions required to support a persistent wet climate. Orbital and in-situ data sets have revealed the existence of > 400 paleolakes on Mars, which can be subdivided into open- and closed-basin lakes. Open-basin lakes require that sufficient water accumulated to fill and overtop the basin-confining topography, providing a minimum constraint on required water volumes. Conversely, closed-basin lakes provide maximum water volumes since the absence of an outlet breach generally implies they did not overflow. Importantly, a subset of both open- and closed-basin lakes are fed by valley networks inferred to have been sourced by precipitation during the era of valley network formation > 3.7 Ga and may be used to quantitatively constrain precipitation and aridity during early Mars.

Stucky De Quay, G.↗
DOE OSTI2020

Phosphoproteome Response to Dithiothreitol Reveals Unique Versus Shared Features of Saccharomyces cerevisiae Stress Responses

To cope with sudden changes in the external environment, the budding yeast Saccharomyces cerevisiae orchestrates a multifaceted response that spans many levels of physiology. Several studies have interrogated the transcriptome response to endoplasmic reticulum (ER) stress and the role of regulators such as the Ire1 kinase and Hac1 transcription factors. However, less is known about responses to ER stress at other levels of physiology. Here, we used quantitative phosphoproteomics and computational network inference to uncover the yeast phosphoproteome response to the reducing agent dithiothreitol (DTT) and the upstream signaling network that controls it. We profiled wild-type cells and mutants lacking IRE1 or MAPK kinases MKK1 and MKK2, before and at various times after DTT treatment. In addition to revealing downstream targets of these kinases, our inference approach predicted new regulators in the DTT response, including cell-cycle regulator Cdc28 and osmotic-response kinase Rck2, which we validated computationally. Our results also revealed similarities and surprising differences in responses to different stress conditions, especially in the response of protein kinase A targets. Furthermore, these results have implications for the breadth of signaling programs that can give rise to common stress response signatures.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

An orthologous gene coevolution network provides insight into eukaryotic cellular and genomic structure and function

The evolutionary rates of functionally related genes often covary. We present a gene coevolution network inferred from examining nearly 3 million orthologous gene pairs from 332 budding yeast species spanning ~400 million years of evolution. Network modules provide insight into cellular and genomic structure and function. Examination of the phenotypic impact of network perturbation using deletion mutant data from the baker’s yeast Saccharomyces cerevisiae, which were obtained from previously published studies, suggests that fitness in diverse environments is affected by orthologous gene neighborhood and connectivity. Mapping the network onto the chromosomes of S. cerevisiae and Candida albicans revealed that coevolving orthologous genes are not physically clustered in either species; rather, they are often located on different chromosomes or far apart on the same chromosome. The coevolution network captures the hierarchy of cellular structure and function, provides a roadmap for genotype-to-phenotype discovery, and portrays the genome as a linked ensemble of genes.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Learned adaptive properties for mitigation of weight perturbations in embedded spiking networks

Recent years have seen an increased importance of neural network inference in edge-based scenarios, which impose size and power constraints requiring novel computing devices. These same edge scenarios may require operating over long periods of time, or exposure to extreme environments, resulting in a drift of neural network weights that cause degraded performance. In searching for ways to develop neural network approaches that perform robustly under these conditions, we propose a biologically-inspired mechanism for the dynamic adaptation of within-neuron parameters that is guided by a global context signal carrying information about perturbations and variability in incoming stimuli. Specifically, we demonstrate that adaptive voltage thresholds or neuronal time constants, when informed by a global context signal, can enable network-level mechanisms to recover from perturbed synaptic weights. Consistent with prior literature, the context-modulated approach is effective for recurrent, but not feedforward networks, by modulating network level dynamics. We demonstrate this approach successfully recovers performance in image classification tasks and spatiotemporal tracking tasks under idealized and Gaussian noise as well as for realistic perturbations from a memristive device when exposed to ionizing radiation. Finally, we discuss how this approach enables the design of robust and energy-efficient neuromorphic systems that perform well, even in resource-constrained scenarios with extreme environments such as edge processing.

