skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.


Search for: All records

Creators/Authors contains: "Zhang, Mei"

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. It is challenging to deploy 3D Convolutional Neural Networks (3D CNNs) on mobile devices, specifically if both real-time execution and high inference accuracy are in demand, because the increasingly large model size and complex model structure of 3D CNNs usually require tremendous computation and memory resources. Weight pruning is proposed to mitigate this challenge. However, existing pruning is either not compatible with modern parallel architectures, resulting in long inference latency or subject to significant accuracy degradation. This paper proposes an end-to-end 3D CNN acceleration framework based on pruning/compilation co-design called Mobile-3DCNN that consists of two parts: a novel, fine-grained structured pruning enhanced by a prune/Winograd adaptive selection (that is mobile-hardware-friendly and can achieve high pruning accuracy), and a set of compiler optimization and code generation techniques enabled by our pruning (to fully transform the pruning benefit to real performance gains). The evaluation demonstrates that Mobile-3DCNN outperforms state-of-the-art end-to-end DNN acceleration frameworks that support 3D CNN execution on mobile devices, Alibaba Mobile Neural Networks and Pytorch-Mobile with speedup up to 34 × with minor accuracy degradation, proving it is possible to execute high-accuracy large 3D CNNs on mobile devices in real-time (or even ultra-real-time). 
    more » « less
    Free, publicly-accessible full text available July 22, 2026
  2. Context. Solar observations of carbon monoxide (CO) indicate the existence of lower-temperature gas in the lower solar chromosphere. We present an observation of pores, and quiet-Sun, and network magnetic field regions with CO 4.66 μm lines by the Cryogenic Infrared Spectrograph (CYRA) at Big Bear Solar Observatory. Aims. We used the strong CO lines at around 4.66 μm to understand the properties of the thermal structures of lower solar atmosphere in different solar features with various magnetic field strengths. Methods. Different observations with different instruments were included: CO 4.66 μm imaging spectroscopy by CYRA, Atmospheric Imaging Assembly (AIA) 1700 Å images, Helioseismic and Magnetic Imager (HMI) continuum images, line-of-sight (LOS) magnetograms, and vector magnetograms. The data from 3D radiation magnetohydrodynamic (MHD) simulation with the Bifrost code are also employed for the first time to be compared with the observation. We used the Rybicki-Hummer (RH) code to synthesize the CO line profiles in the network regions. Results. The CO 3-2 R14 line center intensity changes to be either enhanced or diminished with increasing magnetic field strength, which should be caused by different heating effects in magnetic flux tubes with different sizes. We find several “cold bubbles” in the CO 3-2 R14 line center intensity images, which can be classified into two types. One type is located in the quiet-Sun regions without magnetic fields. The other type, which has rarely been reported in the past, is near or surrounded by magnetic fields. Notably, some are located at the edge of the magnetic network. The two kinds of cold bubbles and the relationship between cold bubble intensities and network magnetic field strength are both reproduced by the 3D MHD simulation with the Bifrost and RH codes. The simulation also shows that there is a cold plasma blob near the network magnetic fields, causing the observed cold bubbles seen in the CO 3-2 R14 line center image. Conclusions. Our observation and simulation illustrate that the magnetic field plays a vital role in the generation of some CO cold bubbles. 
    more » « less
  3. null (Ed.)
  4. Carbon fibers (CFs) are an important engineering material due to their superior mechanical, electrical, and thermal properties. Majority of them are produced from the thermal conversion of polyacrylonitrile (PAN)-based fibers. In order to promote the CF manufacturing speed and offer the possibility to control the microstructure of the fibers, an alternative technology for carbonization of stabilized PAN fiber are explored by laser processing technology. In this work, we investigated the relationship between the laser process and the properties of fibers. Laser irradiation introduces the structural changes in the stabilized PAN fibers. The appearance of D band and G band in Raman spectrum verifies the existence of graphite structures in the laser scanned fibers. The characteristic peaks in FTIR disappear when the high laser energy condition is engaged, which indicates diminishing of non-carbon bonds. Laser treatment also introduces an obvious shrinkage in fiber diameter. The condition of laser irradiation could influence the electrical and mechanical properties of fibers. A new approach to convert stabilized PAN fiber into carbon fiber was demonstrated. 
    more » « less
  5. Abstract Two families of DNA glycosylases (YtkR2/AlkD, AlkZ/YcaQ) have been found to remove bulky and crosslinking DNA adducts produced by bacterial natural products. Whether DNA glycosylases eliminate other types of damage formed by structurally diverse antibiotics is unknown. Here, we identify four DNA glycosylases—TxnU2, TxnU4, LldU1 and LldU5—important for biosynthesis of the aromatic polyketide antibiotics trioxacarcin A (TXNA) and LL-D49194 (LLD), and show that the enzymes provide self-resistance to the producing strains by excising the intercalated guanine adducts of TXNA and LLD. These enzymes are highly specific for TXNA/LLD-DNA lesions and have no activity toward other, less stable alkylguanines as previously described for YtkR2/AlkD and AlkZ/YcaQ. Similarly, TXNA-DNA adducts are not excised by other alkylpurine DNA glycosylases. TxnU4 and LldU1 possess unique active site motifs that provide an explanation for their tight substrate specificity. Moreover, we show that abasic (AP) sites generated from TxnU4 excision of intercalated TXNA-DNA adducts are incised by AP endonuclease less efficiently than those formed by 7mG excision. This work characterizes a distinct class of DNA glycosylase acting on intercalated DNA adducts and furthers our understanding of specific DNA repair self-resistance activities within antibiotic producers of structurally diverse, highly functionalized DNA damaging agents. 
    more » « less