Search NASA⌕ Search

Engineering topics

Cao, Yu

Publications and source records attributed to Cao, Yu.

At least 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Noise-Based Selection of Robust Inherited Model for Accurate Continual Learning

There is a growing demand for an intelligent system to continually learn knowledge from a data stream. Continual learning requires both the preservation of previous knowledge (i.e., avoiding catastrophic forgetting) and the acquisition of new knowledge. Different from previous works that focus only on model adaptation (e.g., regularization, network expansion, memory rehearsal, etc.), we propose a novel training scheme named acquisitive learning (AL), which emphasizes both the knowledge inheritance and knowledge acquisition. AL starts from an elaborately selected model with pre-trained knowledge (the inherited model) and then adapts it to new data using segmented training. The selection is achieved by injecting random noise to various inherited models for better model robustness, which promises higher accuracy in further knowledge acquisition. The approach is validated by the visualization of the loss landscape and quantitative roughness measurement. The combination of the selective inherited model and knowledge acquisition reduces catastrophic forgetting by 10X on the CIFAR-100 dataset.

Du, Xiaocong↗
DOE OSTI2019

Single-Net Continual Learning with Progressive Segmented Training

There is an increasing need of continual learning in dynamic systems, such as the self-driving vehicle, the surveillance drone, and the robotic system. Such a system requires learning from the data stream, training the model to preserve previous information and adapt to a new task, and generating a single-headed vector for future inference. Different from previous approaches with dynamic structures, this work focuses on a single network and model segmentation to prevent catastrophic forgetting. Leveraging the redundant capacity of a single network, model parameters for each task are separated into two groups: one important group which is frozen to preserve current knowledge, and secondary group to be saved (not pruned) for a future learning. A fixed-size memory containing a small amount of previously seen data is further adopted to assist the training. Without additional regularization, the simple yet effective approach of Progressive Segmented Training (PST) successfully incorporates multiple tasks and achieves the state-of-the-art accuracy in the single-head evaluation on CIFAR-10 and CIFAR-100 datasets. Moreover, the segmented training significantly improves computation efficiency in continual learning at the edge.

Du, Xiaocong↗
DOE OSTI2019

Monolithic Integration of Self-Biased C -Band Circulator on SiC Substrate for GaN MMIC Applications

We have designed and fabricated the first self-biased circulator operating in C-band and monolithically integrated it with Qorvo's GaN MMIC technology by embedding a FeNi-based magnetic nanowire composite (MNC) in a 100 $μ$m thick SiC substrate. This integrated microstrip circulator shows a circulation frequency centered at 5.7 GHz, with an insertion loss of 2.7 dB and isolation of 14 dB. Additionally, the circulator demonstrated power handling of 9.6 W with continuous wave and at least 40W under pulsed conditions. This letter points a path in integrating miniaturized circulators into full-duplex GaN T/R MMICs with stringent form factor limits.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Observation of Microheterogeneity in Highly Concentrated Nonaqueous Electrolyte Solutions

The development of models to describe structure and dynamics of nonaqueous electrolyte solutions is challenging, and experimental observations are needed to form a foundation. In this work, neutron scattering is used to probe molecular dynamics in nonaqueous organic electrolytes. Two solutions were compared: one contained symmetrical electrolyte molecules prone to crystallize, and one contained desymmetrized electrolyte molecules preferring disordered states. For the latter, calorimetry and neutron data show that a disordered fluid persists to very low temperatures at high concentrations. Upon heating, localized cold crystallization occurs, leading to burst nucleation of microcrystalline solids within fluid phases. Our results suggest molecular clustering and point to solvation inhomogeneities and molecular crowding in these concentrated fluids.

