Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “end-to-end”

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 235 records · Page 13
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

End-to-end orientation estimation from 2D cryo-EM1images

Cryo-electron microscopy (cryo-EM) is a Nobel Prize-winning technique for deter-mining high-resolution 3D structures of biological macromolecules. A 3D structure is reconstructed from hundreds of thousands of noisy 2D projection images. However, existing 3D reconstruction methods are still time-consuming, and one of the major computational bottlenecks is to recover the unknown orientation of the particle in16each 2D image. The dominant methods typically exploit expensive global search on each image to estimate the missing orientations. Here, a novel end-to-end supervised learning method is introduced to directly recover the missing orientations from 2D cryo-EM images. A neural network is used to approximate the mapping from images to orientations. Furthermore, a robust loss function is proposed for optimizing the parameters of the network, which can handle both asymmetric and symmetric 3D structures. Experiments on synthetic datasets with various symmetry types confirm that the neural network is capable of recovering orientations from 2D cryo-EM images, and the results on one real cryo-EM dataset further demonstrate its potential in more challenging imaging conditions.

3D reconstruction↗
DOE OSTI2024

Towards FAIR Workflows for Federated Experimental Sciences

A de-centralized, peer-to-peer AI metadata framework is demonstrated which can enable end-to-end metadata & lineage tracking for distributed Machine Learning pipelines spanning edge, High Performance Computing, and cloud environments. With a specific example of end-to-end microscopy algorithm and datasets, the proposed method shows how to enable reproducibility, audit trail, provenance of metadata artifacts. The emerging needs of automation in experimental sciences, ML-centric workflows, and FAIR metadata management across federated compute environments is addressed.

machine learning↗
DOE OSTI2024

Metric Learning to Accelerate Convergence of Operator Splitting Methods

Recent developments in machine learning have led to promising advances in accelerating the solution of constrained optimization problems. Increasing demand for real-time decision-making capabilities in applications such as artificial intelligence and optimal control has led to a variety of proposed strategies for learning to produce fast solutions to optimization problems. For example, recent works have shown that it is possible to accelerate the convergence of optimization algorithms by learning to select their parameters, such as gradient descent stepsizes. This work proposes a new approach, in which the underlying metric spaces of proximal operator splitting algorithms are learned to maximize convergence rate. While prior works in optimization theory have derived optimal metrics in simple cases, no such result exists for many practical problem forms including general Quadratic Programming (QP). This paper shows how differentiable optimization can enable the end-to-end learning of proximal metrics, enhancing the convergence of proximal algorithms for QP problems beyond what is possible based on known theory. Additionally, the results illustrate a strong connection between the learned proximal metrics and active constraints at the optima, leading to an interpretation in which the predicted proximal metrics can be viewed as a form of active set prediction.

King, Ethan [BATTELLE (PACIFIC NW LAB)]↗
DOE OSTI2024

MassiveGNN: Efficient Training via Prefetching for Massively Connected Distributed Graphs

Graph Neural Networks (GNN) are indispensable in learning from graph-structured data, yet their rising computational costs, especially on massively connected graphs, pose significant challenges in terms of execution performance. To tackle this, distributed-memory solutions such as partitioning the graph to concurrently train multiple replicas of GNNs are in practice. However, approaches requiring a partitioned graph usually suffer from communication overhead and load imbalance, even under optimal partitioning and communication strategies due to irregularities in the neighborhood minibatch sampling. This paper proposes practical trade-offs for improving the sampling and communication overheads for representation learn- ing on distributed graphs (using popular GraphSAGE architecture) by developing a parameterized prefetch and eviction scheme on top of the state-of-the-art Amazon DistDGL distributed GNN framework, demonstrating about 15–40% improvement in end-to-end training performance on the NERSC Perlmutter supercomputer for various OGB datasets.