context modulation↗
DOE OSTI2024

Radiation-Induced Noise Resilience of Neuromorphic Architectures

Neuromorphic event-based networks use asynchronous time-dependent information to extract features from input data that can allow for edge-based distributed applications such as object recognition. The noise resilience properties of such networks, especially in the context of space applications, are yet to be explored. In this paper, we use the hierarchy of time surfaces (HOTS) algorithm, which is one of the neuromorphic algorithms, to understand the least and most resilient modules in a neuromorphic network. The HOTS algorithm relies on the computing of time surfaces that maps the temporal delays between neighboring pixels into normalized features that involve many computations that are also found in other neuromorphic networks such as exponential decays, distance computations, etcetera. We implemented HOTS on a Digilent PYNQ board with a Xilinx Zynq 7020 system on a chip, and we subjected the boards running the HOTS network inference to neutron radiation at the Los Alamos Neutron Science Center. Furthermore, we used simulation models from our previous similar experiments on the event-based sensor to create a neutron induced noise model to quantify the effect of this noise on the overall performance of the network. This experiment provides the preliminary measurements of the reliability of the HOTS algorithm and proposes methods to create a more reliable HOTS architecture in future spacecraft missions.

Engineering↗
DOE OSTI2021

GPU-Accelerated Machine Learning Inference as a Service for Computing in Neutrino Experiments

Machine learning algorithms are becoming increasingly prevalent and performant in the reconstruction of events in accelerator-based neutrino experiments. These sophisticated algorithms can be computationally expensive. At the same time, the data volumes of such experiments are rapidly increasing. The demand to process billions of neutrino events with many machine learning algorithm inferences creates a computing challenge. We explore a computing model in which heterogeneous computing with GPU coprocessors is made available as a web service. The coprocessors can be efficiently and elastically deployed to provide the right amount of computing for a given processing task. With our approach, Services for Optimized Network Inference on Coprocessors (SONIC), we integrate GPU acceleration specifically for the ProtoDUNE-SP reconstruction chain without disrupting the native computing workflow. With our integrated framework, we accelerate the most time-consuming task, track and particle shower hit identification, by a factor of 17. This results in a factor of 2.7 reduction in the total processing time when compared with CPU-only production. For this particular task, only 1 GPU is required for every 68 CPU threads, providing a cost-effective solution.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2023

Climate change–induced stress disrupts ectomycorrhizal interaction networks at the boreal–temperate ecotone

The interaction networks formed by ectomycorrhizal fungi (EMF) and their tree hosts, which are important to both forest recruitment and ecosystem carbon and nutrient retention, may be particularly susceptible to climate change at the boreal–temperate forest ecotone where environmental conditions are changing rapidly. Here, we quantified the compositional and functional trait responses of EMF communities and their interaction networks with two boreal ( Pinus banksiana and Betula papyrifera ) and two temperate ( Pinus strobus and Quercus macrocarpa ) hosts to a factorial combination of experimentally elevated temperatures and reduced rainfall in a long-term open-air field experiment. The study was conducted at the B4WarmED (Boreal Forest Warming at an Ecotone in Danger) experiment in Minnesota, USA, where infrared lamps and buried heating cables elevate temperatures (ambient, +3.1 °C) and rain-out shelters reduce growing season precipitation (ambient, ~30% reduction). EMF communities were characterized and interaction networks inferred from metabarcoding of fungal-colonized root tips. Warming and rainfall reduction significantly altered EMF community composition, leading to an increase in the relative abundance of EMF with contact-short distance exploration types. These compositional changes, which likely limited the capacity for mycelial connections between trees, corresponded with shifts from highly redundant EMF interaction networks under ambient conditions to less redundant (more specialized) networks. Further, the observed changes in EMF communities and interaction networks were correlated with changes in soil moisture and host photosynthesis. Collectively, these results indicate that the projected changes in climate will likely lead to significant shifts in the traits, structure, and integrity of EMF communities as well as their interaction networks in forest ecosystems at the boreal–temperate ecotone.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Disentangling Sources of Gene Tree Discordance in Phylogenomic Data Sets: Testing Ancient Hybridizations in Amaranthaceae s.l

