SEARCH · Search NASA
Results for “DNN”
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.
Throughput-Oriented and Accuracy-Aware DNN Training with BFloat16 on GPU
Deep Neural Networks (DNNs) have transformed the field of artificial intelligence and achieved extraordinary success in many areas. The training of DNNs is commonly compute and memory-intensive, which has resulted in several optimizations in the training phase. Among them, reduced precision is a typical and widely used technique to accelerate DNN training and reduce memory requirements. However, applying a widely adopted reduced precision format such as Float16 to all involved operations in DNN training is not optimal as the use of Float16 in some operations can hurt model accuracy. Meanwhile, additional optimizations including loss scaling and autocast techniques can mitigate the accuracy loss but lead to inherent overhead and inadequate use of reduced precision. In this work, we leverage another reduced precision format, BFloat16, and introduce a throughput-oriented and accuracy-aware approach to maximize the performance potential of DNN training. Since the high throughput provided by BFloat16 format is accompanied by low precision of the floating-point representation, this approach achieves high throughput by using BFloat16 on all DNN operations and avoids the accuracy loss through a customized accuracy-aware normalization. Results show that our approach outperforms the state-of-the-art mixed-precision training by 1.21x on an NVIDIA A100 GPU.
A DNN surrogate unsteady aerodynamic model for wind turbine loads calculations
A recurrent deep-neural network (DNN) surrogate model capable of modeling the unsteady aerodynamic response and dynamic stall behavior of wind turbine blades has been developed and validated for use in engineering design codes. The model is trained using a subset of the oscillating airfoil experiments conducted at the Ohio State University wind tunnel. The predictions from our DNN model show excellent agreement with the measured data and, in all cases, a marked improvement over the state-of-the-art unsteady aerodynamic models. The DNN-based unsteady aerodynamics model was integrated with OpenFAST to perform full-turbine load computations for the NREL-5MW rotor. The largest differences are observed for the inboard stations, particularly in the pitching moment response, when using the new surrogate model compared to the other models available in OpenFAST.
DNN-based policies for stochastic AC OPF
We report a prominent challenge to the safe and optimal operation of the modern power grid arises due to growing uncertainties in loads and renewables. Stochastic optimal power flow (SOPF) formulations provide a mechanism to handle these uncertainties by computing dispatch decisions and control policies that maintain feasibility under uncertainty. Most SOPF formulations consider simple control policies such as affine policies that are mathematically simple and resemble many policies used in current practice. Motivated by the efficacy of machine learning (ML) algorithms and the potential benefits of general control policies for cost and constraint enforcement, we put forth a deep neural network (DNN)-based policy that predicts the generator dispatch decisions in real time in response to uncertainty. The weights of the DNN are learnt using stochastic primal–dual updates that solve the SOPF without the need for prior generation of training labels and can explicitly account for the feasibility constraints in the SOPF. The advantages of the DNN policy over simpler policies and their efficacy in enforcing safety limits and producing near optimal solutions are demonstrated in the context of a chance constrained formulation on a number of test cases.
Statistical learning framework for safety and failure analysis of a DNN-based autonomous aircraft system
Deep Neural Networks (DNNs) and Machine Learning technology is increasingly used for safety-critical applications in the Aerospace domain. To ensure safe operations, the DNN and the system must undergo rigorous verification and validation, including advanced statistical analyses. Performance and safety of the DNN and system behavior must not only be analyzed for the nominal case, but under numerous off-nominal and failure cases. In this paper we will describe how our statistical learning framework SYSAI can efficiently perform such analyses using the tool’s unique combination of advanced learning modeling and statistical analysis techniques. SYSAI can effectively explore the high-dimensional state and failure space of the system under test; geometrical shape detection of safety regions and boundaries support explainability of the results to the designer. In this paper, we report experiments and results obtained with a vision-based DNN control system (ACT) that is capable of autonomously steering an aircraft down a runway.
