Quality prediction, as the basis of quality control, is dedicated to predicting quality indices of the manufacturing process. In recent years, data-driven deep learning methods have received a lot of ...attention due to their accuracy, robustness, and convenience for the prediction of quality indices. However, the existing studies mainly focus on the quality prediction of a single machine, while ignoring dependency relationships among multiple machines in multistage manufacturing process. To tackle the above issues, a novel path enhanced bidirectional graph attention network (PGAT) is proposed in this article. PGAT models the dependencies among machines into directed graphs and introduces graph attention network to encode the dependencies. Nonetheless, it is difficult for graph neural networks to encode long-distance dependencies. Hence, dependency path information is introduced into the features of machines. Moreover, in order to solve the label noise problem that often occurs in actual industrial dataset, a masked loss function is devised. With its help, batch training with noisy labels can be achieved, which improves the training efficiency. Furthermore, experiments are conducted on a public quality prediction dataset collected from an actual production line. PGAT achieves the state-of-the-art results on this dataset, which confirms the effectiveness of PGAT. Additionally, the experimental results demonstrate the critical role of modeling dependency relationships among machines.
Field effect relies on the nonlinear current–voltage relation in semiconductors; analogously, materials that respond nonlinearly to an optical field can be utilized for optical modulation. For ...instance, nonlinear optical (NLO) materials bearing a saturable absorption (SA) feature an on–off switching behavior at the critical pumping power, thus enabling ultrafast laser pulse generation with high peak power. SA has been observed in diverse materials preferably in its nanoscale form, including both gaped semiconductor nanostructures and gapless materials like graphene; while the presence of optical bandgap and small carrier density have limited the active spectral range and intensity. We show here that solution-processed plasmonic semiconductor nanocrystals exhibit superbroadband (over 400 THz) SA, meanwhile with large modulation depth (∼7 dB) and ultrafast recovery (∼315 fs). Optical modulators fabricated using these plasmonic nanocrystals enable mode-locking and Q-switching operation across the near-infrared and mid-infrared spectral region, as exemplified here by the pulsed lasers realized at 1.0, 1.5, and 2.8 μm bands with minimal pulse duration down to a few hundreds of femtoseconds. The facile accessibility and superbroadband optical nonlinearity offered by these nonconventional plasmonic nanocrystals may stimulate a growing interest in the exploiting of relevant NLO and photonic applications.
This paper proposes a multi-head neural network (MHNN) model with unsymmetrical constraints for remaining useful life (RUL) prediction of industrial equipment. Generally, the existing deep learning ...methods proposed for RUL prediction utilize symmetrical constraint loss functions such as the mean squared error function to calculate training errors. However, if the predicted RUL is much larger than the actual value in some safety–critical applications, severe damage may occur. To address this issue, an unsymmetrical constraint function is proposed as the loss function in this work that penalizes the late predictions (i.e., the predicted RUL is larger than the actual RUL) more strongly. In addition, an adjustable parameter is added to this function to adjust the model’s attention to the late predictions. In MHNN model, the bidirectional gated recurrent units (BGRU) and self-attention mechanism are employed to extract temporal features from the condition monitoring data. In addition, the structure of the multi-head neural network is adopted in the proposed model, helping to capture more degradation information by means of multiple identical and parallel networks. The proposed method is validated against a commonly used turbofan engine dataset. Compared with other latest methods on the same dataset, the proposed method is proven to be superior. Taking the FD004 dataset as an example, the score obtained by MHNN is 24.09% lower than that obtained by the best existing method.
•Four metamorphic stages have been identified within the Daqingshan pelitic granulites.•Phase equilibria modeling has constrained a clockwise P–T path.•The Daqingshan Complex was involved in ...collisional orogen between the Yinshan and the Ordos Blocks.
