Search

The study of defects and strain in GaN substrates and epilayers for vertical device development using synchrotron X-ray topography techniques is presented. Synchrotron monochromatic beam X-ray topography (SMBXT) studies show that ammonothermal-grown GaN substrate wafers have the lowest dislocation densities while patterned hydride vapor phase epitaxy (HVPE) GaN reveal a starkly heterogeneous distribution of dislocations with large areas containing low threading dislocation densities and HVPE GaN substrates contain high dislocation densities. Epitaxial growth does not nucleate new defects but interactions of dislocations with point defects can cause changes in dislocation configurations. Rocking curve Analysis by Dynamical Simulation (RADS) is shown to be effective in analysis of implantation by correlating the depth profile of strain with the doping profile. Removal of lattice damage induced by implantation requires annealing with proper capping or under high pressures to limit loss of material and lattice distortion. Inductively coupled plasma (ICP) etching may introduce strain in GaN material that can be partially recovered by TBCl etching. Diffusion doping does not introduce new strains in GaN material while neutron transmutation doping at low levels can lead to curing of defects while higher levels lead to irradiation damage and nucleation of new defects.

Regularization that incorporates the linear combination of empirical loss and explicit regularization terms as the loss function has been frequently used for many machine learning tasks. The explicit regularization term is designed in different types, depending on its applications. While regularized learning often boost the performance with higher accuracy and faster convergence, the regularization would sometimes hurt the empirical loss minimization and lead to poor performance. To deal with such issues in this work, we propose a novel strategy, namely Gr adients O rthogonal D ecomposition ( GrOD ), that improves the training procedure of regularized deep learning. Instead of linearly combining gradients of the two terms, GrOD re-estimates a new direction for iteration that does not hurt the empirical loss minimization while preserving the regularization affects, through orthogonal decomposition. We have performed extensive experiments to use GrOD improving the commonly used algorithms of transfer learning [ 2 ], knowledge distillation [ 3 ], and adversarial learning [ 4 ]. The experiment results based on large datasets, including Caltech 256 [ 5 ], MIT indoor 67 [ 6 ], CIFAR-10 [ 7 ], and ImageNet [ 8 ], show significant improvement made by GrOD for all three algorithms in all cases.

As a devastating wheat disease, Fusarium head blight (FHB) causes large losses in yield and quality. The purpose of this study was to detect the genetic loci for FHB resistance based on a genome-wide association study (GWAS) using two panels genotyped by the Wheat 660 K SNP array and Wheat 90 K SNP array. Phenotypic investigation results showed that 43.6–51.6% of the test varieties showed an infection type of more than 3, indicating that substantial improvement in the level of resistance was urgently required. GWAS detected 1264 SNPs associated with FHB response distributed on all wheat chromosomes. Fourteen QTL were stably detected in two environments. gFHB-1B (582.9~607.9 Mb on chromosome 1B) and gFHB-5A (577.9~596.5 Mb on 5A) were two novel resistance loci that were detected in two wheat panels. Haplotype analysis of both gFHB-1B and gFHB-5A indicated that varieties with gFHB-1B_R and gFHB-5A_R had consistently less FHB than varieties carrying a one or no resistance QTL. The application frequency of the FHB-resistant alleles gFHB-1B and gFHB-5A showed a gradually decreasing trend in FHB resistance breeding. This study identified new genetic loci for the control of FHB and provides useful information for marker-assisted selection in wheat FHB resistance.

Weak thruster fault detection problem is investigated for Autonomous Underwater Vehicle (AUV) with being less than 10% loss of effectiveness fault in a thruster. For the weak thruster fault, the fault features are small in general, meanwhile the fault features are also difficult to be distinguished from the external disturbances including measurement noise. In this paper, a weak thruster fault detection method is developed based on Bayesian network and hidden Markov model. In the proposed method, the fault features are enhanced at first and then the fault detection is achieved based on the occurrence probability. Specifically, Bayesian network is applicable to enhance weak signals, but the prior knowledge is difficult to obtain for an AUV. To solve this problem, this paper introduces sparse decomposition into Bayesian network to enhance the weak thruster fault features. And then, a hidden Markov model is established to provide the likelihood probability between the observation sequence and the predict one. The fault detection is achieved according to comparing the probability and the threshold. Finally, pool-experimental results on an AUV are presented to verify the effectiveness of the developed fault detection method.

