Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Deep Neural Networks”

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 991 records · Page 55

Flux Improvement based on Machine Learning for the CERES FluxByCldTyp Data Product

The NASA Clouds and the Earth's Radiant Energy System (CERES) product provides over 20 years of accurately observed top-of-the-atmosphere (TOA) and surface flux data record for climate monitoring and diagnostic studies. The interaction between clouds and radiation interaction is a key factor that dominate climate feedbacks but is not well understood. To further advance our understanding of the cloud-radiation interaction, a new CERES FluxByCldTyp (FBCT) product has been developed that contains radiative fluxes by cloud-type, which can provide more stringent constraints when validating models and reveal more insight into the interactions between clouds and climate. For CERES partly cloudy and multiple cloud-type footprints, the FBCT product utilizes Moderate Resolution Imaging Spectroradiometer (MODIS) narrow-band (NB) imager channel radiances partitioned by cloud-type within a CERES footprint to estimate the cloud-type broadband fluxes. The MODIS multi-channel derived broadband fluxes were compared with the CERES observed footprint fluxes and were found to be within 1% and 2.5% for LW and SW, respectively, as well as being mostly free of cloud property dependencies. The FBCT all-sky and clear-sky monthly averaged fluxes were found to be consistent with the CERES SSF1deg product. This study takes advantage of recent progress in machine learning (ML) field by applying deep neural network algorithm to improve fluxes based on MODIS NB radiances. The preliminary study shows ML produce are an improvement over the current FBCT Edition 4 NB2BB algorithm. Furthermore, unlike Ed4 NB2BB, the new ML method convert NB radiances directly to broadband fluxes. For future Ed5, new NB radiances are proposed and used by ML to improve fluxes calculation. Prelimary results show significant LW improvement.

Moguo Sun↗

Improvement of Radiative Fluxes for the CERES FluxByCldTyp Data Product Based on Machine Learning Technique

The NASA Clouds and the Earth's Radiant Energy System (CERES) product provides over 20 years of accurately observed top-of-the-atmosphere and surface flux data for climate studies. The interaction between clouds and radiation interaction is a key factor that dominate climate feedbacks but is not well understood. To further advance our understanding of the cloud-radiation interaction, a new CERES FluxByCldTyp (FBCT) product has been developed that contains radiative fluxes by cloud-type, which can provide more stringent constraints when validating models. The FBCT product utilizes Moderate Resolution Imaging Spectroradiometer (MODIS) narrow-band (NB) imager channel radiances partitioned by cloud-type within a CERES footprint to estimate their broadband fluxes. The MODIS multi-channel derived broadband fluxes were compared with the CERES observed footprint fluxes and were found to be within 1% and 2.5% for LW and SW, respectively, as well as being mostly free of cloud property dependencies. The FBCT all-sky and clear-sky monthly averaged fluxes were found to be consistent with the CERES SSF1deg product. This study takes advantage of recent progress in machine learning (ML) field by applying deep neural network algorithm to improve fluxes based on MODIS NB radiances. The preliminary study shows ML produce are an improvement over the current FBCT Edition 4 NB2BB algorithm. Furthermore, unlike Ed4 NB2BB, the new ML method convert NB radiances directly to broadband fluxes. For future Ed5, new NB radiances are proposed and used by ML to improve fluxes calculation. Preliminary results show significant LW improvement.

Sun, Moguo↗

Improvement of Radiative Fluxes for the CERES FluxByCldTyp Data Product Based on Machine Learning Technique

The NASA Clouds and the Earth's Radiant Energy System (CERES) product provides over 20 years of accurately observed top-of-the-atmosphere (TOA) and surface flux data record for climate monitoring and diagnostic studies. The interaction between clouds and radiation interaction is a key factor that dominate climate feedbacks but is not well understood. To further advance our understanding of the cloud-radiation interaction, a new CERES FluxByCldTyp (FBCT) product has been developed that contains radiative fluxes by cloud-type, which can provide more stringent constraints when validating models and reveal more insight into the interactions between clouds and climate. For CERES partly cloudy and multiple cloud-type footprints, the FBCT product utilizes Moderate Resolution Imaging Spectroradiometer (MODIS) narrow-band (NB) imager channel radiances partitioned by cloud-type within a CERES footprint to estimate the cloud-type broadband fluxes. The MODIS multi-channel derived broadband fluxes were compared with the CERES observed footprint fluxes and were found to be within 1% and 2.5% for LW and SW, respectively, as well as being mostly free of cloud property dependencies. The FBCT all-sky and clear-sky monthly averaged fluxes were found to be consistent with the CERES SSF1deg product. This study takes advantage of recent progress in machine learning (ML) field by applying deep neural network algorithm to improve fluxes based on MODIS NB radiances. The preliminary study shows ML produce are an improvement over the current FBCT Edition 4 NB2BB algorithm. Furthermore, unlike Ed4 NB2BB, the new ML method convert NB radiances directly to broadband fluxes. For future Ed5, new NB radiances are proposed and used by ML to improve fluxes calculation. Prelimary results show significant LW improvement.

