Search NASA⌕ Search

Engineering topics

Kim, Edward

Publications and source records attributed to Kim, Edward.

At least 19 records

Event-to-Video Conversion for Overhead Object Detection

Collecting overhead imagery using an event camera is desirable due to the energy efficiency of the image sensor compared to standard cameras. However, event cameras complicate downstream image processing, especially for complex tasks such as object detection. In this paper, we investigate the viability of event streams for overhead object detection. We demonstrate that across a number of standard modeling approaches, there is a significant gap in performance between dense event representations and corresponding RGB frames. We establish that this gap is, in part, due to a lack of overlap between the event representations and the pre-training data that the object detectors were initially trained on through a number of experiments. Then, apply an off-the-shelf event-to-video conversion tool that converts event streams into gray-scale video to close this gap. We demonstrate that this approach results in a large performance increase, outperforming even event-specific object detection techniques on our overhead target task. These results suggest that better aligning event representations with existing large pre-trained models may result in greater short-term performance gains compared to end-to-end event-specific architectural improvements.

machine learning (ML), computer vision, Neuromorph↗

Dictionary Learning with Accumulator Neurons

The Locally Competitive Algorithm (LCA) uses local competition between non-spiking leaky integrator neurons to infer sparse representations, allowing for potentially real-time execution on massively parallel neuromorphic architectures such as Intel's Loihi processor. Here, we focus on the problem of inferring sparse representations from streaming video using dictionaries of spatiotemporal features optimized in an unsupervised manner for sparse reconstruction. Non-spiking LCA has previously been used to achieve unsupervised learning of spatiotemporal dictionaries composed of convolutional kernels from raw, unlabeled video. We demonstrate how unsupervised dictionary learning with spiking LCA (\hbox{S-LCA}) can be efficiently implemented using accumulator neurons, which combine a conventional leaky-integrate-and-fire (\hbox{LIF}) spike generator with an additional state variable that is used to minimize the difference between the integrated input and the spiking output. We demonstrate dictionary learning across a wide range of dynamical regimes, from graded to intermittent spiking, for inferring sparse representations of both static images drawn from the CIFAR database as well as video frames captured from a DVS camera. On a classification task that requires identification of the suite from a deck of cards being rapidly flipped through as viewed by a DVS camera, we find essentially no degradation in performance as the LCA model used to infer sparse spatiotemporal representations migrates from graded to spiking. We conclude that accumulator neurons are likely to provide a powerful enabling component of future neuromorphic hardware for implementing online unsupervised learning of spatiotemporal dictionaries optimized for sparse reconstruction of streaming video from event based DVS cameras.

artificial intelligence↗

Data-driven materials research enabled by natural language processing and information extraction

Given the emergence of data science and machine learning throughout all aspects of society, but particularly in the scientific domain, there is increased importance placed on obtaining data. Data in materials science are particularly heterogeneous, based on the significant range in materials classes that are explored and the variety of materials properties that are of interest. This leads to data that range many orders of magnitude, and these data may manifest as numerical text or image-based information, which requires quantitative interpretation. The ability to automatically consume and codify the scientific literature across domains - enabled by techniques adapted from the field of natural language processing - therefore has immense potential to unlock and generate the rich datasets necessary for data science and machine learning. This review focuses on the progress and practices of natural language processing and text mining of materials science literature and highlights opportunities for extracting additional information beyond text contained in figures and tables in articles. Here, we discuss and provide examples for several reasons for the pursuit of natural language processing for materials, including data compilation, hypothesis development, and understanding the trends within and across fields. Current and emerging natural language processing methods along with their applications to materials science are detailed. We, then, discuss natural language processing and data challenges within the materials science domain where future directions may prove valuable.

36 MATERIALS SCIENCE↗

SI-Traceable Calibration of Satellite Microwave Radiometers

Background (what): SI (International System of Units)-traceable Microwave Radiometer calibration; Motivation (why): NWP (Numerical Weather Prediction), FCDR (Fundamental Climate Data Record); Technology (how): NIST (National Inst. of Standards and Technology) blackbody target for ; Standards: Status & Future Plans. Development of a National Standard for Microwave Brightness Temperature (TB) at NIST.

MW Radiometer TB Calibration↗

Analysis of Soil Freeze/Thaw Signatures During Slapex F/T Campaign

Permanently frozen and seasonally frozen soils occur over a large portion of the Earth's land surface. Changes in the freeze/thaw state of the land surface reflects major changes in thermal and hydraulic properties as well as acting as a "switch" for many ecological processes. In short, soil freeze/thaw state is a fundamental land surface variable in the water and energy cycles, and it connects to the carbon cycle. Surface freeze/thaw state is observable by passive and active microwave sensors. For example, NASA's Soil Moisture Active Passive (SMAP) mission includes a freeze/thaw data product. Such satellite sensing offers routine all-season and all-weather global observations of soil freeze/thaw state with the application of suitable algorithms. We describe early finding from the SLAPex Freeze/Thaw campaign, believed to be the first airborne campaign of its type, focusing on soil freeze/thaw.