Exploring Flexible Communications for Streamlining DNN Ensemble Training Pipelines
Parallel training of a Deep Neural Network (DNN) ensemble on a cluster of nodes is a common practice to train multiple models in order to construct a model with a higher prediction accuracy. Existing ensemble training pipelines can perform a great deal of redundant operations, resulting in unnecessary CPU usage, or even poor pipeline performance. In order to remove these redundancies, we need pipelines with more communication flexibility than existing DNN frameworks can provide. This project investigates a series of designs to improve pipeline flexibility and adaptivity, while also increasing performance. We implement our designs using Tensorflow with Horovod, and test it using several large DNNs in a large scale GPU cluster, Titan supercomputer at Oak Ridge National Lab. Our results show that the CPU time spent during training is reduced by 2-11X. Furthermore, our implementation can achieve up to 10X speedups when CPU core limits are imposed. Our best pipeline also reduces the average power draw of the ensemble training process by 5-16% when compared to the baseline.
Exploring flexible communications for streamlining DNN ensemble training pipelines
Parallel training of a Deep Neural Network (DNN) ensemble on a cluster of nodes is a common practice to train multiple models in order to construct a model with a higher prediction accuracy, or to quickly tune the parameters of a training model. Existing ensemble training pipelines perform a great deal of redundant operations, resulting in unnecessary CPU usage, or even poor pipeline performance. In order to remove these redundancies, we need pipelines with more communication flexibility than existing DNN frameworks can provide. This project investigates a series of designs to improve pipeline flexibility and adaptivity, while also increasing performance. We implement our designs using Tensorflow with Horovod, and test it using several large DNNs in a large scale GPU cluster, the Titan supercomputer at Oak Ridge National Lab. Our results show that with the new flexible communication schemes, the CPU time spent during training is reduced by 2-11X. Furthermore, our implementation can achieve up to 10X speedups when CPU core limits are imposed. Our best pipeline also reduces the average power draw of the ensemble training process by 5--16% when compared to the baseline.
DNN Expert Profile: Meet the NGFP Class of 2015
After their first month on the job, we caught up with several NNSA Graduate Fellowship Program (NGFP) Fellows assigned to DNN. We learned more about their career interests and life experiences and what they have been up to in their fellowships, which started in June 2015.
TrioSim: A Lightweight Simulator for Large-Scale DNN Workloads on Multi-GPU Systems
Deep Neural Networks (DNNs) have become increasingly capable of performing tasks ranging from image recognition to content generation. The training and inference of DNNs heavily rely on GPUs, as GPUs' massively parallel architecture delivers extremely high computing capability. With the growing complexity of DNNs and the size of training datasets, training DNNs with a large number of GPUs is becoming a prevalent strategy. Researchers have been exploring how to design software and hardware systems for GPU farms to achieve the best utilization, efficiency, and DNN accuracy during training or inference. However, when designing and deploying such systems, designers usually rely on testing on physical hardware platforms equipped with many GPUs, incurring high costs that are almost prohibitive for system designers to test different configurations and designs, even for highly resourceful companies. While an alternative solution is to test on GPU simulators, they are often too slow for these l
Phase I Final Technical Report: High-Resolution Gamma-ray Imaging and Spectroscopy for Special Nuclear Material Assay (GeGI-DNN)
The Phase I program focused on the development, implementation, and evaluation of spectral and spatial algorithms to improve the intrinsic image quality and efficiency of GeGI-DNN for count-starved (low activity) measurement scenarios. As a direct result of the Phase-I achievements, Pinhole imaging efficiency was dramatically improved, Compton imaging performance was improved by incorporating a statistical treatment of the Compton image, and quantitative image analysis routines were developed and implemented.
DNN Approximation of Nonlinear Finite Element Equations
We investigate the potential of applying (D)NN ((deep) neural networks) for approximating nonlinear mappings arising in the finite element discretization of nonlinear PDEs (partial differential equations). As an application, we apply the trained DNN to replace the coarse nonlinear operator thus avoiding the need to visit the fine level discretization in order to evaluate the actions of the true coarse nonlinear operator. The feasibility of the studied approach is demonstrated in a two level FAS (full approximation scheme) used to solve a nonlinear diffusion-reaction PDE.