Machine Leanring, high performance comptuing, grap↗
DOE OSTI2021

Invited: Bambu: an Open-Source Research Framework for the High-Level Synthesis of Complex Applications

This paper presents the open-source High-Level Synthesis research framework Bambu. The framework provides an open-source starting point to experiment with new ideas across High-Level Synthesis, high-level verification and debugging, FPGA/ASIC design, design flow space exploration, and parallel hardware accelerator design. The tool accepts as input standard C/C++ specifications and compiler intermediate representations (IRs) coming from the well-known Clang/LLVM and GCC com- pilers. The broad spectrum and flexibility of input formats allow the electronic design automation (EDA) research community to explore and integrate new transformations and optimizations. The easily extendable modular framework already includes many op- timizations and HLS benchmarks. The integration with synthesis and verification backends (commercial and open-source) allows researchers to quickly test any new finding and easily obtain performance and resource usage metrics for a given application. Different FPGA devices are supported from several different vendors: AMD/XILINX, Intel/Altera, Lattice Semiconductor, and NanoXplore. Finally, integration with the OpenRoad open-source end-to-end silicon compiler perfectly fits with the recent push towards open-source EDA.

Ferrandi, Fabrizio↗
DOE OSTI2023

ML-CGRA: An Integrated Compilation Framework to Enable Efficient Machine Learning Acceleration on CGRAs

Coarse-Grained Reconfigurable Arrays (CGRAs) can achieve higher energy-efficiency than general-purpose processors and accelerators or fine-grained reconfigurable devices, while maintaining adaptability to different computational patterns. CGRAs have shown some success as a platform to accelerate machine learning (ML) thanks to their flexibility, which allows them to support new models not considered by fixed accelerators. However, current solutions for CGRAs employ low level instruction-based compiler approaches and lack specialized compilation infrastructures from high-level ML frameworks that could leverage semantic information from the models, limiting the ability to efficiently map them on the recon- figurable substrate. This paper proposes ML-CGRA, an integrated compilation framework based on the MLIR infrastructure that en- ables efficient ML acceleration on CGRAs. ML-CGRA provides an end-to-end solution for mapping ML models on CGRAs that out- performs conventional approaches by 3.15× and 6.02 × on 4×4 and 8×8 CGRAs, respectively.

machine learning↗
DOE OSTI2022

An Algorithmic and Software Pipeline for Very Large Scale Scientific Data Compression with Error Guarantees

Efficient data compression is becoming increasingly critical for storing scientific data because many scientific applications produce vast amounts of data. This paper presents an end-to-end algorithmic and software pipeline for data compression that guarantees both error bounds on primary data (PD) and derived data, known as Quantities of Interest (QoI).We demonstrate the effectiveness of the pipeline by compressing fusion data generated by a large-scale fusion code, XGC, which produces tens of petabytes of data in a single day. We demonstrate that the compression is conducted by setting aside computational resources known as staging nodes, and does not impact the simulation performance. For efficient parallel I/O, the pipeline uses ADIOS2, which many codes such as XGC already use for their parallel I/O. We show that our approach can compress the data by two orders of magnitude while guaranteeing high accuracy on both the PD and the QoIs. Further, the amount of resources required by compression is a few percent of the resources required by simulation while ensuring that the compression time for each stage is less than the corresponding simulation time.This pipeline consists of three main steps. The first step decomposes the data using domain decomposition into small subdomains. Each subdomain is then compressed independently to achieve a high level of parallelism. The second step uses existing techniques that guarantee error bounds on the primary data for each subdomain. The third step uses a post-processing optimization technique based on Lagrange multipliers to reduce the QoI errors for data corresponding to each subdomain. The Lagrange multipliers generated can be further quantized or truncated to increase the compression level. All of the above characteristics of our approach make it highly practical to apply on-the-fly compression while guaranteeing errors on QoIs that are critical to the scientists.

Banerjee, Tania↗
DOE OSTI2025

A Parallel Alternative for Energy-Efficient Neural Network Training and Inferencing

Energy efficiency of training and inferencing with large neural network models is a critical challenge facing the future of sustainable large-scale machine learning workloads. This paper introduces an alternative strategy, called phantom parallelism, to minimize the net energy consumption of traditional tensor (model) parallelism, the most energy-inefficient component of large neural network training. The approach is presented in the context of feed-forward network architectures as a preliminary, but comprehensive, proof-of-principle study of the proposed methodology. We derive new forward and backward propagation operators for phantom parallelism, implement them as custom autograd operations within an end-to-end phantom parallel training pipeline and compare its parallel performance and energy-efficiency against those of conventional tensor parallel training pipelines. Formal analyses that predict lower bandwidth and FLOP counts are presented with supporting empirical results on up to 256 GPUs that corroborate these gains. Experiments are shown to deliver ∼50% reduction in the energy consumed to train FFNs using the proposed phantom parallel approach when compared with conventional tensor parallel methods. Additionally, the proposed approach is shown to train smaller phantom models to the same model loss on smaller GPU counts as larger tensor parallel models on larger GPU counts offering the possibility for even greater energy savings.