Gene tree discordance in large genomic data sets can be caused by evolutionary processes such as incomplete lineage sorting and hybridization, as well as model violation, and errors in data processing, orthology inference, and gene tree estimation. Species tree methods that identify and accommodate all sources of conflict are not available, but a combination of multiple approaches can help tease apart alternative sources of conflict. Here, using a phylotranscriptomic analysis in combination with reference genomes, we test a hypothesis of ancient hybridization events within the plant family Amaranthaceae s.l. that was previously supported by morphological, ecological, and Sanger-based molecular data. The data set included seven genomes and 88 transcriptomes, 17 generated for this study. We examined gene-tree discordance using coalescent-based species trees and network inference, gene tree discordance analyses, site pattern tests of introgression, topology tests, synteny analyses, and simulations. We found that a combination of processes might have generated the high levels of gene tree discordance in the backbone of Amaranthaceae s.l. Furthermore, we found evidence that three consecutive short internal branches produce anomalous trees contributing to the discordance. Overall, our results suggest that Amaranthaceae s.l. might be a product of an ancient and rapid lineage diversification, and remains, and probably will remain, unresolved. This work highlights the potential problems of identifiability associated with the sources of gene tree discordance including, in particular, phylogenetic network methods. Our results also demonstrate the importance of thoroughly testing for multiple sources of conflict in phylogenomic analyses, especially in the context of ancient, rapid radiations. We provide several recommendations for exploring conflicting signals in such situations.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Benchmarking the Performance of Neuromorphic and Spiking Neural Network Simulators

Software simulators play a critical role in the development of new algorithms and system architectures in any field of engineering. Neuromorphic computing, which has shown potential in building brain-inspired energy-efficient hardware, suffers a slow-down in the development cycle due to a lack of flexible and easy-to-use simulators of either neuromorphic hardware itself or of spiking neural networks (SNNs), the type of neural network computation executed on most neuromorphic systems. While there are several openly available neuromorphic or SNN simulation packages developed by a variety of research groups, they have mostly targeted computational neuroscience simulations, and only a few have targeted small-scale machine learning tasks with SNNs. Evaluations or comparisons of these simulators have often targeted computational neuroscience-style workloads. In this work, we seek to evaluate the performance of several publicly available SNN simulators with respect to non-computational neuroscience workloads, in terms of speed, flexibility, and scalability. We evaluate the performance of the NEST, Brian2, Brian2GeNN, BindsNET and Nengo packages under a common front-end neuromorphic framework. Our evaluation tasks include a variety of different network architectures and workload types to mimic the computation common in different algorithms, including feed-forward network inference, genetic algorithms, and reservoir computing. We also study the scalability of each of these simulators when running on different computing hardware, from single core CPU workstations to multi-node supercomputers. Our results show that the BindsNET simulator has the best speed and scalability for most of the SNN workloads (sparse, dense, and layered SNN architectures) on a single core CPU. However, when comparing the simulators leveraging the GPU capabilities, Brian2GeNN outperforms the others for these workloads in terms of scalability. NEST performs the best for small sparse networks and is also the most flexible simulator in terms of reconfiguration capability NEST shows a speedup of at least 2x compared to the other packages when running evolutionary algorithms for SNNs. The multi-node and multi-thread capabilities of NEST show at least 2x speedup compared to the rest of the simulators (single core CPU or GPU based simulators) for large and sparse networks. We conclude our work by providing a set of recommendations on the suitability of employing these simulators for different tasks and scales of operations. We also present the characteristics for a future generic ideal SNN simulator for different neuromorphic computing workloads.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Accelerating Hamiltonian Monte Carlo for Bayesian inference in neural networks and neural operators