Pelitic granulites are cropped out in the Daqingshan Complex of the Khondalite Belt, a Paleoproterozoic tectonic belt in the Western Block of the North China Craton. Petrological studies show that these granulites contain four distinct metamorphic assemblages. The core of a garnet porphyroblast, along with fine-grained inclusions of biotite+plagioclase+K-feldspar+quartz±sillimanite±ilmenite, defines the prograde metamorphic (M1) stage. The peak (M2) assemblage consists of garnet (mantle)+biotite+K-feldspar+plagioclase+quartz±sillimanite±orthopyroxene±ilmenite±magnetite±rutile. Peak metamorphism was followed by a near-isothermal decompression (M3) and the development of coronae of garnet+biotite+cordierite+plagioclase+quartz±K-feldspar±sillimanite±spinel±ilmenite±magnetite (M3) in the Crd–Grt–Bt gneisses during the following garnet breakdown reactions: garnet+sillimanite+melt→cordierite+biotite+Fe-oxide and garnet+melt→biotite+quartz+plagioclase. Retrograde cooling (M4) assemblages are represented by biotite+muscovite+sillimanite+quartz+plagioclase+K-feldspar±ilmenite. Quantitative phase equilibria modeling in the system Na2O–CaO–K2O–FeO–MgO–Al2O3–SiO2–H2O–TiO2–Fe2O3 was applied to obtain P–T conditions of <780°C and <9kbar for M1, 840–880°C and 9–11kbar for M2, 800–870°C and 5.0–7.5kbar for M3, and <660°C and 4.1–6.9kbar for M4. The combination of the mineral assemblages, mineral compositions, and metamorphic reaction histories in the Daqingshan pelitic granulites defines a clockwise P–T path that involves periods of near-isothermal decompression and late cooling that followed the peak granulite-facies metamorphism. This result is consistent with the tectonic history of the Daqingshan Complex in the Khondalite Belt, which involved continent–continent subduction and collision followed by exhumation and cooling. This suggests a continent–continent collisional event between the Yinshan and Ordos blocks, which became amalgamated to form the Paleoproterozoic Khondalite Belt in the Western Block of the North China Craton.
•The inherited metamorphic cores and retrogressive rims can be identified in single zircon grains.•Retrogressive metamorphic mineral inclusions have been identified in zircon grains.•A ...near-isothermal decompressional clockwise P–T–t path for the Jiaobei HP mafic granulites is defined.
High-pressure (HP) mafic granulites in the Jiaobei terrane are composed predominantly of garnet mafic granulites, garnet–hypersthene granulites, and garnet amphibolites, and they are found as irregular lenses or deformed dike swarms within tonalitic–trondhjemitic–granodioritic gneisses and granitic gneisses. The HP mafic granulites contain four distinct metamorphic assemblages, of which the early prograde assemblage (M1) is represented by the cores of garnets, together with mineral inclusions of clinopyxene+plagioclase±quartz, and it formed at 740–770°C and 0.90–1.00GPa. In contrast, the peak assemblage (M2) consists of high-Ca cores in garnet, high-Al cores in clinopyroxene, and high-Na cores in plagioclase in the matrix, which formed under P–T conditions of 850–880°C and 1.45–1.65GPa. The peak metamorphism was followed by near-isothermal decompression (M3), which resulted in the development of orthopyroxene+clinopyxene+plagioclase±quartz±amphibole±magnetitie symplectites or coronas surrounding some garnet grains, with P–T conditions of 780–830°C and 0.65–0.85GPa. Surrounding some garnet grains are symplectites of amphibole+plagioclase+quartz±magnetitie, which formed during a cooling retrograde stage (M4) with P–T conditions of 590–650°C and 0.62–0.82GPa. An integrated study involving laser Raman analysis of mineral inclusions, cathodoluminescence imaging, and in situ U–Pb dating of zircons shows that the protolith ages of the HP mafic granulites are mainly 2550–2500Ma, and that the timing of the peak metamorphism of the HP mafic granulites ranges from 1900 to 1860 Ma, as recorded by the cores of metamorphic zircons. On the other hand, the medium- to low-pressure granulite–amphibolite facies retrogression occurred mainly at 1860–1820Ma, as recorded by the rims of some single zircon grains and the zircon grains that contain inclusions of garnet+orthopyroxene+plagioclase+sphene. The combination of petrography, mineral compositions, metamorphic reaction history, thermobarometry, and geochronology defines a near-isothermal decompressional clockwise P–T–t path for the Jiaobei HP mafic granulites, suggesting that the Jiaobei terrane underwent initial crustal thickening during 1950–1860Ma, followed by relatively rapid exhumation, cooling, and retrogression in the period 1860–1820Ma. This tectonothermal path was probably generated by subduction and collision-related tectonic processes.
A series of third-order nonlinear optical chromophores containing identical π-bridge and electron-donating groups, but with different electron-withdrawing groups have been designed and synthesized. ...The ultraviolet absorption, density functional theory calculations, and the third-order nonlinear optical (NLO) properties of these materials have been investigated and systematically studied. A theoretical study showed that the third-order NLO properties were enhanced by the increasing electron-withdrawing ability in accordance with the decreasing energy gap between the highest occupied molecular orbital and the lowest unoccupied molecular orbital. The transferred charge to the substituent is increased and the affect on the electronic reallocation is reinforced with the increasing ability of the compounds to attract electrons. The third-order NLO susceptibilities, the second-order hyperpolarizabilities and response times of compounds were 2.829–6.034 × 10−13 esu, 0.737–2.005 × 10−31 esu, 41.92–68.71 fs Our results can be used to develop an efficient design strategy for NLO materials.
Display omitted
•A series of ferrocenyl Schiff base derivatives were prepared.•The formed chromophores exhibited large third-order nonlinearity.•An efficient design strategy for potential nonlinear optical materials is proposed.