Seal, Sudip [ORNL] (ORCID:0000000332330656)↗
DOE OSTI2023

A Pulse Generation Framework with Augmented Program-aware Basis Gates and Criticality Analysis

Near-term intermediate scale quantum (NISQ) de- vices are subject to considerable noise and short coherence time. Consequently, it is critical to minimize circuit execution latency. Traditionally, each basis gate of a transpiled circuit is decoded into a fixed episode of the device control pulses. Recently, people started to investigate merged pulse generation for customized gates through quantum optimal control (QOC). However, existing QOC approaches face the challenges of (i) restricted search space due to prohibitive compilation overhead; (ii) suboptimal end-to-end performance due to aggressive local optimization and falsely introduced dependency among the customized gates; (iii) inadequate adaptivity towards system calibration, which is critical for NISQ devices. In this work, we propose PAQOC, a novel QOC framework that can (i) automatically detect frequently encountered gate patterns in the logical circuit by modeling the problem as a subgraph mining process and reuse these patterns to enable much larger search space exploration (i.e., program aware); (ii) systemically construct customized gate-set based on the impact to the overall program latency (i.e., criticality-aware); and (iii) quickly adapt to system re-calibration thanks to the small-scale pattern-based gate generation (i.e., adaptivity-aware). PAQOC achieves a good tradeoff between circuit performance and compilation time, allowing fully automatic, single stop, ad- hoc customized pulse generation for more efficient execution of user programs on NISQ devices. Evaluations using fifteen applications show that PAQOC can achieve on average 1.95× speedup of the circuit latency and achieve on average 36.7% reduction in compilation overhead. With PAQOC, circuits can run faster with reduced noise, allowing deeper circuits to be tested within the coherence time of present NISQ platforms.

Chen, Yanhao↗
DOE OSTI2021

Automated Generation of Integrated Digital and Spiking Neuromorphic Machine Learning Accelerators

The growing numbers of application areas for artificial intelligence (AI) methods have led to an explosion of domain-specific accelerators that could support every new machine learning (ML) algorithm advancement, clearly highlighting the need for a capability to quickly and automatically transition from algorithm definition to hardware implementation and explore design space along a variety of SWaP (size, weight and Power). The software defined architectures (SODA) synthesizer implements a compiler-based modular infrastructure for the end-to-end generation of machine learning accelerators from high-level frameworks to hardware description language. At the same time, neuromorphic computing, by mimicking how the brain operates, promises to perform artificial intelligence tasks at efficiencies orders of magnitude higher than the current conventional tensor-processing based accelerators, as demonstrated by a variety of specialized designs leveraging Spiking Neural Networks (SNNs). Nevertheless, the mapping of an artificial neural network (ANN) to solutions supporting SNNs is still a non-trivial and very device-specific task, and completely lack the possibility to design hybrid systems that integrate conventional and spiking neural models. In this paper we discuss the support for such an integrated generation leveraging the SODA Synthesizer framework and its modular structure. In particular, we present a new MLIR dialect (part of the SODA frontend) that allows expressing spiking neural network features (e.g., available resources, spiking sequences, analog signal reading, etc.) and illustrate how it enables mapping to Spiking Neurons and deployment to the related specialized hardware (which, in the digital domain, could be generated through the other existing layers of the SODA Synthesizer). We then discuss the opportunities for even deeper integration afforded by the hardware compilation infrastructure, providing a path towards the generation of complex heterogeneous artificial intelligence systems.