SLAPex↗

On-Orbit Special Testing of NOAA-20/JPSS-1 ATMS

The second Advanced Technology Microwave Sounder (ATMS) recently launched November 2017 on the Joint Polar Satellite System-l satellite (JPSS-l), now re-named NOAA-20. It joins the first ATMS flight unit aboard the Suomi NPP (S-NPP) satellite, as well as older sounders-the Advanced Microwave Sounding Units A & B (AMSU-A/B) and Microwave Humidity Sounder (MHS)-on polar-orbiting operational weather satellites. Together, these sounders provide critical all-weather temperature and humidity profile information for Numerical Weather Prediction (NWP) models. This paper presents results from a number of special post-launch tests used to characterize the instrument and provide unique calibration information. These special tests-long stares, alternate techniques for lunar intrusion mitigation and geolocation, spacecraft maneuvers, special scan modes, comparisons with NWP models-require non-standard modes of operation or data analysis, and can only be conducted during commissioning, prior to the start of regular forecast observations.

Kim, Edward↗

Materials Synthesis Insights from Scientific Literature via Text Extraction and Machine Learning

In the past several years, Materials Genome Initiative (MGI) efforts have produced myriad examples of computationally designed materials in the fields of energy storage, catalysis, thermoelectrics, and hydrogen storage as well as large data resources that are used to screen for potentially transformative compounds. The bottleneck in high-Throughput materials design has thus shifted to materials synthesis, which motivates our development of a methodology to automatically compile materials synthesis parameters across tens of thousands of scholarly publications using natural language processing techniques. To demonstrate our framework's capabilities, we examine the synthesis conditions for various metal oxides across more than 12 thousand manuscripts. We then apply machine learning methods to predict the critical parameters needed to synthesize titania nanotubes via hydrothermal methods and verify this result against known mechanisms. Lastly, we demonstrate the capacity for transfer learning by using machine learning models to predict synthesis outcomes on materials systems not included in the training set and thereby outperform heuristic strategies.

36 MATERIALS SCIENCE↗

Overview of SnowEx Year 1 Activities

SnowEx is a multi-year airborne snow campaign with the primary goal of addressing the question: How much water is stored in Earths terrestrial snow-covered regions? Year 1 (2016-17) focused on the distribution of snow-water equivalent (SWE) and the snow energy balance in a forested environment. The year 1 primary site was Grand Mesa and the secondary site was the Senator Beck Basin, both in western, Colorado, USA. Nine sensors on five aircraft made observations using a broad range of sensing techniques, active and passive microwave, and active and passive optical infrared to determine the sensitivity and accuracy of these potential satellite remote sensing techniques, along with models, to measure snow under a range of forest conditions. SnowEx also included an extensive range of ground truth measurements in-situ manual samples, snow pits, ground based remote sensing measurements, and sophisticated new techniques. A detailed description of the data collected will be given and some preliminary results will be presented.

snow↗

Pre-Launch Radiometric Performance Characterization of the Advanced Technology Microwave Sounder on the Joint Polar Satellite System-1 Satellite

The Advanced Technology Microwave Sounder (ATMS) is a space-based, cross-track radiometer for operational atmospheric temperature and humidity sounding, utilizing 22 channels over a frequency range from 23 to 183 gigahertz. The ATMS for the Joint Polar Satellite System-1 has undergone two rounds of re-work in 2014-2015 and 2016, following performance issues discovered during and following thermal vacuum chamber (TVAC) testing at the instrument and observatory level. Final shelf-level testing, including measurement of pass band characteristics and spectral response functions, was completed in December 2016. Final instrument-level TVAC testing and calibration occurred during February 2017. Here we will describe the instrument-level TVAC calibration process, and illustrate with results from the final TVAC calibration effort.

