RESUMEN
In polycrystalline metals, plastic deformation is accompanied by lattice rotations resulting from dislocation glide. Following these rotations in three dimensions requires nondestructive methods that so far have been limited to grain sizes at the micrometer scale. We tracked the rotations of individual grains in nanograined nickel by using three-dimensional orientation mapping in a transmission electron microscope before and after in situ nanomechanical testing. Many of the larger-size grains underwent unexpected lattice rotations, which we attributed to a reversal of rotation during unloading. This inherent reversible rotation originated from a back stress-driven dislocation slip process that was more active for larger grains. These results provide insights into the fundamental deformation mechanisms of nanograined metals and will help to guide strategies for material design and engineering applications.
RESUMEN
Real-world data often exhibits a long-tailed distribution, in which head classes occupy most of the data, while tail classes only have very few samples. Models trained on long-tailed datasets have poor adaptability to tail classes and the decision boundaries are ambiguous. Therefore, in this paper, we propose a simple yet effective model, named Dual-Branch Long-Tailed Recognition (DB-LTR), which includes an imbalanced learning branch and a Contrastive Learning Branch (CoLB). The imbalanced learning branch, which consists of a shared backbone and a linear classifier, leverages common imbalanced learning approaches to tackle the data imbalance issue. In CoLB, we learn a prototype for each tail class, and calculate an inter-branch contrastive loss, an intra-branch contrastive loss and a metric loss. CoLB can improve the capability of the model in adapting to tail classes and assist the imbalanced learning branch to learn a well-represented feature space and discriminative decision boundary. Extensive experiments on three long-tailed benchmark datasets, i.e., CIFAR100-LT, ImageNet-LT and Places-LT, show that our DB-LTR is competitive and superior to the comparative methods.
Asunto(s)
Benchmarking , Aprendizaje , Reconocimiento en PsicologíaRESUMEN
The use of single-crystal substrates as templates for the epitaxial growth of single-crystal overlayers has been a primary principle of materials epitaxy for more than 70 years. Here we report our finding that, though counterintuitive, single-crystal 2D materials can be epitaxially grown on twinned crystals. By establishing a geometric principle to describe 2D materials alignment on high-index surfaces, we show that 2D material islands grown on the two sides of a twin boundary can be well aligned. To validate this prediction, wafer-scale Cu foils with abundant twin boundaries were synthesized, and on the surfaces of these polycrystalline Cu foils, we have successfully grown wafer-scale single-crystal graphene and hexagonal boron nitride films. In addition, to greatly increasing the availability of large area high-quality 2D single crystals, our discovery also extends the fundamental understanding of materials epitaxy.
RESUMEN
Singular value decomposition (SVD) is one of the most effective algorithms in recommender systems (RSs). Due to the iterative nature of SVD algorithms, one big challenge is initialization that has a major impact on the convergence and performance of RSs. Unfortunately, existing SVD algorithms in the literature typically initialize the user and item features in a random manner; thus, data information is not fully utilized. This work addresses the challenge of developing an efficient initialization method for SVD algorithms. We propose a general neural embedding initialization framework, where a low-complexity probabilistic autoencoder neural network initializes the features of user and item. This framework supports explicit and implicit feedback data sets. The design details of our proposed framework are elaborated and discussed. Experimental results show that RSs based on our proposed initialization framework outperform the state-of-the-art methods in rating prediction. Moreover, regarding item ranking, our proposed framework shows an improvement of at least 2.20% ~5.74% than existing SVD algorithms and other matrix factorization methods in the literature.
RESUMEN
The Hall-Petch relationship, according to which the strength of a metal increases as the grain size decreases, has been reported to break down at a critical grain size of around 10 to 15 nanometres1,2. As the grain size decreases beyond this point, the dominant mechanism of deformation switches from a dislocation-mediated process to grain boundary sliding, leading to material softening. In one previous approach, stabilization of grain boundaries through relaxation and molybdenum segregation was used to prevent this softening effect in nickel-molybdenum alloys with grain sizes below 10 nanometres3. Here we track in situ the yield stress and deformation texturing of pure nickel samples of various average grain sizes using a diamond anvil cell coupled with radial X-ray diffraction. Our high-pressure experiments reveal continuous strengthening in samples with grain sizes from 200 nanometres down to 3 nanometres, with the strengthening enhanced (rather than reduced) at grain sizes smaller than 20 nanometres. We achieve a yield strength of approximately 4.2 gigapascals in our 3-nanometre-grain-size samples, ten times stronger than that of a commercial nickel material. A maximum flow stress of 10.2 gigapascals is obtained in nickel of grain size 3 nanometres for the pressure range studied here. We see similar patterns of compression strengthening in gold and palladium samples down to the smallest grain sizes. Simulations and transmission electron microscopy reveal that the high strength observed in nickel of grain size 3 nanometres is caused by the superposition of strengthening mechanisms: both partial and full dislocation hardening plus suppression of grain boundary plasticity. These insights contribute to the ongoing search for ultrastrong metals via materials engineering.