Moguo Sun↗

TruePAL – An AI Assistant for First Responder Safety

This paper presents the development of an AI assistant, Trusted and Explainable Artificial Intelligence for Saving Lives (TruePAL), to provide real-time warning of risks of potential crashes to the first responders. The TruePAL system employs an AI and deep learning technology for saving first responders and roadside crews lives in and around active traffic. A deep neural network (DNN) and a Non-Axiomatic Reasoning System (NARS) are implemented as an AI system. A mobile app with AI interface is developed to perform verbal communication with the first responders. The TruePAL team has developed an explainable AI approach by opening up the DNN blackbox to extract the activation filters of various features and parts of the targeted objects. The combination of DNN and NARS makes the TruePAL system explainable to the users. TruePAL ingests on-board cameras, radar, and other sensor signals, analyzes the environment and traffic patterns to generate timely warning to drivers and roadside crews to avoid crashes. The TruePAL team, in collaboration with the Miami/Dade Police Dept., has designed five use cases and multiple sub-scenarios in a CARLA driving simulator to test the capability of TruePAL in timely warning to the first responder drivers in potential crash scenarios. We have successfully demonstrated its capability of timely warning in over a dozen scenarios based on the use cases. The preliminary test simulation results show that TruePAL could provide the drivers and crew members advanced warning before a crash occurs.

Chow, Edward↗

Assumption Generation for the Verification of Learning-Enabled Autonomous Systems

Providing safety guarantees for autonomous systems is difficultas these systems operate in complex environments that require the use of learning-enabled components, such as deep neural networks (DNNs) for visual perception. DNNs are hard to analyze due to their size (they can have thousands or millions of parameters), lack of formal specifications (DNNs are typically learnt from labeled data, in the absence of any formal or informal requirements), and sensitivity to small changes in the environment. We present an assume-guarantee style compositional approach for the formal verification of system-level safety properties of such autonomous systems. Our insight is that we can analyze the system in the absence of the DNN perception components by automatically synthesizing assumptions on the DNN behaviour that guarantee the satisfaction of the required safety properties. The synthesized assumptions are the weakest in the sense that they characterize the output sequences of all the possible DNNs that, plugged into the autonomous system, guarantee the required safety properties. The assumptions can be leveraged as run-time monitors over a deployed DNN to guarantee the safety of the overall system; they can also be mined to extract local specifications for use during training and testing of DNNs. We illustrate our approach on a case study taken from the autonomous airplanes domain that uses a complex DNN for perception

Autonomous systems↗

Understanding Relationships Between Satellite, Model, and Ground-Based Surface Temperature Characterizations From Overcast to Clear Conditions in Support of Satellite Remote Sensing of Clouds and Radiation

Accurate and consistent global estimates of cloud coverage and their properties are fundamental to long-term Earth radiation budget (ERB) monitoring efforts like the Clouds and the Earth’s Radiant Energy System (CERES) project. Cloud detection algorithms often apply thresholding approaches to identify where clouds occur by comparing satellite-measured radiances with those that are expected under cloud-free conditions. In addition, once a cloud is detected, the derivation of cloud optical and microphysical properties also requires knowledge of the background radiances below the cloud. In the infrared, knowledge of the surface emissivity and the expected skin temperature under both cloudy and cloud-free conditions is needed. These traits are generally well known over the oceans. Over land, however, comparisons between satellite-derived land surface temperature (LST) with that characterized in numerical weather analyses reveal large differences in many parts of the world, often exceeding 5 K, which can lead to significant satellite cloud detection and cloud property retrieval errors. Furthermore, clouds have a dramatic influence on the LST, and therefore characterization of that model parameter also depends on the capability of the model to accurately resolve clouds. Thus, the LST characterized in models is, at times, a poor approximation for what would otherwise be observed, thereby impeding accurate satellite cloud retrievals. As a result, we seek to develop a more robust method for estimating the LST required for satellite cloud characterizations. This effort is accomplished through a combination of surface emission/air temperature relationship studies in all-sky conditions using ground measurement stations, along with deep neural network (DNN) estimates of expected LST under overcast and cloud-free conditions. We demonstrate that substituting DNN-predicted LST for that generated by numerical models can mitigate model-inherent diurnal dependencies and reduce overall bias and uncertainty relative to satellite/ground observations by 0.5–4 K and 0.5–2 K, respectively. It is expected that this work will lead to improved satellite cloud retrievals that enhance ERB monitoring efforts.