25 ENERGY STORAGE↗
DOE OSTI2019

Optimization of Tensile and Corrosion Properties of Dissimilar Friction Stir Welded AA2024-7075 Joints

In this work, we performed some joining experiments of dissimilar aluminum alloys AA2024-T351 and AA7075-T651 using friction stir welding (FSW) technique with different process parameters (rotational speed, welding speed and plunge depth). Response surface methodology based on a central composite rotatable design was used to establish the mathematical models predicting the tensile properties and corrosion rate of dissimilar FSW joints with some high confidence level. Analysis of variance method was employed to verify the developed models. Besides, the influences of FSW process parameters on tensile and corrosion properties of joints were analyzed. In order to obtain a desirable combination property of the dissimilar FSW joints, the optimum process parameters after experimental verification were proposed as: rotational speed 1495 rev/min, welding speed 187 mm/min and plunge depth 0.03 mm. This verification illustrated that the developed models are appropriate for the modeling and optimization of process.

36 MATERIALS SCIENCE↗
DOE OSTI2018

Tunnel barrier schottky

A diode includes: a semiconductor substrate; a cathode metal layer contacting a bottom of the substrate; a semiconductor drift layer on the substrate; a graded aluminum gallium nitride (AlGaN) semiconductor barrier layer on the drift layer and having a larger bandgap than the drift layer, the barrier layer having a top surface and a bottom surface between the drift layer and the top surface, the barrier layer having an increasing aluminum composition from the bottom surface to the top surface; and an anode metal layer directly contacting the top surface of the barrier layer.

25 ENERGY STORAGE↗
DOE OSTI2018

Tunnel barrier schottky

A diode includes: a semiconductor substrate; a cathode metal layer contacting a bottom of the substrate; a semiconductor drift layer on the substrate; a graded aluminum gallium nitride (AlGaN) semiconductor barrier layer on the drift layer and having a larger bandgap than the drift layer, the barrier layer having a top surface and a bottom surface between the drift layer and the top surface, the barrier layer having an increasing aluminum composition from the bottom surface to the top surface; and an anode metal layer directly contacting the top surface of the barrier layer.

42 ENGINEERING↗
DOE OSTI2017

Suppression of OGT by microRNA24 reduces FOXA1 stability and prevents breast cancer cells invasion

O-GlcNAc transferase (OGT) catalyzes the addition of O-GlcNAc to certain serine or threonine residue on a wide variety of cytosolic and nuclear proteins and regulates cellular activities such as signaling and transcription. Although there are emerging evidences that OGT plays important roles in breast cancer metastasis, the underlying mechanism is not fully understood. In this study, we demonstrated that up-regulation of OGT correlates with breast cancer cells invasion. Over-expression of OGT stimulates cells invasion, while OGT silence exhibits the opposite effects. OGT is further identified as a target of microRNA24 (miR24). miR24 down-regulates OGT expression and subsequently suppresses cells invasion. Re-expression of OGT significantly rescues miR24-mediated invasion repression. Furthermore, our data showed that FOXA1 is subjected to O-GlcNAcylation, which instabilizes FOXA1 protein and promotes breast cancer cells invasion. In conclusion, our results demonstrated that miR24 inhibits breast cancer cells invasion by targeting OGT and reducing FOXA1 stability. These results also indicated that OGT might be a potential target for the diagnosis and therapy of breast cancer metastasis.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2017

Emodin suppresses TGF-β1-induced epithelial-mesenchymal transition in alveolar epithelial cells through Notch signaling pathway

Pulmonary fibrosis is characterized by the destruction of lung tissue architecture and the formation of fibrous foci, currently has no satisfactory treatment. Emodin is a component of Chinese herb that has been reported to be medicament on pancreatic fibrosis and liver fibrosis. However, its role in pulmonary fibrosis has not been established yet. In the present study, we investigated the hypothesis that Emodin plays an inhibitory role in TGF-β1 induced epithelial-mesenchymal transition (EMT) of alveolar epithelial cell, and Emodin exerts its effect through the Notch signaling pathway. Emodin inhibits the proliferation of Rat alveolar type II epithelial cells RLE-6TN in a concentration-dependent manner; reduces the expression of Collagen I, α-SMA and Vimentin, promotes the expression of E-cadherin. Moreover, Emodin could regulate the expression patterns of the Notch signaling pathway-related factors and reduce the Notch-1 nucleus translocation. Knockdown of Notch-1 enhances the inhibitory effect of Emodin on TGF-β1-induced EMT in RLE-6TN cells. In conclusion, the data of the present study suggests that Emodin suppresses TGF-β1-induced EMT in alveolar epithelial cells through Notch signaling pathway and shows the potential to be effective in the treatment of pulmonary fibrosis. - Highlights: • Emodin inhibits TGF-β1-induced EMT in alveolar epithelial cells. • Emodin regulates the expression patterns of the Notch signaling pathway-related factors. • Emodin inhibits TGF-β1-induced Notch-1 nucleus translocation and activation.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