Curzel, Serena↗
DOE OSTI2023

HCPerf: Driving Performance-Directed Hierarchical Coordination for Autonomous Vehicles

The rapid development of autonomous driving poses new research challenges to the on-vehicle computing system. In particular, the execution time of autonomous driving tasks highly depends on the specific driving environment. For instance, the execution time of configurable sensor fusion increases significantly as the scene becomes complex, which leads to end-to-end deadline misses from sensing to control and may cause accidents. Thus, a framework that can effectively utilize the system resources to guarantee the end-to-end deadlines of autonomous driving tasks as well as effectively prioritize the responsiveness and throughput of the control commands is crucial for autonomous driving. In this paper, we propose HCPerf, a performance-directed hierarchical coordination framework that intelligently coordinates the autonomous driving tasks with high execution time variation and complex dependencies according to the driving performance in real-time. Specifically, HCPerf mainly consists of two coordinators. The internal coordinator intelligently schedules the tasks according to the driving performance of the vehicle in order to help them meet the end-to-end deadlines while well prioritizing the responsiveness and throughput of the control commands. At the same time, the external coordinator dynamically tunes the rates of tasks according to the schedulability in order to efficiently utilize the system resource. We conduct extensive experiments on both simulation and hardware testbeds with the representative autonomous driving application. The results show that HCPerf can effectively improve the driving performance by 7.69%-45.94% in different driving scenarios.

Ma, Jialiang↗
DOE OSTI2023

DENSECL: Haze Mitigation Using Dense Blocks and Contrastive Loss Regularization

Haze, which occurs as a result of the scattering of light in the atmosphere by small particles, diminishes the visibility of scene objects, inflicting important image applications such as object detection. To address the problem, this paper introduces a new physics-based end-to-end deep learning approach to haze mitigation in outdoor scenes, including those in airborne images. The proposed model named DenseCL is designed with dense blocks and adopts a contrastive loss function as an additional regularization. The model also maintains the cycle consistency by remapping the dehazed outputs into a hazy image using the physics-based light scattering function. DenseCL has been trained with publicly available outdoor images and demonstrates outstanding performance on outdoor, indoor, and remotely sensed nonhomogeneous haze satellite images.

Mitra, Somosmita↗
DOE OSTI2021

Vehicle Tracking with Crop-based Detection

End-to-end production of vehicle tracking data from video in real-time and with high accuracy remains a challenging problem due to the computational cost of object detection on each frame. In this work we present Tracking with Crop-based Detection, a method for speeding object tracking in constrained contexts (with stable cameras and relatively-predictable object motion) such as vehicle traffic monitoring. We leverage this context to provide a strong prior for object locations, which we use to 1.) boost detection speed by detecting objects only in regions corresponding to object priors on most frames and 2.) inform the selection of the detector output for each object. We evaluate Crop-based Detection as an extension to the KIOU object tracker (Crop-KIOU) on the UA-DETRAC dataset. The proposed tracker outperforms all other reported algorithms in terms of PR-MOTA, PR-MOTP, and mostly tracked objects on the UA-DETRAC benchmark, establishing a new state-of-the-art. Relative to tracking by detection with KIOU, Crop-KIOU achieves a 26% higher frame-rate and increases accuracy. Furthermore, Tracking with Crop-based Detection can be combined with frame skipping; we show a 149% increase in framerate relative to KIOU with no decrease in accuracy using this combination of methods.

Gloudemans, Derek↗
DOE OSTI2024

Disruption of Commercial Solar Inverter System by TLS Proxy Man-in-the-Middle Attack

Transport Layer Security (TLS) is a cryptographic protocol that encrypts communication data, providing end-to-end communication encryption and authentication. Currently, TLS is widely adopted for securing communication between servers and end devices, including solar inverter systems. Therefore, users/operators can securely access the solar inverters through a web user interface (WebUI) application programmable interface (API) on a PC or server over TLS-enabled Wi-Fi or Ethernet. However, the security of the TLS-based network becomes compromised if it is breached by a TLS proxy man-in-the-middle (MITM) exploit. This report explores potential vulnerabilities in a commercial solar inverter system that leverages a TLS proxy MITM and discusses the impacts through assume-breached penetration testing. Furthermore, the paper explores recommended mitigation methods against the TLS proxy MITM exploit in solar inverters.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

A Weakly-Supervised, Multitask Deep Learning Framework for Shadow Mitigation in Remote Sensing Imagery

We propose a weakly-supervised, multitask framework for training a convolutional neural network to solve the problem of cloud shadow mitigation given only cloud and shadow masks as labels. The network minimizes the Wasserstein distance between shadows and their proximal sunlit neighborhoods, generating a supervisory signal directly from within the input image. We extract further utility from the shadow mask through multitask learning by introducing an auxiliary task of shadow segmentation. Our approach is advantageous since it performs mitigation in an end-to-end framework which requires only a shadowed image for inference. We apply this process to the Landsat 8 OLI SPARCS validation data set and demonstrate plausible results.