B Scarino↗

From Simulation to Reality With Random Noise

The challenging environment of autonomous vehicle (AV) navigation necessitates certain functions be performed by deep neural networks. Optimizing these models involves collecting vast quantities of domain-specific training data and ensuring that the dataset is representative of expected conditions. High-fidelity simulation plays a vital role in making this process feasible, allowing a wide range of scenarios to be explored at low cost. However, learning from simulation introduces subtle biases into models, which can degrade real-world performance in unpredictable ways. This effect can be mitigated with learning schemes specialized to bridge distributional shifts (transfer learning). Given the complex nature of these methods, the underlying models, and their environments, meaningfully evaluating performance is notstraight forward. Many unrelated factors can effect an improvement in generalization accuracy, but a full ablation analysis is often difficult. To tease out signal from noise, it is necessary to understand how transfer learning performance is affected by noise itself. The goals of this paper are (i) to establish a domain randomization baseline for a simple classification transfer learning task and (ii) to validate the RRAV testbed as a platform for further research in sim-to-real learning. We generate imagery from a simulation of NASA Ames Research Center and train a small convolutional neural network (ConvNet) to classify position relative to a centerline. Further models are trained with different types of noise progressively added to the data. The models are deployed aboard the on-site test vehicle to test real-world performance. In our experiments, we find that such naive domain randomization raises sim-to-real accuracy from 64% to 79%, while training directly on real data yields an 89% accuracy ceiling. These results suggest that the isolated mechanism of domain randomization can significantly improve generalization.

simulation↗

Low-cost scalable quartz crystal microbalance array for environmental sensing

Proliferation of environmental sensors for internet of things (IoT) applications has increased the need for low-cost platforms capable of accommodating multiple sensors. Quartz crystal microbalance (QCM) crystals coated with nanometer-thin sensor films are suitable for use in high-resolution (~1 ng) selective gas sensor applications. We demonstrate a scalable array for measuring frequency response of six QCM sensors controlled by low-cost Arduino microcontrollers and a USB multiplexer. Gas pulses and data acquisition were controlled by a LabVIEW user interface. We test the sensor array by measuring the frequency shift of crystals coated with different compositions of polymer composites based on poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS) while films are exposed to water vapor and oxygen inside a controlled environmental chamber. Our sensor array exhibits comparable performance to that of a commercial QCM system, while enabling high-throughput 6 QCM testing for under $1,000. We use deep neural network structures to process sensor response and demonstrate that the QCM array is suitable for gas sensing, environmental monitoring, and electronic-nose applications.

Anazagasty, Cristain↗

SATYA: Defending Against Adversarial Attacks Using Statistical HypothesisTesting

The paper presents a new defense against adversarial attacks for deep neural networks. We demonstrate the effectiveness of our approach against the popular adversarial image generation method DeepFool. Our approach uses Wald’s Sequential Probability Ratio Test to sufficiently sample a carefully chosen neighborhood around an input image to determine the correct label of the image. On a benchmark of 50,000 randomly chosen adversarial images generated by DeepFool we demonstrate that our method SATYA is able to recover the correct labels for 95.76% of the images for CaffeNet and 97.43% of the correct label for GoogLeNet.

