Search NASA⌕ Search

Engineering topics

Huang, Xinyi

Publications and source records attributed to Huang, Xinyi.

6 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Transcriptional regulation of the raffinose family oligosaccharides pathway in Sorghum bicolor reveals potential roles in leaf sucrose transport and stem sucrose accumulation

Bioenergy sorghum hybrids are being developed with enhanced drought tolerance and high levels of stem sugars. Raffinose family oligosaccharides (RFOs) contribute to plant environmental stress tolerance, sugar storage, transport, and signaling. To better understand the role of RFOs in sorghum, genes involved in myo -inositol and RFO metabolism were identified and relative transcript abundance analyzed during development. Genes involved in RFO biosynthesis ( SbMIPS1, SbInsPase, SbGolS1, SbRS ) were more highly expressed in leaves compared to stems and roots, with peak expression early in the morning in leaves. SbGolS, SbRS, SbAGA1 and SbAGA2 were also expressed at high levels in the leaf collar and leaf sheath. In leaf blades, genes involved in myo -inositol biosynthesis ( SbMIPS1, SbInsPase ) were expressed in bundle sheath cells, whereas genes involved in galactinol and raffinose synthesis ( SbGolS1, SbRS ) were expressed in mesophyll cells. Furthermore, SbAGA1 and SbAGA2 , genes that encode neutral-alkaline alpha-galactosidases that hydrolyze raffinose, were differentially expressed in minor vein bundle sheath cells and major vein and mid-rib vascular and xylem parenchyma. This suggests that raffinose synthesized from sucrose and galactinol in mesophyll cells diffuses into vascular bundles where hydrolysis releases sucrose for long distance phloem transport. Increased expression (>20-fold) of SbAGA1 and SbAGA2 in stem storage pith parenchyma of sweet sorghum between floral initiation and grain maturity, and higher expression in sweet sorghum compared to grain sorghum, indicates these genes may play a key role in non-structural carbohydrate accumulation in stems.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

A Visual Designer of Layer‐wise Relevance Propagation Models

Abstract Layer‐wise Relevance Propagation (LRP) is an emerging and widely‐used method for interpreting the prediction results of convolutional neural networks (CNN). LRP developers often select and employ different relevance backpropagation rules and parameters, to compute relevance scores on input images. However, there exists no obvious solution to define a “best” LRP model. A satisfied model is highly reliant on pertinent images and designers' goals. We develop a visual model designer, named as VisLRPDesigner, to overcome the challenges in the design and use of LRP models. Various LRP rules are unified into an integrated framework with an intuitive workflow of parameter setup. VisLRPDesigner thus allows users to interactively configure and compare LRP models. It also facilitates relevance‐based visual analysis with two important functions: relevance‐based pixel flipping and neuron ablation. Several use cases illustrate the benefits of VisLRPDesigner. The usability and limitation of the visual designer is evaluated by LRP users.

Huang, Xinyi↗
DOE OSTI2021

A Visual Designer of Layer-wise Relevance Propagation Models

Layer-wise Relevance Propagation (LRP) is an emerging and widely-used method for interpreting the prediction results of convolutional neural networks (CNN). LRP developers often select and employ different relevance backpropagation rules and parameters, to compute relevance scores on input images. However, there exists no obvious solution to define a "best" LRP model. A satisfied model is highly reliant on pertinent images and designers' goals. We develop a visual model designer, named as VisLRPDesigner, to overcome the challenges in the design and use of LRP models. Various LRP rules are unified into an integrated framework with an intuitive workflow of parameter setup. VisLRPDesigner thus allows users to interactively configure and compare LRP models. It also facilitates relevance-based visual analysis with two important functions: relevance-based pixel flipping and neuron ablation. Several use cases illustrate the benefits of VisLRPDesigner. The usability and limitation of the visual designer is evaluated by LRP users.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Dual-Solvent Li-Ion Solvation Enables High-Performance Li-Metal Batteries

Novel electrolyte designs to further enhance the lithium (Li) metal battery cyclability are highly desirable. Here, fluorinated 1,6-dimethoxyhexane (FDMH) is designed and synthesized as the solvent molecule to promote electrolyte stability with its prolonged –CF 2 – backbone. Meanwhile, 1,2-dimethoxyethane is used as a co-solvent to enable higher ionic conductivity and much reduced interfacial resistance. Combining the dual-solvent system with 1 m lithium bis(fluorosulfonyl)imide (LiFSI), high Li-metal Coulombic efficiency (99.5%) and oxidative stability (6 V) are achieved. Using this electrolyte, 20 µm Li||NMC batteries are able to retain ≈80% capacity after 250 cycles and Cu||NMC anode-free pouch cells last 120 cycles with 75% capacity retention under ≈2.1 µL mAh -1 lean electrolyte conditions. Such high performances are attributed to the anion-derived solid-electrolyte interphase, originating from the coordination of Li-ions to the highly stable FDMH and multiple anions in their solvation environments. This work demonstrates a new electrolyte design strategy that enables high-performance Li-metal batteries with multisolvent Li-ion solvation with rationally optimized molecular structure and ratio.

25 ENERGY STORAGE↗
DOE OSTI2020

Interactive Visual Study of Multiple Attributes Learning Model of X-Ray Scattering Images

Existing interactive visualization tools for deep learning are mostly applied to the training, debugging, and refinement of neural network models working on natural images. However, visual analytics tools are lacking for the specific application of x-ray image classification with multiple structural attributes. In this paper, we present an interactive system for domain scientists to visually study the multiple attributes learning models applied to x-ray scattering images. It allows domain scientists to interactively explore this important type of scientific images in embedded spaces that are defined on the model prediction output, the actual labels, and the discovered feature space of neural networks. Users are allowed to flexibly select instance images, their clusters, and compare them regarding the specified visual representation of attributes. The exploration is guided by the manifestation of model performance related to mutual relationships among attributes, which often affect the learning accuracy and effectiveness. The system thus supports domain scientists to improve the training dataset and model, find questionable attributes labels, and identify outlier images or spurious data clusters. Case studies and scientists feedback demonstrate its functionalities and usefulness.

97 MATHEMATICS AND COMPUTING↗