Background
To compare three different curve‐fitting methods for intravoxel incoherent motion (IVIM) analysis in breast cancer.
Methods
Diffusion‐weighted imaging was acquired in 30 patients with ...breast cancer using seven b‐values (0–800 s/mm2). Three curve‐fitting methods were used for biexponential IVIM analysis: a. Direct estimation of D (diffusion coefficient), D* (pseudodiffusion coefficient) and f (perfusion fraction) (Method 1), b. Estimation of D first and then D* and f (Method 2), c. Estimation of D and f first and then D* (Method 3). Goodness‐of‐fit, parameter precision (coefficient of variance CV), parameter difference and correlation with relative enhancement ratio (RER) and initial area under the curve (IAUC) from dynamic contrast‐enhanced (DCE) MRI of the three methods were determined and compared.
Results
Among the three biexponential methods, Method 1 best described most of the pixels (63.20% based on R2; 44.52% based on Akaike Information Criteria). The CV of D calculated from Method 2/3 (14.95%/13.90%), the CV of D* from Method 2 (77.04%) and the CV of f from Method 3 (80.87%) were the lowest among the three methods. Significant difference was observed for each IVIM‐derived parameter calculated from all the three methods (P = 0.000–0.005). Only the perfusion‐related f value calculated from Method 2 was correlated with RER (r = 0.548; P = 0.002) or IAUC (r = 0.561; P = 0.001).
Conclusion
IVIM‐derived parameters differ depending on the calculation methods. The two‐step fitting method with D value estimation first was correlated with DCE MRI perfusion. J. Magn. Reson. Imaging 2015;42:362–370.
Early screening and diagnosis of breast cancer can not only detect hidden diseases in time, but also effectively improve the survival rate of patients. Therefore, the accurate classification of ...breast cancer images becomes the key to auxiliary diagnosis.
In this paper, on the basis of extracting multi-scale fusion features of breast cancer images using pyramid gray level co-occurrence matrix, we present a Self-Attention Random Forest (SARF) model as a classifier to explain the importance of fusion features, and can perform adaptive refinement processing on features, thus, the classification accuracy can be improved. In addition, we use GridSearchCV technique to optimize the hyperparameters of the model, which greatly avoids the limitation of artificially selected parameters.
To demonstrate the effectiveness of our method, we perform validation on the breast cancer histopathological image-BreaKHis. The proposed method achieves an average accuracy of 92.96% and a micro average AUC value of 0.9588 for eight-class classification, and an average accuracy of 97.16% and an AUC value of 0.9713 for binary classification on BreaKHis dataset.
For the sake of verify the universality of the proposed model, we also conduct experiments on MIAS dataset. An excellent average classification accuracy is 98.79% on MIAS dataset. Compared to other state-of-the-art methods, the experimental results demonstrate that the performance of the proposed method is superior to that of others. Furthermore, we can analyze the influence of different types of features on the proposed model, and provide theoretical basis for further optimization of the model in the future.
Orthogonal upconversion nanoparticles (UCNPs) have received increasing research attention due to their unique optical performance. However, the establishment of the orthogonal UCNPs emitting primary ...red/green/blue (RGB) UCL remains a great challenge. Herein, RGB‐switchable NaErF4‐cored core–multishell UCNPs (RGB‐UCNPs) are constructed through the coordinative implementation of outside‐in cascaded photon harvesting and interfacial energy‐transfer (IET) inhibition. By modulating several constructional factors, the power‐density‐independent red, green, and blue luminescence in high color purity is readily available in the elaborated RGB‐UCNPs under 1550, 808, and 980 nm excitations, respectively. Notably, full‐color emissions, involving red, yellow, green, cyan, blue, magenta and white, are dynamically implemented in this single architecture through manipulating the excitation power of the combined 1550/808/980 nm lasers. These emission profiles of RGB‐UCNPs make them promising in broad photonic applications. As a proof of concept, the viability of the logicalized information encryption strategy and smartphone APP‐assisted multimodal luminescent QR code anti‐counterfeiting technique is demonstrated, which are both established based on directly useable RGB‐UCNPs with assistance of a self‐built combined laser system. This work not only opens up a new avenue to gain RGB‐switchable UCL and flexible full‐color output in a single nanostructure, but also provides enlightenment for developing the application scenarios of multicolor‐tunable UCNPs.
The excitation‐power‐independent orthogonal RGB‐switchable NaErF4‐cored core–multishell UCNPs (RGB‐UCNPs) are elaborately constructed through the cooperative implementation of cascaded excitation‐photon harvesting and interfacial energy‐transfer (IET) restriction. The power‐density‐independent R/G/B emissions in high color purity together with full‐color emission with a wide gamut are readily accessible in the well‐engineered RGB‐UCNPs under ternary 1550, 808, and 980 nm excitations.