Hamiltonian Monte Carlo (HMC) is a powerful and accurate method to sample from the posterior distribution in Bayesian inference. However, HMC techniques are computationally demanding for Bayesian neural networks due to the high dimensionality of the network’s parameter space and the non-convexity of their posterior distributions. Therefore, various approximation techniques, such as variational inference (VI) or stochastic gradient MCMC, are often employed to infer the posterior distribution of the network parameters. Such approximations introduce inaccuracies in the inferred distributions, resulting in unreliable uncertainty estimates. In this work, we propose a hybrid approach that combines inexpensive VI and accurate HMC methods to efficiently and accurately quantify uncertainties in neural networks and neural operators. The proposed approach leverages an initial VI training on the full network. We examine the influence of individual parameters on the prediction uncertainty, which shows that a large proportion of the parameters do not contribute substantially to uncertainty in the network predictions. This information is then used to significantly reduce the dimension of the parameter space, and HMC is performed only for the subset of network parameters that strongly influence prediction uncertainties. This yields a framework for accelerating the full batch HMC for posterior inference in neural networks. We demonstrate the efficiency and accuracy of the proposed framework on deep neural networks and operator networks, showing that inference can be performed for large networks with tens to hundreds of thousands of parameters. Finally, we show that this method can effectively learn surrogates for complex physical systems by modeling the operator that maps from upstream conditions to wall-pressure data on a cone in hypersonic flow.

Bayesian inference↗
DOE OSTI2023

A 3D Implementation of Convolutional Neural Network for Fast Inference

Low latency inference has many applications in edge machine learning. In this paper, we present a run-time configurable convolutional neural network (CNN) inference ASIC design for low-latency edge machine learning. By implementing a 5-stage pipelined CNN inference model in a 3D ASIC technology, we demonstrate that the model distributed on two dies utilizing face-to-face (F2F) 3D integration achieves superior performance. Our experimental results show that the design based on 3D integration achieves 43% better energy-delay product when compared to the traditional 2D technology.

Miniskar, Narasinga Rao↗
DOE OSTI2025

Overexpression of plasma membrane SUT1 in poplar alters lateral sucrose partitioning in stem and promotes leaf necrosis

Abstract In Populus and many other tree species, photoassimilate sucrose diffuses down a concentration gradient via symplastically connected mesophyll cells to minor vein phloem for long‐distance transport. There is no evidence for apoplastic phloem‐loading in Populus . However, plasma membrane sucrose transporters (SUT1 and SUT3) orthologous to those associated with apoplastic phloem loading are expressed in vascular tissues of poplar. While SUT3 functions in sucrose import into developing xylem, the role of SUT1 remains unclear. Here, we overexpressed PtaSUT1 in Populus tremula x P. alba to examine the effects on sucrose partitioning in transgenic plants. Overall leaf sucrose levels were similar between wild type and transgenic lines. Stem sucrose levels were not changed in bark but were significantly reduced in the adjacent xylem, suggesting hindered intercellular sucrose trafficking from the phloem to the developing xylem. Fully expanded leaves of transgenic plants deteriorated prematurely with declining photosynthesis prior to severe necrotic spotting. Necrotic spotting advanced most rapidly in the distal portion of mature leaves and was accompanied by sharp hexose increases and sharp sucrose decreases there. Leaf transcriptome profiling and network inference revealed the down‐regulation of copper proteins and elevated expression of copper microRNAs prior to noticeable leaf injury. Our results suggest ectopic expression of PtaSUT1 altered sucrose partitioning in stems with systemic effects on leaf health and copper homeostasis mediated in part by sucrose‐sensitive copper miRNAs.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service

Computing demands for large scientific experiments, such as the CMS experiment at the CERN LHC, will increase dramatically in the next decades. To complement the future performance increases of software running on central processing units (CPUs), explorations of coprocessor usage in data processing hold great potential and interest. Coprocessors are a class of computer processors that supplement CPUs, often improving the execution of certain functions due to architectural design choices. We explore the approach of Services for Optimized Network Inference on Coprocessors (SONIC) and study the deployment of this as-a-service approach in large-scale data processing. In the studies, we take a data processing workflow of the CMS experiment and run the main workflow on CPUs, while offloading several machine learning (ML) inference tasks onto either remote or local coprocessors, specifically graphics processing units (GPUs). With experiments performed at Google Cloud, the Purdue Tier-2 computing center, and combinations of the two, we demonstrate the acceleration of these ML algorithms individually on coprocessors and the corresponding throughput improvement for the entire workflow. This approach can be easily generalized to different types of coprocessors and deployed on local CPUs without decreasing the throughput performance. We emphasize that the SONIC approach enables high coprocessor usage and enables the portability to run workflows on different types of coprocessors.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