RESUMEN
AZ31 Mg alloy was processed by accumulative roll-bonding (ARB) and hot rolling (HR), respectively, followed by annealing. Layered bimodal structures characterized by an alternative distribution of fine-grained layers and coarse-grained layers were obtained in the ARB samples, while mixed bimodal structures were achieved in the HR samples. The ARB samples have superior combinations of high strength and good elongation compared to the HR samples, indicating a clear effect of layered bimodal structures on mechanical properties of the alloy. The strength of the ARB samples is related to the grain size; while the ductility is attributed to the activity of non-basal slip and the strong backstress.
RESUMEN
To obtain some basic laws for bird-strike resistance of composite materials in aeronautical application, the high-velocity impact behaviors of composite laminates with different materials were studied by numerical methods. The smoothed particle hydrodynamics (SPH) and finite element method (FEM) coupling models were validated from various perspectives, and the numerical results were comparatively investigated. Results show that the different composite materials have relatively little effect on projectile deformations during the bird impact. However, the impact-damage distributions can be significantly different for different composite materials. The strength parameters and fracture energy parameters play different roles in different damage modes. Lastly, modal frequency was tentatively used to explain the damage behavior of the composite laminates, for it can manifest the mass and stiffness characteristics of a dynamic structure. The dynamic properties and strength properties jointly determine the impact-damage resistance of composite laminates under bird strike. Future optimization study can be considered from these two aspects.
RESUMEN
This corrects the article DOI: 10.1038/srep35345.
RESUMEN
Strength and ductility are mutually exclusive if they are manifested as consequence of the coupling between strengthening and toughening mechanisms. One notable example is dislocation strengthening in metals, which invariably leads to reduced ductility. However, this trend is averted in metastable austenitic steels. A one-step thermal mechanical treatment (TMT), i.e. hot rolling, can effectively enhance the yielding strength of the metastable austenitic steel from 322 ± 18 MPa to 675 ± 15 MPa, while retaining both the formability and hardenability. It is noted that no boundaries are introduced in the optimized TMT process and all strengthening effect originates from dislocations with inherited thermal stability. The success of this method relies on the decoupled strengthening and toughening mechanisms in metastable austenitic steels, in which yield strength is controlled by initial dislocation density while ductility is retained by the capability to nucleate new dislocations to carry plastic deformation. Especially, the simplicity in processing enables scaling and industrial applications to meet the challenging requirements of emissions reduction. On the other hand, the complexity in the underlying mechanism of dislocation strengthening in this case may shed light on a different route of material strengthening by stimulating dislocation activities, rather than impeding motion of dislocations.
RESUMEN
Tuberculosis, caused by Mycobacterium tuberculosis complex (MTBC) infections, is one of the most widespread infectious diseases worldwide. Nontuberculous mycobacteria (NTM) also cause chronic pulmonary infections, however, NTM infection is generally overlooked.This study analyzed the frequencies of MTBC and NTM clinical isolates from 181,132 specimens obtained from patients in Taiwan suspected of having a pulmonary mycobacterial infection from 2002 to 2014. The resistant rates to 4 first-line antibiotics (isoniazid, ethambutol, rifampicin, and streptomycin) of 9079 clinical MTBC isolates were also examined by the modified agar proportion method.Overall, the mycobacterial isolation rate was 8.65%, and this consisted of MTBC isolation rate of 5.01% and NTM isolation rate of 3.63%. The prevalence of MTBC isolates among the identified mycobacterial strains could be seen to decrease significantly from 82.5% in 2002 to 41.18% in 2014. Notably, the corresponding NTM prevalence increased 3.36 fold from 17.54% in 2002 to 58.82% in 2014. The frequencies of MTBC and NTM isolates showed a reciprocal trend with the crossing over occurring in the years 2010 and 2011. Although the resistance rates of the MTBC isolates to isoniazid and streptomycin were relatively stable over the study period, resistance rates of the MTBC isolates against rifampicin and ethambutol fluctuated across the study period. Overall, the incidence of multidrug resistance was relatively consistent at about 1.74%.The diagnosis, identification, and susceptibility tests for NTM should be standardized and integrated into appropriate clinical settings to cope with the increase in NTM infections. In addition, the documentation of the antibiotic resistance rates of MTBC clinical isolates to the antibiotic treatments most often clinically prescribed over a decade provides valuable clues and reference points for effective mycobacterial control.