Raj, Sunny↗

Toward a Multimodal-Deep Learning Retrieval System for Monitoring Nuclear Proliferation Activities

As the amount of available data increases, the human ability to locate, process, and analyze it is strained and eventually overwhelmed. To address this challenge for nonproliferation analysts, the authors have been creating a large-scale multimodal retrieval system to help analysts triage and search open source science, technology, and news data for indicators of nuclear proliferation capabilities and activities. The system relies on a set of deep neural networks (DNNs) trained to evaluate conceptual similarities across data modalities, such as text, image, and video. These DNNs can be used to search and prioritize data, according to a nuclear fuel cycle (NFC) process template, that are conceptually closest to the seed query items regardless of data modality. The DNNs have been trained to map conceptually related words, sentences, and images to nearby points in a multimodal feature space, enabling intra- and intermodal retrieval via nearest-neighbor calculations to seed query points. The authors evaluate the system’s ability to retrieve NFC-related data that have been purposely hidden in collections of unrelated background data. They demonstrate quantitative and qualitative results for text-to-image, image-to-image, and image-to-video retrievals. This paper discusses data challenges confronting nonproliferation analysts, describes the DNNs designed to retrieve multimodal data that are proximal in the semantic space, and demonstrates their effectiveness by applying the DNNs to nonproliferation-specific multimodal data sets.

97 MATHEMATICS AND COMPUTING↗

Autoencoder-Based Optical Wireless Communications Systems

In this study, we propose deep neural network autoencoders for capturing the end-to-end performance of the single- and multi-user optical wireless communications (OWC) systems. We compare the end-toend performance of the proposed autoencoders (learningbased OWC systems) with the state-of-the art model-based OWC systems in terms of the block error rate (BLER) metric. Our numerical results indicate that the proposed learning-based OWC system outperforms the model-based counterparts in both single- and multi-user settings.

99 GENERAL AND MISCELLANEOUS↗

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.

Pittman, Randall↗

Deep Learning-Based Queue-Aware Eco-Approach and Departure System for Plug-In Hybrid Electric Buses at Signalized Intersections: A Simulation Study

Eco-Approach and Departure (EAD) has been considered as a promising eco-driving strategy for vehicles traveling in an urban environment, where information such as signal phase and timing (SPaT) and geometric intersection description is well utilized to guide vehicles passing through intersections in the most energy-efficient manner. Previous studies formulated the optimal trajectory planning problem as finding the shortest path on a graphical model. While this method is effective in terms of energy saving, its computation efficiency can be further enhanced by adopting machine learning techniques. In this paper, we propose an innovative deep learning-based queue-aware eco-approach and departure (DLQ-EAD) system for a plug-in hybrid electric bus (PHEB), which is able to provide an online optimal trajectory for the vehicle considering both the downstream traffic condition (i.e. traffic lights, queues) and the vehicle powertrain efficiency. Based on optimal solutions obtained from the graph-based trajectory planning algorithm (GTPA), a deep neural network (DNN) is developed to learn the optimal vehicle speed for the next time step given its current state. It is demonstrated that the trained DNN can provide energy-efficient trajectories with high computational efficiency and high flexibility adopting to dynamic changes in the surrounding environment. To address the impact of downstream traffic, a queue prediction model is further developed using data from radars and connected vehicles (CVs), as well as signal timing data from SPaT messages. A comprehensive simulation study in the microscopic traffic modeling software PTV VISSIM shows that the proposed DLQ-EAD can achieve 18.7%-24.0% energy efficiency improvements for a single PHEB on various traffic congestion levels. The proposed queue prediction model can be of practical significance even at low penetration rates of CVs. Specifically, additional energy savings of 2.0%-8.2% can be further achieved with 20% vehicles in the network.

Ye, Fei↗

Enforcing constraints for time series prediction in supervised, unsupervised and reinforcement learning

We assume that we are given a time series of data from a dynamical system and our task is to learn the flow map of the dynamical system. We present a collection of results on how to enforce constraints coming from the dynamical system in order to accelerate the training of deep neural networks to represent the flow map of the system as well as increase their predictive ability. In particular, we provide ways to enforce constraints during training for all three major modes of learning, namely supervised, unsupervised and reinforcement learning. In general, the dynamic constraints need to include terms which are analogous to memory terms in model reduction formalisms. Such memory terms act as a restoring force which corrects the errors committed by the learned flow map during prediction. For supervised learning, the constraints are added to the objective function. For the case of unsupervised learning, in particular generative adversarial networks, the constraints are introduced by augmenting the input of the discriminator. Finally, for the case of reinforcement learning and in particular actor-critic methods, the constraints are added to the reward function. In addition, for the reinforcement learning case, we present a novel approach based on homotopy of the action-value function in order to stabilize and accelerate training. We use numerical results for the Lorenz system to illustrate the various constructions.