Couwenhoven, Scott↗
DOE OSTI2020

Leveraging Fine-Grained Occupancy Estimation Patterns for Effective HVAC Control

As occupancy sensing technologies become mature, various occupancy sensors are increasingly deployed in commercial buildings for pervasive occupancy monitoring. These sensors provide occupant-count data, which contains rich spatiotemporal information about occupancy patterns. With long-term occupant-count data collected from a commercial building, we design three different predictive models that capture the occupancy dynamics and examine how a model predictive control of the HVAC system benefits from actual occupancy count prediction. Our analysis reveals that mispredictions of occupancy states, especially false positives and false negatives, may introduce inefficient control that leads to energy waste or user discomfort. To address this issue, we take a step further to design an adaptive model predictive controller that minimizes inefficient control actions according to misprediction types and distributions. A comprehensive evaluation is performed in OpenBuild and EnergyPlus simulators to study the effectiveness of the proposed end-to-end control strategy. The evaluation shows that the proposed solution reduces energy consumption by 29.5% while improving the average weighted occupants comfort by 86.7% in Predicted Mean Vote (PMV) over the fixed schedule strategy.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Comparative Study of Large Language Model Architectures on Frontier

Large language models (LLMs) have garnered significant attention in both the AI community and beyond. Among these, the Generative Pre-trained Transformer (GPT) has emerged as the dominant architecture, spawning numerous variants. However, these variants have undergone pre-training under diverse conditions, including variations in input data, data preprocessing, and training methodologies, resulting in a lack of controlled comparative studies. Here we meticulously examine two prominent open-sourced GPT architectures, GPT-NeoX and LLaMA, leveraging the computational power of Frontier, the world’s first Exascale supercomputer. Employing the same materials science text corpus and a comprehensive end-to-end pipeline, we conduct a comparative analysis of their training and downstream performance. Our efforts culminate in achieving state-of-the-art performance on a challenging materials science benchmark. Furthermore, we investigate the computation and energy efficiency, and propose a computationally efficient method for architecture design. To our knowledge, these pre-trained models represent the largest available for materials science. Our findings provide practical guidance for building LLMs on HPC platforms.

Yin, Junqi↗
DOE OSTI2024

Pretraining Billion-Scale Geospatial Foundational Models on Frontier

As AI workloads increase in scope, generalization capability becomes challenging for small task-specific models and their demand for large amounts of labeled training samples increases. On the contrary, Foundation Models (FMs) are trained with internet-scale unlabeled data via self-supervised learning and have been shown to adapt to various tasks with minimal fine-tuning. Although large FMs have demonstrated significant impact in natural language processing and computer vision, efforts toward FMs for geospatial applications have been restricted to smaller size models, as pretraining larger models requires very large computing resources equipped with state-of-the-art hardware accelerators. Current satellite constellations collect 100+TBs of data a day, resulting in images that are billions of pixels and multimodal in nature. Such geospatial data poses unique challenges opening up new opportunities to develop FMs. We investigate billion scale FMs and HPC training profiles for geospatial applications by pretraining on publicly available data. We studied from end-to-end the performance and impact in the solution by scaling the model size. Our larger 3B parameter size model achieves up to 30% improvement in top1 scene classification accuracy when comparing a 100M parameter model. Moreover, we detail performance experiments on the Frontier supercomputer, America's first exascale system, where we study different model and data parallel approaches using PyTorch's Fully Sharded Data Parallel library. Specifically, we study variants of the Vision Transformer architecture (ViT), conducting performance analysis for ViT models with size up to 15B parameters. By discussing throughput and performance bottlenecks under different parallelism configurations, we offer insights on how to leverage such leadership-class HPC resources when developing large models for geospatial imagery applications.

Tsaris, Aristeidis (aris)↗