The high-frequency ultrasonic transducers with larger bandwidths yield excellent imaging performance in the biomedical field. However, achieving perfect acoustic impedance matching from the piezo-element to the target medium in the operating frequency spectrum is still a challenge. Conventional matching layers are mostly fabricated by only one or two uniform materials which are limited by their acoustic property. We propose a novel composite matching layer with gradient acoustic impedance based on a 1-3 gradient composite structure and multilevel matching theory. The proposed gradient-composite matching layer applied for ultrasonic transducer provides efficient impedance matching and ultrawide bandwidth which can significantly improve the quality of biomedical imaging. The active aperture size of the matching layer is 5× 5 mm

Identifying novel molecular mechanisms of exhausted CD8 T cells (Tex) is a key goal of improving immunotherapy of cancer and other diseases. However, high-throughput interrogation ofin vivoTexcan be costly and inefficient.In vitromodels of Texare easily customizable and quickly generate high cellular yield, offering an opportunity to perform CRISPR screening and other high-throughput assays. We established anin vitromodel of chronic stimulation and benchmarked key phenotypic, functional, transcriptional, and epigenetic features against bona fidein vivoTex. We leveraged this model ofin vitrochronic stimulation in combination with pooled CRISPR screening to uncover transcriptional regulators of T cell exhaustion. This approach identified several transcription factors, including BHLHE40.In vitroandin vivovalidation defined a role for BHLHE40 in regulating a key differentiation checkpoint between progenitor and intermediate subsets of Tex. By developing and benchmarking anin vitromodel of Tex, we demonstrate the utility of mechanistically annotatedin vitromodels of Tex, in combination with high-throughput approaches, as a discovery pipeline to uncover novel Texbiology.

Background Intravenous fluid therapy is important for pediatric and adult patients in intensive care units (ICUs). However, medical professionals continue to struggle to determine the most appropriate fluids to obtain the best possible outcomes for each patient. Objective We conducted a meta-analysis involving cohort studies and randomized controlled trials (RCTs) to compare the influence of balanced crystalloid solutions and normal saline among patients in ICUs. Patients and methods Studies that compared balanced crystalloid solutions and saline in ICU patients from databases including PubMed, Embase, Web of Science, and Cochrane Library were systematically searched up to July 25, 2022. The primary outcomes were mortality and renal-related outcomes, which included major adverse kidney events within 30 days (MAKE30), acute kidney injury (AKI), new receipt of renal replacement therapy (RRT), maximum creatinine increasing, maximum creatinine level, and final creatinine level ≥ 200% of baseline. Service utilization including length of hospital stay, ICU stay, ICU-free days and ventilator-free days were also reported. Results A total of 13 studies (10 RCTs and 3 cohort studies) involving 38,798 patients in ICUs met the selection criteria. Our analysis revealed that each subgroup had no significant difference in mortality outcomes among ICU patients between balanced crystalloid solutions and normal saline. A significant difference was detected between the adult groups (odds ratio [OR], 0.92; 95% confidence interval [CI], [0.86, 1.00]; p = 0.04) indicating that the AKI in the balanced crystalloid solutions group was lower than that in the normal saline group. Other renal-related outcomes, such as MAKE30, RRT, maximum creatinine increasing, maximum creatinine level, and final creatinine level ≥ 200% of baseline showed no significant difference between the two groups. Regarding secondary outcomes, the balanced crystalloid solution group had a longer ICU stay time (WMD, 0.02; 95% CI, [0.01, 0.03]; p = 0.0004 and I2 = 0%; p = 0.96) than the normal saline group among adult patients. Furthermore, children treated with balanced crystalloid solution had a shorter hospital stay time (WMD, − 1.10; 95% CI, [− 2.10, − 0.10]; p = 0.03 and I2 = 17%; p = 0.30) than those treated with saline. Conclusions Compared with saline, balanced crystalloid solutions could not reduce the risk of mortality and renal-related outcomes, including MAKE30, RRT, maximum creatinine increasing, maximum creatinine level, and final creatinine level ≥ 200% of baseline, but the solutions may reduce total AKI incidence among adult patients in ICUs. For service utilization outcomes, balanced crystalloid solutions were associated with a longer length of ICU stay in the adult group and shorter length of hospital stay in the pediatric group.

Background The centromere proteins (CENPs) are the key regulators in mitosis-related protein complexes and involved in kinetochore assembly and chromosome segregation during mitosis, yet they remain largely unexplored in hepatocellular carcinoma (HCC). Methods In our study, we analyzed the potential role of CENPs family numbers by bioinformatic analysis. We detected the expression level of Centromere protein A (CENPA) in HCC tissues and cells by qRT-PCR, IHC and western blotting. Furthermore, CCK-8, colony formation, EdU and flow cytometry assays were used to evaluated the potential function of CENPA in HCC cells. Subcutaneous and orthotopically mouse model were used to explored the oncogenic role of CENPA. Besides, we demonstrated the interaction between CENPA and YY1 by silver staining and co-immunoprecipitation (co-IP) assays. Dual luciferase reporter and ChIP-sequencing (ChIP-seq) assays were used to identify the potential binding sites of CENPA and YY1 on CCND1 and NRP2 promoter regions. Results In this study, CENPA was identified significantly up-regulated in HCC, and the CENPA overexpression was linked to the poor prognosis of HCC patients. CENPA knockdown inhibited the HCC cell proliferation and tumor growth. Mechanistically, CENPA activated YY1 transcription and cooperatively interacted with YY1 to bind at cyclin D1 (CCND1) and neuropilin 2 (NRP2) promoter regions for transcriptional activation. Moreover, we identified a lactylation modification site (K124) of CENPA, which is essential for CENPA to activate its target genes transcriptionally. Conclusions These findings indicate the synergistic mechanism between CENPA and YY1 that promotes CCND1 and NRP2 expression, which leads to HCC progression.