Stinis, Panagiotis↗

Machine Learning for Automated Extraction of Building Geometry

As data science comes to buildings, the promise of using machine learning and novel sources of data has received much attention. Advances in machine learning and computer vision algorithms, combined with increased access to unstructured data (e.g., images and text), have created an opportunity for automated extraction of building characteristics – cost-effectively, and at scale. Acquisition of features such as footprint are time consuming and costly to acquire with today’s manual methods, but can be streamlined through intelligent software-based solutions applied to satellite images. When combined with aerial RGB and thermal images, full 3D geometries and thermal maps can be constructed to determine additional characteristics such as window to wall ratio, height, number of stories and envelope thermal characteristics. In this paper we present three contributions to accelerate these high potential opportunities: (1) a methodical analysis of how these features can be integrated into today’s simulation and data driven software tools to enhance efficiency measure identification and owner/operator decision making; (2) development and accuracy testing of open source deep neural network methods to extract building footprints from satellite imagery, including the curation and application of openly available GIS datasets for training and continued development by others; and (3) an open framework for drone-based image capture and creation of 3D building geometries. This work represents an important bridge between high-level studies that span diverse application areas and those that detail point solutions yet cannot be easily replicated or extended.

Touzani, Samir↗

Towards Efficient Uncertainty estimation in deep learning for robust energy prediction in crystal materials

In material science, recent studies have started to explore the potential of using deep learning to improve property prediction from high-fidelity simulations, e.g, density functional theory (DFT). However, the design spaces are sometimes too large and intractable to sample completely. This results in a critical question that is how to evaluate the confidence and robustness of the prediction. In this paper, we propose an efficient approach to estimate uncertainty in deep learning using a single forward pass and then apply it for robust prediction of the total energy in crystal lattice structures. Our approach is built upon the deep kernel learning (DKL) that originally introduces to leverage the expressiveness of deep neural networks as input with a probabilistic prediction of Gaussian processes (GPs) as output. Existing DKL methods have difficulties in the accuracy of predictive uncertainty, training stability, and scaling to large datasets, which lead to significant barriers in real-world applications. We propose to address these challenges by using an inducing point approximate GP in feature space combined with spectral normalization as a regularization. We finally demonstrate our robust performance on an artificial example and a real-world application from materials chemistry.

Bi, Sirui↗

Machine Learning Tools Set for Natural Gas Fuel Cell System Design

This study is focusing on leveraging the system design tools set for the next-generation solid oxide fuel cell (SOFC) based natural gas fuel cell (NGFC) system. Conventionally, system design and optimization of NGFC systems rely heavily on traditional reduced order model (ROM) techniques and designers’ experience level. For overcoming the technical barriers of system design, multiple multi-physics models and machine learning (ML) tools have been utilized to automate the conceptual design process and enhance the reliability of solutions for the NGFC system. The proposed tools set includes a physics-informed ML tool for automated ROM construction that leverages advances in deep neural networks to significantly reduce ROM prediction error for the NGFC power island compared to traditional approaches. The constructed physics-informed ML ROM can be used in system design, and optimization tools set Institute for the Design of Advanced Energy Systems (IDAES) Process Systems Engineering (PSE) framework. The tools set also provides a user-friendly graphic user interface built within Jupyter Notebooks, and the complete tools set is open-source public available.

Wang, Dewei↗

DeepGhostBusters: Using Mask R-CNN to Detect and Mask Ghosting and Scattered-Light Artifacts from Optical Survey Images

Wide-field astronomical surveys are often affected by the presence of undesirable reflections (often known as "ghosting artifacts" or "ghosts") and scattered-light artifacts. The identification and mitigation of these artifacts is important for rigorous astronomical analyses of faint and low-surface-brightness systems. However, the identification of ghosts and scattered-light artifacts is challenging due to a) the complex morphology of these features and b) the large data volume of current and near-future surveys. In this work, we use images from the Dark Energy Survey (DES) to train, validate, and test a deep neural network (Mask R-CNN) to detect and localize ghosts and scattered-light artifacts. We find that the ability of the Mask R-CNN model to identify affected regions is superior to that of conventional algorithms and traditional convolutional neural networks methods. We propose that a multi-step pipeline combining Mask R-CNN segmentation with a classical CNN classifier provides a powerful technique for the automated detection of ghosting and scattered-light artifacts in current and near-future surveys.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