ATMS↗

NASA's SnowEx Campaign: Observing Seasonal Snow in a Forested Environment

SnowEx is a multi-year airborne snow campaign with the primary goal of addressing the question: How much water is stored in Earth's terrestrial snow-covered regions? Year 1(2016-17) focused on the distribution of snow-water equivalent (SWE) and the snow energy balance in a forested environment. The year 1 primary site was Grand Mesa and the secondary site was the Senator Beck Basin, both in western Colorado, USA. Nine sensors on five aircraft made observations using a broad range of sensing techniques - active and passive microwave, and active and passive optical/infrared - to determine the sensitivity and accuracy of these potential satellite remote sensing techniques, along with models, to measure snow under a range of forest conditions. SnowEx also included an extensive range of ground truth measurements - in-situ manual samples, snow pits, ground based remote sensing measurements, and sophisticated new techniques. A detailed description of the data collected will be given and some preliminary results will be presented.

cryosphere↗

Towards Validation of SMAP: SMAPEX-4 & -5

The L-band (1 - 2 GHz) microwave remote sensing has been widely acknowledged as the most promising method to monitor regional to global soil moisture. Consequently, the Soil Moisture Active Passive (SMAP) satellite applied this technique to provide global soil moisture every 2 to 3 days. To verify the performance of SMAP, the fourth and fifth campaign of SMAP Experiments (SMAPEx-4 -5) were carried out at the beginning of the SMAP operational phase in the Murrumbidgee River catchment, southeast Australia. The airborne radar and radiometer observations together with ground sampling on soil moisture, vegetation water content, and surface roughness were collected in coincidence with SMAP overpasses. The SMAPEx-4 and -5 data sets will benefit to SMAP post-launch calibration andvalidation under Australian land surface conditions.

Ye, Nan↗

SLAPex Freeze/Thaw 2015: The First Dedicated Soil Freeze/Thaw Airborne Campaign

Soil freezing and thawing is an important process in the terrestrial water, energy, and carbon cycles, marking the change between two very different hydraulic, thermal, and biological regimes. NASA's Soil Moisture Active/Passive (SMAP) mission includes a binary freeze/thaw data product. While there have been ground-based remote sensing field measurements observing soil freeze/thaw at the point scale, and airborne campaigns that observed some frozen soil areas (e.g., BOREAS), the recently-completed SLAPex Freeze/Thaw (F/T) campaign is the first airborne campaign dedicated solely to observing frozen/thawed soil with both passive and active microwave sensors and dedicated ground truth, in order to enable detailed process-level exploration of the remote sensing signatures and in situ soil conditions. SLAPex F/T utilized the Scanning L-band Active/Passive (SLAP) instrument, an airborne simulator of SMAP developed at NASA's Goddard Space Flight Center, and was conducted near Winnipeg, Manitoba, Canada, in October/November, 2015. Future soil moisture missions are also expected to include soil freeze/thaw products, and the loss of the radar on SMAP means that airborne radar-radiometer observations like those that SLAP provides are unique assets for freeze/thaw algorithm development. This paper will present an overview of SLAPex F/T, including descriptions of the site, airborne and ground-based remote sensing, ground truth, as well as preliminary results.

BOREAS↗

Cost-Effectiveness Analysis of Intensity Modulated Radiation Therapy Versus 3-Dimensional Conformal Radiation Therapy for Preoperative Treatment of Extremity Soft Tissue Sarcomas

Purpose: Create a cost-effectiveness model comparing preoperative intensity modulated radiation therapy (IMRT) versus 3-dimensional conformal radiation therapy (3DCRT) for extremity soft tissue sarcomas. Methods and Materials: Input parameters included 5-year local recurrence rates, rates of acute wound adverse events, and chronic toxicities (edema, fracture, joint stiffness, and fibrosis). Health-state utilities were used to calculate quality-adjusted life years (QALYs). Overall treatment costs per QALY or incremental cost-effectiveness ratio (ICER) were calculated. Roll-back analysis was performed using average costs and utilities to determine the baseline preferred radiation technique. One-way, 2-way, and probabilistic sensitivity analyses (PSA) were performed for input parameters with the largest impact on the ICER. Results: Overall treatment costs were $17,515.58 for 3DCRT compared with $22,920.51 for IMRT. The effectiveness was higher for IMRT (3.68 QALYs) than for 3DCRT (3.35 QALYs). The baseline ICER for IMRT was $16,842.75/QALY, making it the preferable treatment. The ICER was most sensitive to the probability of local recurrence, upfront radiation costs, local recurrence costs, certain utilities (no toxicity/no recurrence, grade 1 toxicity/no local recurrence, grade 4 toxicity/no local recurrence), and life expectancy. Dominance patterns emerged when the cost of 3DCRT exceeded $15,532.05 (IMRT dominates) or the life expectancy was under 1.68 years (3DCRT dominates). Furthermore, preference patterns changed based on the rate of local recurrence (threshold: 13%). The PSA results demonstrated that IMRT was the preferred cost-effective technique for 64% of trials compared with 36% for 3DCRT. Conclusions: Based on our model, IMRT is the preferred technique by lowering rates of local recurrence, severe toxicities, and improving QALYs. From a third-party payer perspective, IMRT should be a supported approach for extremity soft tissue sarcomas.

62 RADIOLOGY AND NUCLEAR MEDICINE↗