Performance Analysis Tool for HPC and Big Data Applications on Scientific Clusters

Big data is prevalent in HPC computing. Many HPC projects rely on complex workflows to analyze terabytes or petabytes of data. These workflows often require running over thousands of CPU cores and performing simultaneous data accesses, data movements, and computation. It is challenging to analyze the performance involving terabytes or petabytes of workflow data or measurement data of the executions, from complex workflows over a large number of nodes and multiple parallel task executions. To help identify performance bottlenecks or debug the performance issues in large-scale scientific applications and scientific clusters, we have developed a performance analysis framework, using state-ofthe- art open-source big data processing tools. Our tool can ingest system logs and application performance measurements to extract key performance features, and apply the most sophisticated statistical tools and data mining methods on the performance data. It utilizes an efficient data processing engine to allow users to interactively analyze a large amount of different types of logs and measurements. To illustrate the functionality of the big data analysis framework, we conduct case studies on the workflows from an astronomy project known as the Palomar Transient Factory (PTF) and the job logs from the genome analysis scientific cluster. Our study processed many terabytes of system logs and application performance measurements collected on the HPC systems at NERSC. The implementation of our tool is generic enough to be used for analyzing the performance of other HPC systems and Big Data workows.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

High-performance radio frequency transistors based on diameter-separated semiconducting carbon nanotubes

In this paper, we report the high-performance radio-frequency transistors based on the single-walled semiconducting carbon nanotubes with a refined average diameter of ∼1.6 nm. These diameter-separated carbon nanotube transistors show excellent transconductance of 55 μS/μm and desirable drain current saturation with an output resistance of ∼100 KΩ μm. An exceptional radio-frequency performance is also achieved with current gain and power gain cut-off frequencies of 23 GHz and 20 GHz (extrinsic) and 65 GHz and 35 GHz (intrinsic), respectively. These radio-frequency metrics are among the highest reported for the carbon nanotube thin-film transistors. This study provides demonstration of radio frequency transistors based on carbon nanotubes with tailored diameter distributions, which will guide the future application of carbon nanotubes in radio-frequency electronics.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2016

Cross-Domain Recognition by Identifying Joint Subspaces of Source Domain and Target Domain

This paper introduces here a new method to solve the cross-domain recognition problem. Different from the traditional domain adaption methods which rely on a global domain shift for all classes between the source and target domains, the proposed method is more flexible to capture individual class variations across domains. By adopting a natural and widely used assumption that the data samples from the same class should lay on an intrinsic low-dimensional subspace, even if they come from different domains, the proposed method circumvents the limitation of the global domain shift, and solves the cross-domain recognition by finding the joint subspaces of the source and target domains. Specifically, given labeled samples in the source domain, we construct a subspace for each of the classes. Then we construct subspaces in the target domain, called anchor subspaces, by collecting unlabeled samples that are close to each other and are highly likely to belong to the same class. The corresponding class label is then assigned by minimizing a cost function which reflects the overlap and topological structure consistency between subspaces across the source and target domains, and within the anchor subspaces, respectively. We further combine the anchor subspaces to the corresponding source subspaces to construct the joint subspaces. Subsequently, one-versus-rest support vector machine classifiers are trained using the data samples belonging to the same joint subspaces and applied to unlabeled data in the target domain. We evaluate the proposed method on two widely used datasets: 1) object recognition dataset for computer vision tasks and 2) sentiment classification dataset for natural language processing tasks. Comparison results demonstrate that the proposed method outperforms the comparison methods on both datasets.

97 MATHEMATICS AND COMPUTING↗