Utilization of an Si(331) beam conditioner together with an Si(111) double-crystal monochromator (DCM) enables the angular resolution of synchrotron X-ray topography to be increased by an order of magnitude compared with grazing-incidence topography or back-reflection topography conducted with the DCM alone. This improved technique with extremely small beam divergence is referred to as synchrotron X-ray plane-wave topography (SXPWT). This study demonstrates that the rocking curve width of 4H-SiC 0008 in PWT is only 2.5′′ and thus the lattice distortion at the scale of 1′′ will significantly affect the diffracted intensity. This work reports the ultra-high angular resolution in SXPWT which enables detailed probing of the lattice distortion outside the dislocation core in 4H-SiC, where the sign of the Burgers vector can be readily determined through comparison with ray-tracing simulations.

This paper describes gft (general fine-tuning), a little language for deep nets, introduced at an ACL-2022 tutorial. gft makes deep nets accessible to a broad audience including non-programmers. It is standard practice in many fields to use statistics packages such as R. One should not need to know how to program in order to fit a regression or classification model and to use the model to make predictions for novel inputs. With gft, fine-tuning and inference are similar to fit and predict in regression and classification. gft demystifies deep nets; no one would suggest that regression-like methods are “intelligent.”

Mesenchymal stem cells (MSCs) have shown great potential in treating autoimmune diseases due to their immunomodulatory capability, which has been verified in both animal experiments and clinical trials. Psoriasis is a chronic and remitting immune-related disease. Limited studies have demonstrated that MSCs might be an effective therapeutic approach for managing psoriasis, whose underlying mechanism remains to be elucidated. In our present study, human umbilical cord-derived MSCs (hUC-MSCs) were subcutaneously injected into mice with imiquimod (IMQ)-induced psoriasis-like skin inflammation to explore the feasibility of this cellular therapy. The severity of psoriasis-like dermatitis was evaluated by cumulative psoriasis area and severity index score and epidermal thickness of skin tissue sections. Flow cytometric analysis was utilized to detect T helper cells, regulatory T cells, and γδ T cells in skin-draining lymph nodes. Real-time quantitative polymerase chain reaction and enzyme-linked immunosorbent assay were used to assess the expression levels of psoriasis-related cytokines and chemokines in mouse dorsal skin lesions. We discovered that hUC-MSCs drastically diminished the severity of IMQ-induced psoriasis-like dermatitis and suppressed inflammatory cell response. Although the tail vein injection of hUC-MSCs was also effective, it was correlated with higher mortality owing to pulmonary embolism. By comparison, subcutaneous injection with two million hUC-MSCs was identified to be the optimal therapeutic strategy. Furthermore, we uncovered that hUC-MSCs might repress skin inflammation probably through inhibiting interleukin-17-producing γδ T cells. In conclusion, subcutaneous administration of hUC-MSCs might be a promising therapeutic approach for psoriasis. Our findings provide novel insights into the underpinning mechanism of hUC-MSC treatment in the management of psoriasis.

Machine learning (ML) tasks in Internet of Things (IoT) are sensitive to task inference accuracy. In this letter, an ML task offloading scheme is proposed to minimize the total delay of task processing in an edge-intelligence-enabled IoT scenario, while guaranteeing the accuracy requirements of tasks, and taking into account the multiple attributes of tasks, task inference accuracy, and impact of error inference on task processing delay. The problem of wireless channel allocation, and computing resource allocation is modeled along with the task offloading. Considering the high complexity of the optimization problem, we design an algorithm which decomposes the problem into a computing resource allocation sub-problem and a task offloading and channel allocation sub-problem, and then solves them separately. In extensive simulations, the superiority of our scheme is demonstrated in comparisons with 4 other schemes.

Account recovery (usually through a password reset) on many websites has mainly relied on accessibility to a registered email, due to its favorable deployability and usability. However, it makes a user's online accounts vulnerable to a single point of failure when the registered email account is compromised. While previous research focuses on strengthening user passwords, the security risk imposed by email-based password recovery has not yet been well studied. In this paper, we first conduct a measurement study to characterize the password recovery activities in the wild. Specifically, we examine the authentication and password recovery protocols from 239 traffic-heavy websites, confirming that most of them use emails for password recovery. We further scrutinize the security policy of leading email service providers and show that a significant portion of them takes no or marginal effort to protect user email accounts, leaving compromised email accounts readily available for mounting password recovery attacks. Then, we conduct case studies to assess potential losses caused by such attacks. Finally, we propose and implement a lightweight email security enhancement called Secure Email Account Recovery (SEAR) to defend against password recovery attacks by adding an extra layer of protection to password recovery emails.