A Novel Approach for Real-Time Quality Monitoring in Machining of Aerospace Alloy through Acoustic Emission Signal Transformation for DNN
Gamma titanium aluminide (γ-TiAl) is considered a high-performance, low-density replacement for nickel-based superalloys in the aerospace industry due to its high specific strength, which is retained at temperatures above 800 °C. However, low damage tolerance, i.e., brittle material behavior with a propensity to rapid crack propagation, has limited the application of γ-TiAl. Any cracks introduced during manufacturing would dramatically lower the useful (fatigue) life of γ-TiAl components, making the workpiece surface’s quality from finish machining a critical component to product quality and performance. To address this issue and enable more widespread use of γ-TiAl, this research aims to develop a real-time non-destructive evaluation (NDE) quality monitoring technique based on acoustic emission (AE) signals, wavelet transform, and deep neural networks (DNN). Previous efforts have opted for traditional approaches to AE signal analysis, using statistical feature extraction and classification, which face challenges such as the extraction of good/relevant features and low classification accuracy. Hence, this work proposes a novel AI-enabled method that uses a convolutional neural network (CNN) to extract rich and relevant features from a two-dimensional image representation of 1D time-domain AE signals (known as scalograms), subsequently classifying the AE signature based on pedigreed experimental data and finally predicting the process-induced surface quality. The results of the present work show good classification accuracy of 80.83% using scalogram images, in-situ experimental data, and a VGG-19 pre-trained neural network, establishing the significant potential for real-time quality monitoring in manufacturing processes.
DNN Sentinel - March 2022 Edition, Theme -- "Prevent, Counter, Respond"
This was sponsored by NA-80, the Nuclear Response Division's primary customer, but for some reason, it's not listed as a possible selection
CECAV-DNN: Collective Ensemble Comparison and Visualization using Deep Neural Networks
Not Available
MIND-MAC: Multi-Level In-memory Quasi Non-Destructive MAC Operation in Compact 2T-nC FeRAM for Efficient DNN Accelerator
We present MIND-MAC, a compact 2T-nC FeRAM architecture that performs multi-level, quasi-non-destructive in-memory multiply–accumulate (MAC) for deep neural networks. By exploiting voltage-controlled partial domain switching in MFM capacitors and read-transistor amplification, the cell stores multi-bit weights and gates bit-serial inputs to produce an accumulated current on shared lines. We combine TCAD-extracted parasitics with experimentally calibrated ferroelectric models in SPICE to validate device-/circuit-level behavior, and validate multi-level sensing and QNRO with measurements on a fabricated 2T-3C test vehicle. An analytical system model maps MIND-MAC to a 6-GB main-memory in-memory compute (IMC) architecture and benchmarks VGG13 inference in 61.08 ms at 964.99 mJ. Results indicate high density, reduced rewrite overhead, and energy efficiency, positioning 2T-nC FeRAM as a promising IMC candidate for next-generation AI hardware.
A sequential DNN based Baseline Energy Prediction Framework with Long term Error Mitigation
In this paper,we aim to present a novel sequential framework containing two deep network architectures for baseline energy prediction, applied to a simulated commercial building dataset. We have observed that this sequential framework helps to improve long term prediction accuracy, there by mitigating prediction error accumulation over time. The proposed framework utilizes convolution layers to extract features from the input data space without changing spatial relations between variables. These features are memorized by tensor train based gated recurrent units for an accurate long-term prediction. Although, architecture comprising the amalgamation of convolution layers and memory cell have shown promising results in domains such as social media analysis, language modeling, video frame prediction and image recognition, this paper extends its scope to the context of energy applications. Furthermore, the addition of tensor train based gated recurrent units is motivated by the necessity of computational time reduction during the training process, there by making this framework more suitable for future in-field deployment. Lastly, results show the current framework outperforms other existing machine learning based methods, in both short-term and long-term prediction categories.
Goals, Objectives, and Requirements (GOR) of the Ground-based Nuclear Detonation Detection (GNDD) Team for the Office of Defense Nuclear Nonproliferation Research and Development (DNN R&D)
The goal, objectives, and requirements (GOR) presented in this document define a framework for describing research directed specifically by the Ground-based Nuclear Detonation Detection (GNDD) Team of the National Nuclear Security Administration (NNSA). The intent of this document is to provide a communication tool for the GNDD Team with NNSA management and with its stakeholder community. It describes the GNDD expectation that much of the improvement in the proficiency of nuclear explosion monitoring will come from better understanding of the science behind the generation, propagation, recording, and interpretation of seismic, infrasound, hydroacoustic, and radionuclide signals and development of "game-changer" advances in science and technology.