Asunto(s)
Antibióticos Antituberculosos/uso terapéutico , Infecciones por Mycobacterium no Tuberculosas/epidemiología , Tuberculosis Resistente a Múltiples Medicamentos/epidemiología , Tuberculosis Pulmonar/epidemiología , Estudios Transversales , Humanos , Incidencia , Pruebas de Sensibilidad Microbiana , Infecciones por Mycobacterium no Tuberculosas/tratamiento farmacológico , Taiwán , Tuberculosis Resistente a Múltiples Medicamentos/tratamiento farmacológico , Tuberculosis Pulmonar/tratamiento farmacológicoRESUMEN
INTRODUCTION: Rapidly growing mycobacteria (RGM) can cause a broad spectrum of both community and healthcare-associated infections in humans. The aim of this study was to report the clinical management and outcomes of successive patients following cesarean delivery with healthcare-associated surgical site infections (SSIs) caused by RGM. METHODOLOGY: Patients who were admitted to Chung Shan Medical University Hospital, Taichung, Taiwan, between September 2006 and July 2008, and who developed SSIs following cesarean delivery at an obstetrics hospital and were then referred to our hospital, were enrolled. Demographic characteristics of the patients and clinical isolates were obtained retrospectively and an environmental investigation was performed. PCR-restriction fragment length polymorphism (PCR-RFLP) analysis of the hsp65gene and pulsed-field gel electrophoresis (PFGE) of large genomic DNA restriction fragments were applied to differentiate Mycobacterium species. RESULTS: Seventeen patients were diagnosed with RGM infections by microbiology and/or histopathology. Mycobacterial isolates by PCR-RFLP analysis from 15 patients revealed Mycobacterium abscessus (M. abscessus) and M. lentiflavum. Most of the patients received surgical debridement and combination antimicrobial therapy and were eventually cured. CONCLUSIONS: Our study demonstrates the potential that RGM infections have in causing healthcare-associated SSIs. Surgery plus prolonged combination antimicrobial therapy seemed to be an effective option for the management of M. abscessus infections.
Asunto(s)
Cesárea/efectos adversos , Infecciones por Mycobacterium no Tuberculosas/prevención & control , Micobacterias no Tuberculosas/efectos de los fármacos , Micobacterias no Tuberculosas/aislamiento & purificación , Infección de la Herida Quirúrgica/microbiología , Adulto , Antiinfecciosos/farmacología , Proteínas Bacterianas/genética , Chaperonina 60/genética , ADN Bacteriano/aislamiento & purificación , Electroforesis en Gel de Campo Pulsado , Femenino , Humanos , Evaluación del Resultado de la Atención al Paciente , Reacción en Cadena de la Polimerasa , Polimorfismo de Longitud del Fragmento de Restricción , Estudios Retrospectivos , Infección de la Herida Quirúrgica/prevención & control , Taiwán , Adulto JovenRESUMEN
Non-tuberculous mycobacteria (NTM) can cause chronic pulmonary infection, however, NTM infection is generally overlooked. This retrospective study analyzed the frequencies of Mycobacterium tuberculosis complex (MTBC) and NTM clinical isolates from 99 200 specimens of patients suspected with pulmonary mycobacterial infection in Taiwan from 2002-2007. A total of 8024 mycobacterial isolates, including 5349 MTBC and 2675 NTM, were obtained from the 99 200 specimens in the study period. The overall mycobacterial isolation rate was 8.09% (8024/99 200), and the overall MTBC and NTM isolation rate was 5.39% (5349/99 200) and 2.7% (2675/99 200), respectively. Notably, the prevalence of NTM isolates among the identified mycobacteria strains was increased 2.6 fold from 2002 (17.54%, 147/838) to 2007 (45.80%, 659/1439). The frequencies of MTBC and NTM isolates showed a reciprocal trend: the NTM isolation rates were steadily increasing while the overall mycobacterial isolation rates remained stable over the study period. Our results suggest that the diagnosis, identification and susceptibility tests for NTM should be standardized and integrated in clinical routines, for providing the information of NTM infection and prescribing clinical treatment in a more precise and efficient way to reduce the increasing NTM in the studied area.